Public Citizen Testimony on Trump’s Critical Minerals Deals
Updated October 21, 2026
On September 2, 2026, Global Trade Watch Director Melinda St. Louis testified before the full House Ways and Means Committee at a hearing on “Strategic Partnerships to Secure Critical Resources and Supply Chains.”
She was the only witness to testify to the profound lack of transparency, harmful terms in the text made available for review, and risks of corruption and misuse of taxpayer financing in the web of critical minerals deals, frameworks, and memoranda of understanding (MOUs) the Trump administration has negotiated.
In her opening statement, she summarized the contents of a detailed written testimony and a recent Public Citizen report on the secretive web of Trump’s minerals deals.
St. Louis explained that it is a reasonable goal to seek diverse and reliable supply chains for minerals. But the way we do so matters. When we talk about “supply chains,” we are talking about real people: people who work in dangerous mines; Indigenous and other frontline communities whose land might be seized or polluted; children whose health may be harmed. These people are put even more at risk by trade and investment deals that don’t require transparency, that lack strong and enforceable protections, and that fail to respect the sovereignty and development goals of partner countries.
Her testimony focused on three areas:
- The lack of transparency in the Trump administration’s critical minerals deals,
- Harmful terms in the text we’ve seen, and
- Risks of corruption and misuse of taxpayer financing.
Several members of the committee asked questions of St. Louis and raised similar concerns. In opening the hearing, Trade Subcommittee Ranking Member Linda Sánchez (D-Calif.) emphasized many of the same points. She also mentioned two letters that she led, along with 50+ colleagues, letters that Public Citizen has also endorsed.
Rep. Sánchez asked St. Louis to elaborate on her testimony that the number of critical mineral deals that the Trump administration claims to have entered doesn’t match up with their public announcements about concluded deals. St. Louis explained that, instead of the 38 touted deals, a Public Citizen review of publicly available information could only confirm the existence of 27 minerals-specific frameworks or MOUs: 15 MOUs and 12 frameworks. Only two MOUs and three frameworks have been released by the U.S. government. A few others have been made public by the partner government.
Rep. Sánchez also asked about the concerning trend of the Trump administration inking “minerals for security” deals, in which the U.S. offers military support or security guarantees to a country experiencing conflict in exchange for privileged access to the country’s mineral wealth. St. Louis explained that, for example, the deal with the Democratic Republic of the Congo has no meaningful labor or environmental standards, despite the documented child labor and other human rights abuses.
Rep. Lloyd Doggett (D-Texas) highlighted the harmful pollution that poorly regulated mining causes, noting that in Trump’s agreements on reciprocal trade, the word “environment” appears only twice, in reference to the “business environment.” He asked St. Louis about corrupt practices both in mining activities abroad and in potential self-dealing by the Trump administration.
She explained that a firm linked to Eric and Donald Trump Jr. increased its stake in a company just days before it acquired a major stake in a Kazakhstan mine. That mine had been offered up to $1.6 billion in potential financing from the U.S. Development Finance Corporation and Export-Import Bank. The project deal was reportedly ushered through by Secretary Lutnick, including direct outreach between President Trump and Kazakhstan’s president. Secretary Lutnick signed a minerals MOU with the Kazakh government that same month, but it hasn’t been released.
Rep. Danny Davis (D-Ill.) asked how Trump’s adversarial approach to allies like Canada undermines efforts to build resilient critical minerals supply chains. St. Louis explained that once trust is broken, it may not be easy to repair, and countries will find other more reliable partners to do business with. She called for mutually beneficial agreements with high-road standards, unlike what we have seen so far in Trump’s coercive trade agreements or in his hostility to Canada.
Rep. Terri Sewell (D-Ala.) asked about the impacts of Trump’s international agenda on workers and the ability to enforce labor standards abroad. St. Louis emphasized that minerals deals should include a rapid response enforcement mechanism that builds off of the model that congressional Democrats forced Trump to insert into the USMCA. Such a model would hold individual facilities accountable to resolve abusive conditions swiftly.
Rep. Judy Chu (D-Calif.) asked for more detail on potential conflicts of interest within minerals deals and the impacts of this possible misuse of taxpayer money on everyday Americans. St. Louis explained that, in the midst of an affordability crisis and while claiming to be using trade policy to benefit regular Americans, Trump is instead signing secretive deals full of giveaways to the billionaire class.
Rep. Don Beyer (D-Va.) asked if President Trump’s trade agenda is actually supporting long-term partnerships to guarantee mineral supply chains. St. Louis explained that by coercing countries into secretive deals that lack congressional or civil society involvement by either country, U.S. credibility is undermined on the global stage. A more strategic minerals plan would start with a focus on demand reduction so less mining is needed overall and require high standards for labor and the environment.
Public Citizen will be submitting responses to Questions for the Record regarding this hearing in the coming weeks and looks forward to continued engagement with Congress to hold the administration accountable to the goal of securing reliable critical supply chains without sacrificing democratic principles of transparency, equity, and justice.
Update
On October 19, 2026 St. Louis responded to Questions for the Record from the Committee. The response can be found here.
You might be interested in
Stay Updated on Public Citizen
Follow Public Citizen
Support Our Work
Mining Billions in the Dark
- In early August, President Trump announced a $3 billion federal investment in critical minerals companies at a Department of State roundtable event with industry leaders.
- Among the companies participating were firms that are only nine months old, connected to the administration’s family members or donors, associated with past fraud or fraud allegations, or whose financial backers remain largely hidden from public view.
- Many of these decisions result from the Trump administration putting extraordinary financing authority in the hands of a small group of former private-equity executives, including the billionaire financier Deputy Secretary of Defense.
Introduction
In early August, President Trump announced a $3 billion federal investment in critical minerals companies at a Department of State roundtable event with industry leaders. “Very simply, we’re making mining great again,” he said.
The announcement, which included deals from the Defense Department’s Office of Strategic Capital (OSC), U.S. Development Finance Corporation (DFC), and the U.S. Export-Import Bank (EXIM), is the latest salvo in the administration’s sweeping spending on mining, including an unprecedented acquisition of ownership stakes in companies like Vulcan Elements, a start-up rare earth magnet company funded by a venture capital firm where Donald Trump Jr. is a partner. “Magnets. You gotta go with magnets,” Trump said at the State Department event, before Commerce Secretary Howard Lutnick praised Vulcan directly.
Like Vulcan, many of the companies receiving these deals are examples of potential cronyism or are still in the early stages of development. Several have limited commercial production histories, an unproven ability to operate at scale, or remain years away from generating the minerals the administration says are critical to national security. The federal government is increasingly placing taxpayer dollars—and in some cases, direct ownership interests—behind companies that have not demonstrated they can deliver. The result is a model where a small number of politically-connected investors are positioned to benefit from federal financing, while taxpayers absorb much of the risk.
Even the basic details of the August investments appeared to receive limited review, as the White House fact sheet misidentified the names of two of the companies that received a combined $110 million. One of the errors named a completely different company than the intended investment recipient the White House announcement meant to highlight.
It remains unclear how the mining companies were selected or evaluated for government financing, raising questions about what priorities are driving these investments. The deals come as changes to federal disclosure laws make it more difficult to evaluate company ownership and to trace who is benefiting. In August, the Treasury Department’s Financial Crimes Enforcement Network announced it would permanently reduce the information American companies are required to share about who owns them and would destroy the ownership data companies have already reported. This lack of information magnifies the potential for harmful conflicts of interest and makes it harder to determine who ultimately controls the companies receiving public money.
Other companies participating in the August announcement are firms that are only nine months old, connected to the administration’s family members or donors, associated with past fraud or fraud allegations, or whose financial backers remain largely hidden from public view. This comes on the heels of the White House rolling back disclosure requirements designed to protect investors from speculative projects. In March 2025, a White House executive order waived Securities and Exchange Commission (SEC) disclosure requirements for projects backed by federal agencies, including those that mandated publicly-traded companies demonstrate a mine can actually make money.
As the administration moves beyond traditional loans into equity stakes, the government is now responsible for overseeing projects in which it also holds a financial interest. The Defense Department has not outlined how it will prevent those interests from influencing its regulatory or contracting decisions, and that omission has already proven problematic.
The rush to finance private mining companies has already spurred bipartisan criticism. The Republican-led Committee on Armed Services warned earlier this summer that the Pentagon’s push to finance the critical minerals industry has a “high possibility for distorting markets or corruption,” and the Cato Institute cited the report as it cautioned that the rapid expansion of Defense investments circled relationships with Trump.
A Federal Bet on Unproven Ventures
Backing emerging technologies and early-stage mining companies with limited production history shifts risks from private investors onto the government. If those companies fail, taxpayers could be left footing the bill, while their executives walk away with the gains. For example, Strategic Bauxite USA—which the White House initially called Standard Bauxite, an unrelated Turkish company—was given $85 million as an equity investment from the agency’s new Economic Defense Unit.
Strategic Bauxite appears to have been first registered as a business in Wyoming in December 2025. Like 300,000 other companies, its listed address is a small brick building just off the main street of Sheridan. The storefront belongs to a company that farms registrations, many of which have been tied to scamming operations and shell companies perpetuating fraud. Wyoming law does not require the registering agent to disclose the identities of company owners or verify the businesses.
Strategic Bauxite does not appear to have a functioning company website, and its earliest press coverage appears to be when it purchased an existing bauxite mine in Guyana about a month ago. The financial terms were not disclosed, and it does not appear in any previous SEC records.
On his LinkedIn page, Roy Ostrom III is listed as the co-founder, and while his bio claims to have “over $5bn capital raised,” many of his past positions similarly lack public information. He is listed, for example, as a managing partner at Visby Management, a private investing company whose website is a single page displaying only the company logo, and he appears to be the only employee publicly associated with it. A search of SEC filings did not identify any filings with Roy Ostrom III as an executive, director, or beneficial owner.
Last year, Visby attempted to purchase a gold mine in Colombia. The Colombian mining company’s board declined to engage with them and ultimately proceeded with another buyer, so Visby’s financing claims were not independently verified in any regulatory filings. Though the source of Ostrom’s capital remains undisclosed, he previously helped expand a Colombian cannabis company. The Defense Department says its support for Strategic Bauxite will be matched with a $64.5 million “co-investment” from the private sector, but it did not include any details. Further transparency is needed, given that Donald Trump Jr.’s venture capital firm previously held an undisclosed stake in a mining start-up that received a Pentagon loan, after a White House adviser intervened on the company’s behalf.
It’s not the only example of a large bet on a company still in the early stages. Niron Magnetics, a Minnesota startup, received a $150 million conditional loan as part of Trump’s August announcement. While Niron Magnetics has previously received government grants to develop its rare-earth-metal-free magnets, this is the first time the company is receiving a loan from the government. Government grants, like the one Niron previously received through the Department of Energy’s SCALEUP program, are designed to incubate promising ideas without requiring certainty of commercial success. Debt financing, by contrast, requires repayment and ultimately a return on investment—making the government’s decision to lend $150 million substantially different from funding early-stage research. By law, OSC loans must consider national security and economic interests, if the financing would accelerate the project, and its creditworthiness, but those conditional details and loan terms are not required to be disclosed to the public. Niron has not yet launched full-scale production, and is still constructing its first plant in Sartell, Minnesota, with a second facility in the planning phase. The company’s CEO, Jonathan Rowntree spoke with the Trump administration about rapidly scaling up production to meet demand for U.S.-made magnets, and in February opened a D.C. office to expand its lobbying, hiring a former assistant to Republican Majority Whip Tom Emmer.
A similar story is playing out with the largest loan announced at the roundtable, a $1.4 billion conditional loan to Sila Nanotechnologies to scale its silicon-anode battery technology for AI and defense. While Sila previously received much smaller research and development support during the Obama and Biden administration to develop EV batteries, those sums pale in comparison. One of the company’s investors, venture capital firm 8VC, is owned by Joe Lonsdale, a Palantir co-founder who donated $1 million to the Trump-aligned America PAC.
Table 1: Deals Announced at Mining Roundtable
| Company | Federal agency | Financial Backing | Potential Issues |
|---|---|---|---|
| Strategic Bauxite | Industrial Base Analysis and Sustainment Program (IBAS) | $85M equity investment | Limited ownership transparency and corporate history. Strategic Bauxite USA was reportedly formed in Wyoming in December 2025 at an address tied to other shell companies. Its co-founder Roy Ostrom III’s investment history and sources of capital are not readily verifiable through public means. |
| Niron Magnetics | Office of Strategic Capital (OSC) | $150M conditional loan | Early-stage tech company. Niron is a private company that has not yet established full-scale commercial production. Niron expanded its lobbying activities this year by hiring former assistant to Republican Majority Whip Tom Emmer. |
| Sila Nanotechnologies | OSC | $1.4B conditional loan | Early-stage tech company. Venture capital firm 8VC is an investor in Sila. Joe Lonsdale, who donated $1M to Trump’s America PAC, founded 8VC. |
| Sunrise Energy Metals | OSC | $400M conditional loan | Private gain through stock appreciation. Friedland, the company’s largest shareholder, exercised 3 million options just days before the loan announcement. As of Aug 16, Sunrise’s stock has risen more than 7,000% since Trump took office, increasing Friedland’s net worth by hundreds of millions. |
| 5E Advanced Materials | Export-Import Bank (EXIM) | $8M loan | Political access. 5E paid $120,000 to the Bernhardt Group, the lobbying firm founded by former Trump Interior Secretary David Bernhardt, to lobby on issues including the company’s effort to obtain critical-mineral status for boron. |
| Westwater Resources | EXIM | $25M loan | Political access. Westwater has a longstanding financial relationship with Cantor Fitzgerald, the former firm of Commerce Secretary Howard Lutnick. Cantor served as its ATM equity-sales agent and investment banker. |
| Global Advanced Metals | EXIM | $25M loan | No competitive bidding. The company previously received two sole-sourced—meaning no competitive bidding—contracts from the Trump administration. Despite the purported goal of expanding domestic supplies, this loan is to expand processing of tantalum and niobium, neither of which are mined in the United States. |
| Harena Rare Earths | Development Finance Corporation | $4.8M project development funding | Limited corporate history. Founded in 2022, Harena Rare Earths is planning its first mine in Madagascar. With limited mining experience, the company identified MP Materials and USA Rare Earth as downstream partners, both linked to the Trump administration. Paul Jacobi, a partner at the Wexford Capital, an investor in Harena, attended a high-profile crypto dinner hosted by Donald Trump in May 2025 |
Source: Public Citizen research
Perhaps the most risky move of the Trump administration’s mining mania is its aggressive push to advance deep-sea mining, an unproven, nascent technology with profound environmental harms. At the industry roundtable, Tom Albanese, a pioneer of the speculative effort to extract minerals from the seafloor, presented Trump with a solid gold replica of a deep-sea polymetallic nodule. The real versions are mineral-rich rocks on the ocean floor that contain critical minerals like nickel, cobalt, and copper, but no deep sea mining operation has successfully launched anywhere in the world. The absurdity of the prop shouldn’t obscure the seriousness of the administration’s support for the industry. Despite several countries calling for a moratorium, the Trump administration is moving aggressively to open U.S. waters to the industry. The Interior Department’s Bureau of Ocean Energy Management is currently considering selling deep-sea mining leases off the coast of Virginia to Odyssey Marine Exploration, the company currently merging with Albanese’s American Ocean Minerals.
Albanese has previously faced allegations of inflating asset values. While he was CEO of Rio Tinto, he helped sell investors on the promise of a Mozambique coal mine. Rio Tinto bought it for $3.7 billion but sold it for just $50 million several years later. The SEC charged Albanese and Rio Tinto with fraud in 2017. The company ultimately paid $28 million to settle the case, with Albanese personally paying a fine of just $50,000—a slap on the wrist for a multi-millionaire. This year, Trump’s SEC dismissed the fraud case against the company’s former CFO.
Political Connections, Public Money
For some companies, the critical minerals push has already translated into huge profits. The biggest emerging winners are those with connections to the White House. At the roundtable, Sunrise Energy Metals received a $400 million conditional loan for scandium production for military uses. Last year, the company gained momentum from the administration’s push to expand domestic mineral production, including a letter of interest from the U.S. Export Import Bank, with what the company called an “attractive repayment term.” Since Trump took office, Sunrise’s stock has increased by 7,000 percent.
The company has courted ties to Trump’s business world: Its board appointed Peter Arnell, a designer and branding executive, as a non-executive director in May, just two days after the White House created a novel position for him as U.S. Chief Brand Architect. Arnell’s advertising firm worked for Trump in the 1990s, and he has previously described Trump as “one of the greatest choreographers of business and image that the business world has ever seen.”
Another key figure behind Sunrise’s ascent is Robert Friedland, one of the main beneficiaries of Trump’s mining policy. The chairman of Sunrise Energy’s board and the company’s largest shareholder, Friedland’s storied past shows an appetite for lawbreaking and risk-taking for personal gain. He earned the nickname “Toxic Bob” because of two massive cyanide spills involving his ventures—one in Colorado and the other in Guyana. Previously, he was arrested for possessing and attempting to sell what was then $100,000 worth of LSD, and convicted to a two-year federal prison sentence. As a young man, Friedland and Steve Jobs shared a spiritual living experience on an apple farm that would go on to inspire the name of Jobs’s technology company.
At the roundtable, Friedland sat center stage, just to the right of Howard Lutnick. Over the course of Trump’s second term, Friedland increased his holdings in Sunrise significantly, roughly tripling his position. At the end of 2024, he held almost 12 million shares; he now holds over 33 million, a stake worth about $381 million. Just a week before the August announcement, Friedland exercised an option to buy an additional 3 million shares. The rapid increase in his holdings and the appreciation in Sunrise’s stock represents hundreds of millions of dollars in gains.
It’s not the only way Friedland has profited from the Trump administration. At the roundtable, Trump also announced a potential $1 billion investment in another of Friedland’s ventures, Ivanhoe Electric, which is developing a copper mine in Arizona. Trump called it “one of the biggest anywhere in the world.” Separately, I-Pulse, Friedland’s venture into semiconductors, received a $250 million equity investment through the Department of Commerce’s CHIPS Act this year. And another of his companies, Ivanhoe Mines, stands to gain from its potential role as a supplier for Project Vault, the administration’s new critical mineral stockpile.
Political access has become a valuable resource in the race for government-backed funding for critical minerals. 5E Advanced Materials, which received a $8 million loan from EXIM in the August announcement, previously successfully lobbied to get a critical minerals designation for boron. They paid $120,000 to the Bernhardt Group, Trump’s former Secretary of the Interior’s lobbying shop, to make their case in Washington. Boron was added to the critical minerals list in 2025. This kind of critical mineral determination can open up billions in funding from the Department of Energy and the Defense Department, as well as expedited permitting.
Other companies have pursued a different form of access: Relationships with the financial institutions and advisers close to Trump. At the round-table, James Litinsky, CEO of MP Materials, sat a few seats down from Trump and Lutnick. His company received $400 million from the Defense Department last year in exchange for a 15 percent government equity stake. At the event, Litinsky claimed the “American taxpayer has already benefited” from the Pentagon becoming his largest investor. Litinsky himself, along with Trump’s friends and allies, have profited handsomely: MP Materials stock climbed more than 300 percent between Trump’s inauguration and last fall, when Litinsky sold off roughly $40 million in shares, cashing in on the federal investment. Stephen Miller similarly sold his MP Materials shares last August after the deal was announced, and billionaire Gina Rinehart, an avid Trump supporter and Mar-a-Lago guest, is its largest private backer.
Meanwhile, other mining companies have pursued relationships with Cantor Fitzgerald, the investment bank formerly led by Howard Lutnick, before ownership was transferred to his children. One of the loan grantees announced at the roundtable is Westwater Resources, which has had multiple financing relationships with Cantor. Until 2024, the bank served as Westwater’s at-the-market equity sales agent, helping the company sell newly issued shares. In 2024, Westwater reported $1.1 million in net proceeds from approximately 2.3 million shares sold through the arrangement. In 2025, the company again engaged Cantor to help arrange a proposed $150 million debt syndication for its graphite plant. Westwater went on to receive a letter of interest from EXIM last year, before being given $25 million in the August announcement.
Fast-Tracking Pollution for Profit
While lining the pockets of his wealthy friends, Trump’s mining policy is set to create lasting environmental damage. The Trump administration has rapidly expanded the use of FAST-41, a federal permitting coordination process designed to speed reviews for major infrastructure projects. Since taking office, the administration has added twenty mining projects, giving them access to coordinated timelines and streamlining federal review. Several of the companies mentioned at the roundtable have benefited from the administration’s eagerness to cut red tape at the cost of the environment and surrounding communities, including Westwater Resources.
It plans to develop a roughly 41,900-acre graphite deposit in Alabama, building a processing facility to supply material for lithium-ion batteries. In March 2026, the project was added to the FAST-41 program, expediting its permitting process.
Despite the significant federal investment, the project remains in the early stages of environmental assessment, which isn’t expected to conclude until June 2027. Pending reviews include a Clean Water Act analysis of impacts to wetlands and waterways, a state discharge permit covering stormwater and mining-related water, and an Endangered Species Act consultation.
Those impacts could be substantial. The company’s plans involve converting a large forested area into an open-pit sulfide-ore mine near streams and wetlands, potentially contaminating groundwater with acid mine drainage and heavy metals. The project would also generate large quantities of tailings, which can contain toxic metals and require permanent storage in tailings facilities. The nearby Coosa River basin is habitat for several threatened and endangered species, including the federally listed Alabama live-bearing snail.
Another attendee was the CEO of Energy Fuels, Ross Bhappu, who leads an established company that is already a major beneficiary of the Trump administration’s mining subsidies and public lands giveaways. This summer, Trump reduced the size of Bears Ears National Monument by more than 90 percent, allowing mining companies access to about 1 million acres of federal land in southern Utah, just miles away from a uranium mill owned by Energy Fuels. The removal of protected lands not only paved the way for environmental harms, but Bhappu faces Congressional allegations of using the opportunity for financial advantage through potential insider trading. Five days prior to the announcement, before the downsizing of the monument was publicly announced, Bhappu increased his ownership by 41 percent. The Trump administration awarded Energy Fuels a $725 million loan from OSC this June, and fast-tracked permitting for its uranium mine in New Mexico, the Roca Honda Project.
One of the most controversial mining projects added to FAST-41 by Trump is Resolution Copper in Arizona. After a decades-long legal battle over environmental concerns and religious disputes raised by the San Carlos Apache Tribe, Trump’s Forest Service finalized a land swap this spring that granted Australian mining companies Rio Tinto and BHP land from Tonto National Forest with a significant copper reserve in exchange for land elsewhere in Arizona. It was originally authorized by Congress in 2014 in a provision snuck in by then-Sen. John McCain, who received campaign contributions from Rio Tinto. Subsequently, the Biden administration blocked the transfer, rescinding the environmental review. The Trump administration reissued an environmental impact statement that eventually cleared the land swap this March. Victoria Peacey, president and general manager of Resolution Copper, who attended the August roundtable, credited the Trump administration with finally moving the project across the finish line: “We started permitting in 2005, and if it were not for your administration–we finally got it over the line this year,” she told Trump at the event.
Table 2: Attendees at Mining Roundtable
| Trump Administration | ||
| Attendee | Roles | Conflicts of Interest |
| Howard Lutnick | Secretary of Commerce; previous chairman of Cantor Fitzgerald | Lutnick’s former firm Cantor Fitzgerald has been tied to several of the mining deals that the administration has announced, including USA Rare Earths and Cove Kaz Capital’s Tungsten-mine in Kazakhstan. |
| Doug Burgum | Secretary of Interior | Burgum is invested in several energy companies and leases land in his home state, North Dakota. He has a close personal relationship with Harold Hamm, an oil executive and one of Trump’s biggest donors. |
| Jarrod Agen | Director of National Energy Dominance Council | Agen, a former Lockheed Martin employee, continues to hold stock in the company while directing the White House council that sets national policy affecting defense contractors. |
| Private Sector | ||
| Attendee | Roles | Preferential Access |
| Robert Friedland | Board chairman of Sunrise Energy Metals; founder and major shareholder of Ivanhoe Capital, I-Pulse, and Ivanhoe Electric | In May 2025, Friedland met with Trump and Saudi Crown Prince Mohammed bin Salman in Riyadh. In February this year, Friedland met with Trump for the launch of Project Vault, which Ivanhoe mines will supply. |
| Ross Bhappu | CEO of Energy Fuels | Bhappu bought the largest block of Energy Fuels stock in company history just days before Trump cut the Bears Ears National Monument, granting mining companies access to uranium-rich land. House Democrats opened an investigation into potential insider trading. |
| James Litinsky | CEO and chairman of MP Materials | Litinsky’s MP Materials received a $400 million equity stake from the Pentagon, a deal negotiated in part by a lawyer who simultaneously managed Deputy Defense Secretary Stephen Feinberg’s personal trust. |
| Tom Albanese | Chairman of American Ocean Minerals; former CEO of Rio Tinto | Albanese’s company American Ocean Minerals stands to directly benefit from Trump’s EO fast-tracking seabed mining permits. After Odyssey Marine Exploration—the company that later merged with American Ocean Minerals—submitted an unsolicited request for an offshore lease, BOEM opened a Request for Information about a potential lease sale off the coast of Virginia. |
| Victoria Peacey | President and General Manager of Resolution Copper, a joint venture of Rio Tinto and BHP | Peacey’s Resolution Copper project’s permitting was fast-tracked, and its long-contested Oak Flat land transfer completed after Trump officials intervened. |
Source: Public Citizen research.
The Revolving Door at Pentagon’s Investment Arm
Many of these decisions result from the Trump administration putting extraordinary financing authority in the hands of a small group of former private-equity executives. Stephen Feinberg was confirmed as the Deputy Secretary of the Defense Department last March, joining from his position as CEO and co-founder of investment firm Cerberus Capital Management. Feinberg hired former Cerberus employee George Kollitides II to lead the Defense Department’s novel Economic Defense Unit, which signed the Strategic Bauxite deal. Kollitides also serves as vice chair of the Defense’s Office of Strategic Capital investment committee, which authorized conditional loans to Sila, Sunrise, and Niron. In addition, Feinberg brought former Cerberus employee David Lorch to lead OSC.
Cerberus, meanwhile, has aggressively expanded its critical minerals portfolio in recent years, investing in lithium and zinc companies, among others. Those investments closely track the sector now being prioritized by Trump’s minerals agenda. Additionally, the lawyer managing Feinberg’s personal irrevocable trust and charitable foundation also brokered last year’s MP Materials deal with the Pentagon. Given this shared history, greater transparency is needed to ensure these decisions are driven by national priorities, not private relationships.
Kollitides is technically classified as a special government employee, allowing him to simultaneously continue private sector work. That also means no ethics disclosure, agreement, or recusal is available publicly for Kollitides, making it difficult to ascertain whether potential conflicts were identified and addressed before federal funds were committed.
This concern is heightened by Kollitides’s relationship with D. Boral Acquisition I Corp., a special-purpose acquisition company where Kollitides served as director nominee in January 2026, shortly before being appointed to the Defense Department in early February. Just six days after his appointment, D. Boral’s SEC prospectus featured Kollitides’ brand-new Pentagon job, saying he was “responsible for all Department of War investing and economic activities.” D. Boral’s board member, Kevin McGurn, became the interim CEO of Trump Media in April. And in June, D. Boral recommended buying Westwater Resources stock.
The investment bank sponsoring the acquisition company, D. Boral Capital, helped underwrite another special-purpose acquisition company backed by Donald Trump Jr. and Eric Trump last year, which went public in a $345 million IPO.
The potential benefits to Trump’s friends and family from this kind of deal-making, and his personal intervention in mining contracts while president, led several Democratic senators last month to formally request the Department of Defense and other federal agencies to preserve documents related to many of the critical minerals companies participating in the roundtable, including American Ocean Minerals Corporation, Energy Fuels, Westwater, and Vulcan. Several federal agencies have been reported destroying official communications in violation of the Federal Records Act, including, most recently, the Interior Department routinely shredding documents and using disappearing Signal messages.
The administration’s push to secure critical minerals is outpacing oversight. Now, billions of dollars of public resources are flowing toward a select group of companies and executives. Rather than quantifying proven demand or supporting mineral recycling and circularity, the Trump administration is giving away taxpayer dollars without properly vetting the companies involved, and doing little to actually ensure long-term supply chain resilience.
That makes it urgent to understand how these decisions are being made, what relationships helped along the way, and whether these government investments are actually in the interests of the American people.
You might be interested in
Stay Updated on Public Citizen
Follow Public Citizen
Support Our Work
Public Citizen Comments to the Texas Senate Water, Agriculture, & Rural Affairs Committee Regarding Assessing Water Demands of Energy-Intensive Technologies
To: Chairman: Sen. Charles Perry and Water, Agriculture, & Rural Affairs Committee Members
CC: Senator Kevin Sparks, Senator César Blanco, Senator Roland Gutierrez, Senator Adam Hinojosa, Senator Lois Kolkhorst, and Senator José Menéndez
From: Rita Beving, Public Citizen
Re: Assessing Water Demands of Energy-Intensive Technologies in Texas
Public Citizen appreciates the opportunity to submit comments to this committee on the large water demands of energy-intensive technologies in our state. We will focus our comments on both the direct and indirect use of water for data centers.
Data from a recent August 2026 industry publication indicates there are 237 operating data centers in Texas.1 Hundreds more have been announced or are under construction.
Today, a mid-size data center can use up to as much as 300,000 gallons per day for cooling its servers, while a large or hyperscale facility can use up to 5 million gallons per day. 2
According to estimates by the Houston Advanced Research Council (HARC), the direct (on-site cooling) and indirect (electricity generation) water consumption for existing data centers may have totaled as much as 25 billion gallons in 2025. 3
LACK OF REPORTING
The Texas Water Development Board (TWDB) initially adopted its 2027 water plan on July 23rd of this year. Currently, the TWDB five-year plan does not account for individual data center water use in its water plan, although it had issued a survey request to the industry for this and other information this past year. When the agency testified in June, it stated that of the 341 surveys sent this year, only 17% of the data centers responded to the agency.4 Yet the next water plan the agency will be reviewing is for 2032.
The TWDB has been sending data centers surveys since 2023. Only one-third of data centers surveyed in 2024 responded. In 2025, just 17% responded, the same percentage as last year. Though required by law to respond (see Statute 16.012m of the Texas Water Code),5 the consequences for data centers and other entities in noncompliance for answering the agency survey are minimal resulting in a Class C misdemeanor with a maximum fine of $500. 6
Likewise, the Public Utilities Commission also sent out a survey this year in collaboration with the TWDB. It asked operators to provide information about water consumption, water sources, cooling systems, electricity demand and whether facilities are connected to the state’s electric grid or have
their own on-site power source such as a gas-fired power plant. In its June testimony before Natural Resources, the agency testified that it had received responses from only 28 companies, representing 92 data center facilities.7 It was unclear from the PUC’s testimony whether the actual number quoted was the total number of surveys the agency had actually sent out.
INDIRECT WATER USE
Too often the question about data center water usage focuses soley on the actual cooling needs of a facility itself, when there is a need to look at the water-energy nexus of an entire data center campus, both on and off-site.
Data centers are now recognized as the most energy-intensive elements of the digital economy. The total electrical power going into a single data center rack is expected to approach 1 MW by 2028 —equivalent to the peak hour energy use of roughly 250 U.S. homes flowing into a computing unit the size of a bookcase. 8
A February report by Bluefield Research, “The Water-Power Nexus: How Data Centers are Reshaping the U.S. Water Landscape,” notes that the biggest water story is not what is happening onsite with data center facilities, but rather with the development of power plants that accompany them. By 2030, data centers are expected to account for 8.9% of total electrical demand nationwide, up 4.4%, or more than double of 2023. 9
According to an Environmental Integrity Project report, data centers connected to the grid have found that power plants use between 75 to 92% of all the water needed to keep data centers running. 10
In a Global Energy Monitor report just released August 25, 2026, U.S. gas-fired power proposals tied to data centers nearly doubled in the first half of this year. 11 Texas accounts for nearly one-third of this explosive growth with a 51% increase in development during the last 6 months, more than any country worldwide. 12
ERCOT is now tracking more than 438,000 MW of large load requests to connect to the grid. Almost 89% of those requests are from data centers.13 In 2025, ERCOT received 225 large load interconnection requests through mid-November, compared to the 152 requests received during the entire 2022-2024 period combined. 14
As Texas leads the way in the development of data center campuses and the power that supports these facilities, Texas needs to look holistically at the water needed for both data center development and its supporting power generation.
Due to the backlog of connection requests, developers are pivoting to behind-the-meter power and faster-to-install engines to skip supply backlogs. In Texas, many of these developers are turning to gas turbines for onsite backup generation.
Gas turbines are the most expensive and critical component of a natural gas plant, accounting for almost half of the gas generation proposals for data centers nationwide.15 Both simple-cycle and combined-cycle gas turbines are often used for power generation, utilizing water in varying capacities.
In a basic simple-cycle gas turbine plant, water is used to cool the turbine and auxiliary equipment, and is not used for electric generation. The hot exhaust, typically around 600 °C, is routed through a closed‑loop water circuit that absorbs heat from the turbine housing, bearings, and lubricants. After collecting heat, water is pumped to a cooling tower where it releases the thermal load into the air before returning to the system for reuse. 16
Combined-cycle gas turbines, which combine gas and steam-powered turbines, can use an average of 2800 gallons per megawatt hour.17 In a combined‑cycle natural gas plant, water is essential for electricity generation because the exhaust heat from the gas turbine is captured in a heat‑recovery steam generator that produces steam to drive a second turbine. 18 This steam cycle turns water into a power‑producing medium, distinguishing it from a simple‑cycle plant where water serves only for cooling. 19
The 5800-acre FERMI Matador project near Amarillo plans to utilize 90 combined-cycle gas turbines, 16 cooling towers, and 6 emergency generators.20 The full build out of the project is for 11 GW and will be supported by other forms of power including nuclear and solar.
The 1100-acre Stargate project in Abilene has now projected utilizing a total of 51 natural gas turbines. In 2024, the Stargate project initially filed for 10 simple cycle turbines21 and 62 diesel generators. It began operation without public input by securing a minor source permit from TCEQ for this portion of its facility. In July 2026, the project filed for an additional 41 turbines and 18 generators. 22 At full buildout, Stargate will have a capacity of 1.2 GW of power.
The 2000-acre Nexus data center project in Hubbard is implementing the first phase of its campus with 500-612 MW of generation capacity. In full build out, it will be supported by 7.2 GW of onsite natural gas generation. It is utilizing 31 combustion engines, some combined-cycle capable, with 24 reciprocating engines and 6 diesel generators. 23
Due to the lack of reporting from data center projects, the water use for gas turbines coupled with cooling needs for an entire facility’s campus will be unknown unless the water use is tracked in a more complete and comprehensive manner. The tracking of water use needs to include a developer’s projections for the final buildout of these facilities.
DIRECT WATER USE
Most often in discussions about the water needs for data centers, the focus is on the direct water use in the cooling of the servers.
Data centers consume a considerable amount of power which generates heat. The more equipment a facility has, the more heat it generates. Servers and processes within a data center need to maintain optimum temperatures and humidity. Otherwise, excessive heat may damage equipment, causing servers to malfunction or even stop working.
Though cooling methods are evolving, data centers typically use either water- or air-based cooling systems. Water-based systems that are reliant on evaporative cooling use less electricity, while air-based systems consume less water but require more power. Below are some of the typical cooling technologies utilized for data center facilities:
- Air-cooling systems release waste heat directly into the ambient air without using water.
- Direct-to-chip cooling involves circulating liquid directly to the heat-generating components of servers.
- Immersion cooling involves submerging entire servers or components in a non-conductive liquid coolant.
- Closed-loop systems recirculate chilled water or coolant within a closed circuit minimizing water loss and increasing efficiency.
- Hybrid systems combine multiple cooling methods, using evaporative cooling only when necessary. 24
There has been some discussion by some county and state officials about mandating closed loop water cooling systems to minimize water loss. While these systems have been marketed as waterless solutions reducing direct water use to data centers, the systems can take millions of gallons of water to fill. And at some given point in time, that water will need to be replaced.
Closed loop systems can shift the water burden instead to upstream thermal electric demand. Industry experts note that closed loop systems may utilize as much as 40% more electricity than evaporative cooling systems for the same cooling output. 25 This increase may be even higher for hyperscale facilities since power demand for gas fired behind-the-meter generation can reach gigawatts of power. 26
As the data center industry evolves there are other emerging cooling technologies that may improve water efficiencies. But power-hungry solutions coupled with thermal power generation do not solve the water use problem.
To quote an industry expert, “Waterless cooling that drives up electricity demand, and thereby increases water use elsewhere, is not a sustainable solution. It is a redistribution of impact.”
REPORTED OR PERMITTED WATER USE
Below is a list of permitted or reported water requests of the larger data center projects in Texas. It is unclear whether these water requests include the additional water needed for the electric generation for these projects:
FERMI America Matador Project – 2.5 million gallons per day (Permitted-2024)
FERMI America has been approved for up to 2.5 million gallons per day for its 5800-acre Matador project (or approximately 912.5 million gallons annually) with the City of Amarillo and has an option with the city to increase that volume to 5 million gallons per day as the projects builds out.28
FERMI’s data center has been permitted for 90 gas turbines, and in March, submitted an application for an additional 51 turbine units. 29
Note: On August 17, the City of Amarillo implemented its Stage 1 drought contingency plan after its reservoir dropped below 60 percent. During the month of July, FERMI reported its use of 310,000 gallons of water. 30
Meta Platforms El Paso Campus – 1.5 million gallons per day (Permitted)
The Meta Platforms data center is building on a 1,000-acre parcel in northeast El Paso. The water supply agreement between Meta and the city-owned water utility permits the project to use as much as 1.5 million gallons of water per day with a capped allocation at 2.5 million gallons per day. 31
The campus plans on using 813 modular gas generators totaling 366 MW of power. 32
Google’s Midlothian Data Center – 263 million gallons annually (Reported-2025 use)
Google’s 375-acre data center campus in Midlothian first broke ground in 2019, later purchasing an additional 165 acres. In 2024, Google added a fourth building to its campus. 33 According to Google’s own 2026 Environmental Annual Report, the data center withdrew a total of 263 million gallons of water in 2025, consumed 220 million gallons and discharged a total of 43.4 million gallons. In January 2026, it started construction on a fifth building. 34
Nexus Hubbard Plant – Hubbard, Texas – 100 million gallons annually (Permitted)
A 600 MW natural gas-powered data center is planned on a 2000-acre site outside Hubbard with roughly 153 MW of emergency diesel generation at the site. 35 In February 2026, Nexus requested an operating amendment to its annual permit to increase its pumping from to 115 million gallons annual water request from the Prairielands Groundwater District. 36
The plant’s profile notes that it has 31 combustion engines, some combined-cycle capable, 24 reciprocating engines, and 6 diesel generators. 37
RECOMMENDATIONS
Public Citizen recommends the following as a response to the water challenges the state may face with the increase of water not only for data centers, but also for the thermal generation that supports it:
1) Require Detailed Reporting of Direct and Indirect Water Use
Data center facilities should be required to report the total water demand for their entire campus including that needed for power generation or other critical infrastructure for their total buildout. Many campuses phase in their activities, not telling cities, counties, or the state what their expected buildout and final resource requirements will be.
Reliance on voluntary surveys should be replaced with mandatory reporting requirements.
Water data for both current plus electrical demand buildout should be reported. This reporting should be mandated and sent to the PUC, TCEQ, and the Texas Water Development Board (TWDB).
Failure to provide this information in a timely manner to the aforementioned agencies should be met with significant fines, and if needed, denial or delay of a facility’s requested permits. Currently the failure to report to the TWDB is a Class C misdemeanor and a $500 fine.
Cities and counties should be able to secure the same information outlined above in their development agreements.
Finally, this information should be publicly available, and the state code should be amended to ensure that utilities are able to release water use information for these large industrial users.
2) Include Large Users in Water Planning
The TWDB needs to track both data center and electrical infrastructure water usage. Regional planning groups and their respective utilities should be required to track and record large water usage within their respective areas. Likewise, this data should be reported by groundwater districts and river authorities. This data should be publicly available.
The reported actual and projected water usage needs to be utilized in the TWDB’s water planning process which occurs every 5 years. The 2027 water plan has been finalized, and, therefore, the next opportunity for any data to be included would not be until 2032. Data collection needs to begin now to have meaningful projections for future water planning.
3) Energy and Data Center Water Use Should be Metered
Transparency and accessible data make for better planning both for local communities and for the state. All utility-provided energy and water should be metered. All groundwater use for data center campuses should also be metered.
4) Use of Non-Potable Water
Data centers and power plants should pursue gray water, treated wastewater, brackish water or other non-potable water sources first as their primary sources for their cooling systems.
5) Utilize Energy Efficient Technologies
Water demand may be decreased with utilizing the most energy efficient technologies appropriate for a data center facility. The legislature could require the use of Energy Star equipment. The state should explore developing water efficiency standards for data centers and other large industrial facilities.
6) Large Water Users Should Shoulder Infrastructure Costs
- Large industrial and data center users should bear the costs for any new infrastructure required to service them including water distribution, water supply, water treatment, and associated financing costs. This may be done by increased tiered rate charges for large water users or some kind of surcharge.
7) Pretreatment of Wastewater May Be Required
- Wastewater from data centers may contain biocides, glycols, corrosion inhibitors, traces of heavy metals, and PFAs. These chemicals can be somewhat concentrated when discharged and could strain or even overwhelm local water treatment facilities. This wastewater can also be quite warm, up to 140 degrees, due to the heat absorbed through the cooling of servers. The state should consider, especially for large scale data center operations, requiring pretreatment of water before it is discharged back into a system or into waters and streams.
8) Every Texas County Needs to Be in a Groundwater District
The time has also come to require that every county be in a groundwater district in order to protect aquifer and water resources. According to the TWDB, there are 98 groundwater districts that cover 73% of the state.38 For those remaining counties or portions thereof without a groundwater district, a data center or other industrial facility could drill water wells in those areas with no accountability for the volumes of water extracted due to the lack of a permitting authority.
9) Existing Groundwater Districts Need Expanded Authority and Protection
Recent years have seen the increased threats of litigation against groundwater districts. Past hearings in Natural Resources have pointed to a lack of data, funding and staff to meet legal and other ongoing challenges. Groundwater districts need to be shielded from liability to meet drought contingency limits on pumping or to meet future conditions. The state should consider granting groundwater districts additional authority within Chapter 36 of the State Water Code.
10) Update the Rule of Capture
With the rapid development of data center sites and the increasing efforts to commoditize water in Texas, it is time to update the Rule of Capture to protect aquifers and protect water resources from the overpumping of water by individuals or entities. This is especially troubling in areas with no groundwater districts to set limits on water withdrawals.
2 https://natureforward.org/data-centers-and-water-use/
4 https://house.texas.gov/videos/22709 (Temple McKinnon testimony)
5 https://www.twdb.texas.gov/waterplanning/waterusesurvey/index.asp
8 https://compass.beg.utexas.edu/files/publications/Water_Requirements_for_DC_White_Paper.pdf
10 https://environmentalintegrity.org/news/74-gas-power-plants-proposed-for-u-s-data-centers-could-release-as-much-climate-pollution-as-australia/
13 https://www.ercot.com/news/release/06182026-puct-approves-ercots
20 https://insideclimatenews.org/news/25082026/texas-data-center-gas-projects/
21 https://www.citriniresearch.com/p/stargate-a-citrini-field-trip-bde
22 https://www.aterio.io/blog/stargate-files-for-phase-3-expansion-in-abilene
23 https://checkmarkpro.com/assets/Case-Study-Nexus-Data-Center-Campus-Hubbard-Texas.pdf
24 https://waterwelljournal.com/data-centers-and-groundwater/
28 https://amarillotribune.org/2026/08/27/asked-answered-how-much-water-is-fermi-america-using/
30 https://amarillotribune.org/2026/08/27/asked-answered-how-much-water-is-fermi-america-using/
34 https://sustainability.google/reports/google-2026-environmental-report/
35 https://checkmarkpro.com/assets/Case-Study-Nexus-Data-Center-Campus-Hubbard-Texas.pdf
36 https://www.prairielandsgcd.org/wp-content/uploads/2026/02/OP-Hearing-Notice_02.17.26.pdf
37 https://checkmarkpro.com/assets/Case-Study-Nexus-Data-Center-Campus-Hubbard-Texas.pdf
38 https://www.twdb.texas.gov/groundwater/conservation_districts/facts.asp
Stay Updated on Public Citizen
Follow Public Citizen
Support Our Work
Outrage of the Month: Vaccine Chaos and the Resurgence of Measles
Health Letter, September 2026
By Robert Steinbrook, M.D.
Director, Public Citizen's Health Research Group
If you’re not outraged, you’re not paying attention!
Read what Public Citizen has to say about the biggest blunders and outrageous offenses in the world of public health, published monthly in Health Letter.
As of Aug. 20, 2026, 2,777 confirmed measles cases had been reported in the United States this year; 2,289 confirmed cases were reported in all of 2025. The 2026 cases represent a 35-year high, were reported by 46 states and the District of Columbia, included 193 hospitalizations (7% of all cases), and most were associated with a measles outbreak. The continuing epidemic started in Texas in early 2025 with two imported cases and subsequently spread throughout the country.
Of the 2026 measles cases, 94% were in individuals who were unvaccinated or had an unknown vaccination status, 3% had one dose of the measles, mumps and rubella (MMR) vaccine and 3% had two doses, which is the recommendation for children. The cases included 522 (19%) in children younger than 5 years and 1,314 (47%) between 5 and 19 years. To date, two deaths have been reported in 2026; three deaths were reported in 2025.
During the 2024-2025 school year, MMR vaccination coverage was 92.5% for kindergartners. Vaccine coverage has been decreasing and is well below the 95% threshold for coverage with two MMR doses that prevents outbreaks and protects communities against an exceptionally contagious disease that can cause serious illness and death.
In November 2026, the United States is likely to lose its measles elimination status, which it has held since 2000. Measles elimination status entails an uninterrupted and prolonged absence of endemic transmission — that is, no continuous transmission chain for measles for at least one year.
Against the backdrop of an ongoing public health crisis, it is mystifying that the Trump administration in August 2026 issued a tone-deaf executive order calling into question the childhood vaccine schedule. Among other troubling provisions, the executive order called for the MMR vaccine to be split into three separate, single-disease shots that should be administered at separate medical visits.
The MMR vaccine is one of the crowning achievements of modern medicine. A split version of the vaccine doesn’t exist in the United States, is not supported by evidence-based science, and manufacturers have no interest in making one. Splitting the vaccine would further decrease coverage at a time when uptake is already too low, subject children to more needlesticks, increase health care costs and leave more children vulnerable to infectious diseases that vaccines can prevent.
The first individual measles vaccine was released in the United States in 1963. Following the licensing of the mumps vaccine in 1967 and the rubella vaccine in 1969, the Food and Drug Administration approved the combined MMR vaccine in 1971. Of 194 World Health Organization member states, 179 use the combined MMR vaccine. There is no connection between the MMR vaccine or any other childhood vaccine and autism.
Creating vaccine chaos and confusion is exactly the wrong response to the resurgence of measles. Once again, states, medical professional societies and health insurers, now joined by vaccine manufacturers, must push back and protect public health.
You might be interested in
Stay Updated on Public Citizen
Follow Public Citizen
Support Our Work
Product Recalls: July 15, 2026 – August 19, 2026
Health Letter, September 2026
Note: Never stop taking a drug that appears on the product recall list without first talking to your doctor or pharmacist. Most recalls are limited to a single manufacturer and may not be related to the version of a particular drug you are taking. If a recall does relate to the version of the drug you are taking, you should not stop taking the drug until you discuss an alternative treatment with your doctor or pharmacist.
For Class 1 recalls, there is a potential for serious injury or death.
For Class 2 recalls, there is a potential for serious adverse events but a lower chance of the drug causing serious injury or death than in a Class 1 recall.
CLASS I
cetirizine hydrochloride tablets, 5 mg, 100-count bottle. Manufactured by: Unique Pharmaceuticals Labs (A Div. of J.B. Chemicals & Pharmaceuticals, Ltd.). Lot #: GY825029, GY825030, GY825031 or GY825032, exp. date: 2028-OCT.
CLASS II
amlodipine and olmesartan medoxomil tablets, 10 mg/20 mg, 30 tablets per bottle. Manufactured by: Alkem Laboratories Ltd. Lot #: 25121311, exp. date: 3/31/2028.
dicyclomine hydrochloride capsules, 10 mg, 1,000 bottles. Distributed by: Aurobindo Pharma USA, Inc. Batch #: DMC125018A, exp date: 07/2027.
Diovan (valsartan) tablets, 160 mg, 90 tablets. Manufactured by: Patheon Manufacturing Services LLC. Lot #: AV5913C, exp. date: 6/2027.
divalproex sodium delayed-release tablets, 500 mg, 500-tablet bottles. Distributed by: Bionpharma Inc. Lot #: 704C003, exp. date: 12/31/2028.
levothyroxine sodium tablets; 25, 50, 75, 88, 112, 125 or 150 mcg; 10 or 100 tablets (10 x 10 or 10 x 1 unit dose blisters) per carton. Packaged and Distributed by: MAJOR PHARMACEUTICALS. Lot #: N02212, exp. date: 07/2026; Lot #: N02200, exp. date: 10/2026; Lot #: N02172, exp. date: 07/2026; Lot #: N02296 or N02288, exp. date: 12/2026; Lot #: N02310, exp. date: 01/2027; Lot #: N02203, exp. date: 09/2026; Lot #: N02266, exp. date: 11/2026; Lot #: N02167, exp. date: 07/2026; Lot #: N02212A or N02212B, exp. date: 07/2026; Lot #: N02212, exp. date: 07/2026; Lot #: N02200A or N02200B, exp. date: 10/2026; Lot #: N02200, exp. date: 10/2026; Lot #: N02172A or N02172B, exp. date: 07/2026; Lot #: N02288A or N02288B, exp. date: 12/2026; Lot #: N02172, exp. date: 07/2026; Lot #: N02288, exp. date: 12/2026; Lot #: N02266A or N02266B, exp. date: 11/2026; Lot #: N02266, exp. date: 11/2026; Lot #: N02167A, exp. date: 07/2026; Lot #: N02167, exp. date: 07/2026.
mirabegron extended-release tablets, 25 mg, 30-count bottle. Manufactured by: Zydus Lifesciences Ltd. Lot #: E408618, exp. date: 11/30/2026.
For more information about the drug recalls listed above, please visit http://www.accessdata.fda.gov/scripts/ires/index.cfm.
CONSUMER PRODUCTS
For a list of recent recalls of consumer products, please visit the Consumer Product Safety Commission website at http://www.cpsc.gov/en/Recalls/.
You might be interested in
Stay Updated on Public Citizen
Follow Public Citizen
Support Our Work
16 Organizations Call for Expanding Connecticut State CRA to Include Climate Investments
August 31, 2026
Commissioner Jorge Perez
Department of Banking
280 Trumbull Street, 16th Floor
Hartford CT 06103-1800
Dear Commissioner Perez,
Thank you for your attention to the role financial institutions play in meeting the credit needs of the communities where they operate through the examination of state chartered banks and credit unions under Connecticut’s Community Reinvestment Act (CRA). Given the growing impacts of climate change, particularly on low- and moderate-income (LMI) communities, the undersigned organizations encourage the Department of Banking to include climate mitigation and adaptation financing as eligible for positive consideration under Connecticut’s CRA.
Worsening climate disasters as well as chronic climate change impacts threaten the availability of financial services in climate vulnerable geographies. Financial institutions are increasingly withdrawing from climate vulnerable communities in an effort to manage their risk, a practice we expect will accelerate as climate change impacts grow. At a hearing before the U.S. Senate Committee on Banking, Housing, and Urban Affairs last year, former Federal Reserve Chair Jerome Powell testified “If you fast forward 10 or 15 years, there will be regions of the country where you can’t get a mortgage, there won’t be ATMs, banks won’t have branches and things like that.” While withdrawing from communities can help financial institutions manage their exposure to climate risks in the short-run, withdrawal does nothing to reduce climate risk in the financial system over the long term. Moreover, this practice deprives communities of essential financial services that they need to build climate resilience, risks violations of fair lending laws, and exacerbates economic and racial inequality.
The impact of climate change on the availability of financial services is not felt equally across communities. Due to a history of systemic exclusion, including redlining, low-income communities and communities of color are overrepresented in climate-vulnerable areas. LMI communities face greater flood risk and are more exposed to extreme heat than their middle- and upper-income counterparts. Following a climate disaster, LMI communities have fewer resources and less credit access to make needed repairs. Bluelining, or the practice of limiting credit creation and investment in climate-vulnerable areas, often leads to further financial exclusion and magnifies existing inequalities and barriers to traditional finance for LMI communities and communities of color.
Given the disproportionate impacts of climate change on LMI communities and the role of bluelining in exacerbating existing credit barriers, climate-related financing in these communities should be eligible for CRA credit. Eligible loans and investments should extend to both climate mitigation, that reduces greenhouse gas emissions, and climate adaptation, that makes property and communities more resilient to the impacts of climate change.[1] In New York and Illinois, climate-related investments in LMI communities are already eligible under the state’s CRA. As detailed in a 2021 Industry Letter published by the New York State Department of Financial Services, institutions subject to New York’s CRA may receive credit for community development lending or qualified investments related to climate mitigation and adaptation. In Illinois, climate-related product offerings and investments targeted towards LMI communities, including those related to climate resilience and mitigating environmental harm, are included as examples of community development activities eligible for credit under the state’s CRA rule. [2]
Connecticut’s state CRA directs the Department of Banking to assess community reinvestment performance “utilizing the applicable methodology set forth in federal CRA.”[3] There are numerous ways that banks can receive positive consideration for climate mitigation and adaptation financing under the federal CRA regulations in place since 1995. For example, the agencies have explicitly identified financing of “renewable energy, energy-efficient, or water conservation equipment” as CRA-eligible community development if “the benefit from the energy generated is provided to an affordable housing project or a community facility that has a community development purpose.”[4] Other examples include financing that helps meet “essential community needs” such as “a new or rehabilitated flood control measure, such as a levee or storm drain, that serves the community, including low- and moderate-income residents,” as well as financing of community development financial institutions (CDFIs) that help property owners pay for energy efficiency and clean energy upgrades.[5] Furthermore, banks would also receive positive consideration on the lending test for direct loans to LMI households or small businesses in LMI census tracts, to improve the energy efficiency of their homes or businesses.[6]
Clarifying that climate mitigation and adaptation activities are eligible for positive consideration under Connecticut’s CRA is well within the Department of Banking’s authority. It would not subject covered institutions to any new requirements; instead, it would provide further detail on the types of activities covered institutions can pursue for CRA credit. This can be communicated to covered institutions through an industry letter or guidance and does not require amending the state’s CRA rule.
Connecticut has made addressing climate change a priority. The state has set science-aligned emissions reduction milestones, including a commitment to achieving net-zero emissions by 2050 and prioritized green economic development and building climate resilience.[7] Including climate change-related financings as eligible for CRA credit will support the state’s emissions reduction and climate resilience goals and incentivize extending needed investment to the LMI communities and communities of color most impacted by climate change.
Thank you for your attention to this important issue. We welcome the opportunity for further engagement on this topic and collaboration with the Department to address climate change-related impacts to communities in Connecticut.
Sincerely,
Americans for Financial Reform Education Fund
Building Neighborhoods Together
Center for International Environmental Law
Connecticut Citizen Action Group
CT Coalition for Economic and Environmental Justice
Greater Bridgeport Community Enterprises
National Community Reinvestment Coalition
Natural Resources Defense Council
Neighborhood Housing Services of New Haven
NEST
New London Homeless Hospitality Center
North Hartford Partnership
Public Citizen
Sierra Club Connecticut Chapter
The Greenlining Institute
Third Act CT
[1] Examples of climate mitigation investments include but are not limited to investments in rooftop or community solar solar, heatpumps, appliance electrification, and energy efficiency upgrades. Examples of climate adaptation investments include but are not limited to investments in flood mitigation, home hardening, and greening infrastructure.
[2] 38 Ill. Admin. Code § 345 APPENDIX C
[3] Sec. 36a-32. (Formerly Sec. 36-52c). Assessment of banks. Factors to be considered. Community reinvestment performance evaluation.
[4] Community Reinvestment Act; Interagency Questions and Answers Regarding Community Reinvestment; Guidance. Federal Register / Vol. 81, No. 142 / July 25, 2016 / Rules and Regulations. §ll.12(h)—1. This was provided as an example of community development loans. However, the guidance also notes that examples provided of community development loans would also qualify as investments or services. See §ll.12(g)—4
[5] Ibid. §ll.12(g)(4)(iii)—4 and §ll.12(h)—1
[6] Community Reinvestment Act Regulations. Federal Register / Vol. 60, No. 86 / May 4, 1995 / Rules and Regulations. § 25.22 for OCC regulated institutions. § 228.22 for Federal Reserve regulated institutions. § 345.22 for FDIC regulated institutions.
[7] An Act Concerning the Protection of the Environment and the Development of Renewable Energy Sources and Associated Job Sectors, No. 5004, 115th Gen. Assem., 1st Sess. (Conn. 2025).
You might be interested in
Stay Updated on Public Citizen
Follow Public Citizen
Support Our Work
American Society of International Law 2026 Remarks on Trump Administration Global Health Agreements
Peter Maybarduk participated in the panel discussion “Global Governance for Health – Quo Vadis?” at the American Society of International Law’s (ASIL’s) 120th Annual Meeting (April 22-25, 2026). His delivered remarks are below.
Public Citizen is a consumer advocacy group with one million members and supporters across the United States and a 50-year history standing for the public interest before Congress, government agencies, and the courts. I direct our Access to Medicines program, which for the past 15 years has been providing technical assistance all over the world, particularly in low- and middle-income countries. We work to overcome some of the many barriers to treatment, including funding barriers. Public Citizen also has a litigation practice which is able to bring lawsuits when there is a transgression.
Given the urgent need for more perspective on what has been negotiated in the new bilateral health agreements, Public Citizen filed Freedom of Information Act requests with the U.S. State Department seeking disclosure of the memoranda of understanding and the various related agreements. When the State Department did not produce the documents according to the statutory timeline, Public Citizen sued for their production. We are not alone in pursuing these agreements. Partners in the negotiating countries around the world are also pushing for transparency.
Why is obtaining these agreements so important? Two reasons. One is that they tell us what our government is and is not doing to combat infectious disease. The other is that these agreements tell the story of what foreign aid is becoming under Trump.
Of course, these are health agreements. What they achieve or fail to achieve has serious implications for millions of people. We don’t yet know whether we’ll be able to track or ensure their progress or what the accountability measures are. They target infectious diseases inadequately, which we all have a personal interest in fighting successfully.
A year ago, the Trump administration decimated U.S. global health programs. And while many programs survive and are working to rebuild in the aftermath of the cuts, the State Department is restructuring foreign aid under its America First Global Health Strategy. Each text is a piece of that strategy made manifest. Each is important in its own regard for what is being negotiated and what the reality on the ground in different countries will be going forward. Each is also important in terms of understanding how Trump has remade the world and what the Strategy means when it is put into practice. We have to start building that record for our own understanding, and so we can influence it before it’s too late, and also to understand what the Trump administration expects or extracts in return.
Why does the State Department not release the agreements as it signs them? Our view is that the purpose of that is to divide and conquer the partners of the United States.
We are accustomed to working in cooperation with these countries, recognizing that, in health, the more we all know, the better that we all do. But the Trump administration has a very different perspective, treating its negotiating partners more like hostiles and treating health a bit like a conflict, as though U.S. negotiating advantage must be preserved through secrecy.
If one country doesn’t know what another country has negotiated, it’s more difficult to push for improved terms. The State Department’s template agreement, which leaked early on in this process, gives a general awareness of some of the things that the United States is asking for. But how far can you go and how much can you tailor an agreement to suit your country’s needs and interests? That information would be significantly enhanced through publication of the documents. We would learn if the health strategy is real or just paper, if there is accountability, if there are quid pro quo terms, and if the metrics are designed to improve health. With more transparency, countries would be better able to stand up for what they need, and we’d all be able to get better health outcomes.
The Trump administration claims its main complaint about foreign aid is that all the funding goes to a wasteful NGO sector, and that they will instead transfer aid money directly to partner governments. But even from the limited information in the press releases announcing the new agreements, we see grants for private companies who are going to make many millions of dollars off of the same processes. We may just be passing waste in new directions, with too little information made public to tell.
The Trump administration is also pushing extractive exchanges that have no connection to health. This is clearest in the case of Zambia, whose health agreement is reportedly linked with mining concessions. In our view, that’s a new and terrible way of doing health. Where else is the United States putting extractive interests over health? What is the United States asking for in exchange for restoring just a part of the desperately needed aid that was cut last year?
Around 30 agreements have been signed so far. Few of these are available for the public to read. Public Citizen tracks the agreements and annexes that have been shared publicly on its website.[1] Still, there is much more to learn. Beyond Africa, countries in Latin America and Asia have also signed agreements. The main texts don’t appear to be classified and, from the agreements that we’ve seen, include a provision allowing either party to release the text.
We know that this is a U.S. government that likes to move fast and break things. We all should be working together to make these agreements public, because it is important to influence the negotiations before they are concluded and to ensure affected communities can plan for the implementation phase.
Many transparency laws are national, and one of the interesting opportunities from today’s panel is to have conversations across borders about the transparency efforts that are underway from civil society organizations in different countries and how they connect. Lawsuits on each continent are part of the disclosure endeavor.
Activists in Kenya successfully pressed for the release of one of the first agreements, which was immensely helpful in understanding what’s being negotiated. In Zambia, the Chapter One Foundation and LCK Foundation also filed in their analogous freedom of information infrastructure and petitioned their constitutional court. In Lesotho, 21 organizations are warning about the lack of transparency. This transparency push is starting to spread across all the countries that are implicated, and we want to do everything we can to encourage it. Together, we’re going to bring health back into the light.
[1] U.S. Bilateral Health Agreements: Texts and Related Information, Public Citizen (Mar. 11, 2026), https://www.citizen.org/article/u-s-bilateral-health-agreements-case-act-reporting/
You might be interested in
Stay Updated on Public Citizen
Follow Public Citizen
Support Our Work
Response to Senator Ron Wyden’s Data Center Tax Proposal
August 31, 2026
The Honorable Ron Wyden
Ranking Member
Senate Finance Committee
Washington, DC 20002
RE: Public Citizen’s Response to Senator Ron Wyden’s Request for Information: Proposal for the Taxation of Data Centers Senate Finance Committee Democratic Staff White Paper
Dear Ranking Member Wyden,
Thank you for the opportunity to submit a response to your proposal to tax data centers. On behalf of more than one million members and supporters of Public Citizen, we submit the following comments in response to your Request for Information on the Senate Finance Committee’s Democratic Staff White Paper: Proposal for the Taxation of Data Centers. We applaud your leadership to hold the data center industry accountable and to protect the American people.
As the committee references, Big Tech is rapidly expanding a nationwide network of data centers to power artificial intelligence—fueling corporate profits while leaving the public to shoulder the costs.[1] Built at breakneck speed and with little to no government oversight or public input, this development is unfolding without necessary guardrails to protect the public interest. Further unchecked expansion threatens to drive more climate-warming emissions, harm local communities, and impact grid stability and the cost of living with estimates predicting data center electricity demand could double—or even triple—by 2028.[2]
Public Citizen believes this proposal provides an important first step to make Big Tech accountable and pay for the rampant development of data centers. And, we echo warnings of the inadequate existing federal oversight, regulation, or enforcement of AI and data centers to properly respond to community concerns, as well as concerns over the profitability of AI.[3] As you and the committee consider this proposal, we are confident that Congress can embrace holistic federal guardrails that respond to valid concerns about community consent, land use, energy prices, local pollution, and climate change. For further ideas, we hope you consider the policy suggestions in Public Citizen’s December 2025 report.[4]
Response to Senator Wyden’s Data Center Public Investment excise tax proposal:
- Removal of existing unnecessary benefits: Public Citizen supports the proposal to remove federal benefits for data centers, including making data centers ineligible for 100% bonus depreciation, prohibiting data center incentives through opportunity zones, and excluding the rents paid to REITs by new data centers from the tax-favorable REIT regime. Additionally, the committee should consider 1) calculating projected revenue increases from removing benefits of bonus depreciation with the Joint Committee on Taxation (JCT); and 2) working with the Government Accountability Office and JCT to identify other potential federal tax credits and benefits that data centers currently receive. This would allow the committee to fully consider other federal incentives that should be revoked for data centers (for example, research and development tax credits).[5]
- Implementation of a Data Center Public Investment excise tax: Public Citizen supports an excise tax levied on data centers to generate long-term funding to support communities and workers. As the committee considers the current proposal, where the excise tax would function as a gross receipts tax on data center operators, Public Citizen urges committee staff to explore additional tax levies to the gross receipts tax, such as including a federal excise tax on data center energy consumption. Public Citizen suggests coupling an excise tax on energy use with an incentive for the owners and operators to reduce residential rates. For example, the excise tax could be adjusted year-over-year if the data center owner or operator reduces residential energy rates by at least one percent. The revenue generated from the tax could, in part, fund utility rate reductions for moderate- and low-income households, finance clean energy and energy efficiency grants for families, or fund existing or new programs (i.e., a federal utility climate adaptation and resilience fund to disperse funds to states to reduce residential utility rates).As the committee works to identify the “low single-digit rate,” Public Citizen would welcome providing feedback, analysis, or other considerations for this portion of the excise tax proposal.
- Determining what is a data center: Public Citizen supports the first iterations of defining which data centers would be eligible under this proposal, including carving out exemptions for small, local data centers unless they share data with hyperscalers. However, the committee should consider a mechanism to ensure hyperscalers do not use this definition to avoid the excise tax, such as requiring the registration of small, local data centers through the Internal Revenue Service (IRS) and/or certification or registration of owners and operators of hyperscalers that may already own small, local data centers through the IRS. The committee should also consider defining what is considered a “small” or “medium” sized data center. Additionally, the committee should consider additional definitions related to energy and water use when determining the eligibility of data centers and hyperscalers for the excise tax. The definition should encompass how to address both hyperscalers (typically above 100 megawatts [MW]) and data centers that are below that threshold but would still consume large amounts of energy (20MW-100MW). The MW definition should ensure that other industrial large loads are not adversely impacted. Public Citizen defines a relevant data center as 20 MW or greater—consistent with the Federal Energy Regulatory Committee’s (FERC) definition of a large load.[6] We also recommend Congress or FERC set a maximum threshold limit for data centers and hyperscalers. As the committee moves forward with the excise tax proposal, considerations around energy use of data centers should be done in coordination with relevant congressional committees and the Energy Information Administration (EIA), which is currently in the process of gathering energy data from data centers.[7] The EIA should be required and/or encouraged to work with the IRS to ensure accurate data and reporting of energy use, should energy use considerations be included as part of the revenue for the excise tax.
- Addressing pre-existing data centers: In the current excise tax proposal, which focuses on data centers that were placed in service towards the end of 2023, the committee should consider additional mechanisms to close potential loopholes that may allow data centers to use a pre-construction date to avoid the excise tax. For example, a company shouldn’t be allowed to expand an existing facility to avoid the “placed in service” date.
- How to deal with the many business models of data center infrastructure: The current proposal to limit the scope of the tax on the “data center operator,” or the legal owner of the data center and/or the taxpayer operating the data center, is a useful start. Public Citizen would support strong guidelines for ensuring the taxpayer selling the services of the data center or operating the data center be carefully defined to ensure that Big Tech firms do not avoid the excise tax or other federal regulations. For example, the committee should consider how third party business vendors, or private equity businesses, who would qualify as the “owner” but not the sole “operator” of the data center should be taxed. While Big Tech should shoulder most of the tax liability of owning and operating data centers, private equity investors[8] and third party business vendors supporting the investments, ownership, and operations of data centers should also be considered in the excise tax proposal.
- Data centers in space: Public Citizen has serious concerns about proposals to put data centers in space. With that, we support the committee’s work to tax data centers in space that service U.S. customers and companies that try to avoid being treated as a non-U.S. asset. The committee should additionally require U.S. owners and operators to register any data center assets in space to prevent further tax avoidance by claiming that they are not servicing U.S. customers.
- Use of revenue: In addition to the above uses of the revenue from the excise tax should energy use considerations be included (i.e., support for energy bills and clean energy programs), the revenue should also be used to support communities negatively impacted by data centers and workers who lost their jobs due to AI. Other valuable uses of the revenue include supporting AI safety and literacy programs, apprenticeship and job transition programs, or grants for community-created programs in areas adversely impacted by AI and data centers. Public Citizen welcomes further discussion on additional ideas on the use of revenue for the excise tax with the committee.
Additionally, the committee should consider proposing, either as part of this proposal or a separate effort, limiting data center development on federal lands.
Thank you for considering these ideas and policy considerations while moving forward on the taxation proposal for data centers. If you would like to meet and discuss these ideas further or have questions, please reach out to Meghan Pazik, senior policy advocate with Public Citizen’s Climate Program ([email protected]).
Sincerely,
Public Citizen’s Climate and Energy Program
[1] Weise, Karen, Microsoft Backs Up A.I. Spending With $27.2 Billion Quarterly Profit. New York Times (30 June 2025), https://www.nytimes.com/2025/07/30/technology/microsoft-earnings-ai-data-centers.html
[2] Shehabi, Arman et al., 2024 United States Data Center Energy Usage Report, Lawrence Berkeley Lab (Dec. 2024), https://escholarship.org/uc/item/32d6m0d1
[3] Marshall, Christa, AI boom sparks rare warning of ‘significant risks’ to grid, E&E News, (4 May 2026), https://www.eenews.net/articles/ai-boom-sparks-rare-warning-of-significant-risks-to-grid/ ; Leask, Hugh, Bill Gates warns ‘there is no plan’ for the ‘upheaval’ AI will cause, CNBC (26 August 2026), https://www.cnbc.com/2026/08/26/bill-gates-ai-jobs-economic-upheaval.html
[4] Noel, Deanna and Pazik, Meghan, Reining in Big Tech: Policy Solutions to Address the Data Center Buildout, Public Citizen (3 December 2025), https://www.citizen.org/article/reining-in-big-tech-policy-solutions-to-address-the-data-center-buildout/
[5] Mannion, Tom and Schram, Tim, 4 Key OBBBA Provisions Data Centers Need to Know, BDO USA, (30 March 2026), https://www.bdo.com/insights/industries/technology/4-key-obbba-provisions-data-centers-need-to-know
[6] Federal Energy Regulatory Commission, Interconnection of Large Loads to the Interstate Transmission System (Docket No. RM26-4-000), FERC (23 October 2025), https://www.ferc.gov/rm26-4
[7] Energy Information Administration, EIA launches pilot survey on energy use at data centers, EIA (25 March 2026), https://www.eia.gov/pressroom/releases/press585.php
[8] Baker, Jim, Private equity cash fuels data center buildout, Private Equity Stakeholder Project (26 May 2026), https://pestakeholder.org/news/private-equity-cash-fuels-data-center-buildout/ ; Goldman Sachs, Private Markets Are Expected to Have a Growing Role in Data Center Financing, Goldman Sachs (12 June 2026), https://www.goldmansachs.com/insights/articles/private-markets-expected-to-have-growing-role-in-data-center-financing
You might be interested in
Stay Updated on Public Citizen
Follow Public Citizen
Support Our Work
Trump’s Bilateral Agreements Reshape Global Health Funding
By Public Citizen, Partners In Health
In 2025, the Trump administration froze foreign assistance funding, dismantled USAID, and terminated 80 percent of global health funding awards, causing broad disruption to health programs, including those that support HIV/AIDS, tuberculosis, malaria, and maternal and child health.[1] In the aftermath, new bilateral agreements between the United States and partner country governments provide the main framework for moving global health funds under the State Department’s “America First Global Health Strategy,” restructuring and supposedly stabilizing health programs that have been supported by USAID and PEPFAR for decades.
The Trump administration, however, has kept the public in the dark about the full details of its plans for the future of U.S. support for global health. The State Department has negotiated these bilateral agreements (memoranda of understanding, or MOUs) on rushed timelines with limited participation from civil society and affected communities, and has refused to release all of the agreements and related planning documents to the public.
Public Citizen filed a lawsuit against the U.S. State Department for failing to produce records responsive to Freedom of Information Act (FOIA) requests seeking 16 MOUs. Through the suit, Public Citizen obtained 15 MOUs, six of which had not previously been made public.[2] One document, which appears to be Uganda’s MOU, was withheld in its entirety pursuant to FOIA exemptions related to foreign government information and the foreign relations of the United States. Uganda’s MOU was among five previously posted by the State Department.[3]
Below, we analyze the 17 MOUs now publicly available,[4] as well as funding figures from an 18th agreement accessed independently.[5] These MOUs paint a fuller picture of U.S. health assistance and help reveal how the administration plans to spend congressionally appropriated funds for global health.
The analysis finds that:
- Compared to pre-2025 spending, the U.S. government plans a 59% reduction — amounting to about $2 billion — in funding under the MOUs to these 18 countries by 2030, despite Congress ordering global health spending to remain relatively steady;
- Across the 18 countries, U.S. government funding is reduced by between 43% and 97% compared to baseline levels by the final year of the MOUs. Those with the steepest reductions in their MOUs by 2030 are Rwanda (97% reduction in U.S. funding compared to pre-2025), Liberia (84%), Burundi (78%), Madagascar (77%), and Sierra Leone (71%);
- Countries that miss steep co-financing requirements may face penalties which could worsen funding cuts, with some countries set to lose much more than others. At the same time, “performance incentives,” which draw from an unspecified total pool of funding, could redirect tens of millions of dollars to certain countries, creating an avenue to reward some countries while deepening funding disparities; and
- In addition to funding, co-financing penalties, and performance incentives, agreements vary across several other domains, including plans for sustaining health workforce capacity.
The U.S. government plans a 59% reduction in funding under the MOUs by 2030, compared to historical spending
In 2026, in a rebuke to the Trump administration’s proposed $6.2 billion cut to fiscal year 2026 global health funding,[6] Congress enacted a near-stable budget.[7] Yet based on the MOUs, the administration’s plans to actually spend newly appropriated money won’t come close to keeping up with historical levels. Instead, the MOUs warn of a disturbing potential underspend of congressionally appropriated global health dollars.
Compared to historical spending, the U.S. government plans a 59% reduction (amounting to about $2 billion) in funding under the MOUs to these 18 countries by 2030 (figure 1).
*Baseline funding is the sum total of the most recent FY25 State/USAID PEPFAR and USAID-GHP request or FY23 actuals across countries and includes bilateral HIV, TB, malaria, nutrition, MCH, GHS, FP/RH; FY25 rescissions cut down FP/RH by $500 million but the FY26 conferenced SFOPS/NSRP bill & report restores that funding; see KFF Tracker for similar analysis.
The steep decline in funding compared to baseline levels over the course of the MOUs raises concerns about how, or if, the administration intends to spend appropriated money. While spending outside the MOUs (for example, through the State Department’s Annual Program Statement [APS] addenda, which provide supplemental funding to support specific projects)[8] could fill in the funding gap shown in the MOUs, the administration has shown its willingness to delay funding.[9] If the MOUs are any indication, in the coming years, new money intended to fight disease and support health might instead sit unused or be spent in unpredictable ways.
Rapid funding transition makes country-level planning difficult
The U.S. government plans to transition countries away from U.S. funding over just three to five years, even as countries are still dealing with the impact from the Trump administration’s cuts.
Analysis of the country-level changes in U.S. contributions over the course of the MOUs compared to pre-2025 baseline levels shows rapid decreases, and in some cases near-complete planned cessation of U.S. funding by 2030 (figures 2 and 3).
*Figure 3 relies on the same data as figure 2, but shows 2024 as 100% of the baseline and charts variation relative to that point over the five year period. Baseline funding is the sum total of the most recent FY25 State/USAID PEPFAR and USAID-GHP request or FY23 actuals across countries and includes bilateral HIV, TB, malaria, nutrition, MCH, GHS, FP/RH; FY25 rescissions cut down FP/RH by $500 million but the FY26 conferenced SFOPS/NSRP bill & report restores that funding; see KFF Tracker for similar analysis.
Across all 18 countries, U.S. government funding is reduced by between 43% and 97% compared to baseline levels by the final year of the MOUs.[10] Those with the steepest reductions in their MOUs by 2030 are Rwanda (97% reduction in U.S. funding compared to 2024 baseline funding levels), Liberia (84%), Burundi (78%), Madagascar (77%), and Sierra Leone (71%).
Additionally, while co-financing commitments (the amount of funding the co-signatory government is expected to contribute for each year of the agreement) could help with program stability, the planned decreases in U.S. funding are larger for several countries than the funding increases from partner governments. For more than half of countries (11 out of 18), co-financing commitments fall short of the U.S. government’s cuts over five years compared to baseline U.S. government funding. Countries may also struggle to meet these funding expectations. For example, Malawi would have to mobilize new funding equal to 56% of the country’s total health expenditure to meet its annualized co-financing commitment.[11] Countries also face co-financing requirements from the Global Fund and Gavi, representing a substantial additional burden; for instance, in Uganda, annualized co-financing obligations from Global Fund, Gavi and the U.S. MOU on average equal 44% of the total government health budget for 2026.[12]
Funding decreases by the U.S. over such a short period, uncertainty around potential future contributions from the U.S. government outside of the MOUs, and steep co-financing requirements from multiple donors make it harder for countries to sustain health programming and plan for the future.
Harsh penalties and vague incentives make funding more unpredictable
At the same time as the Trump administration plans to decrease its funding, aggressive co-financing commitments and potential penalties increase the risk of funding shocks and could mean even steeper declines in support compared to previous years.
If countries do not meet co-financing expectations, the U.S. government reserves the right to reduce or cease funding. The MOUs also state that the co-investment may not include funding from other donors or multilateral organizations, but must be funds “raised directly” by the country.
Consequences for failing to meet co-financing commitments range from reducing U.S. funding by $1 for every $1 (or equivalent) of the shortfall from the co-signatory country’s expected funding contributions,[13] a 2:1 reduction (in which the U.S. would reduce funding by twice as much as the shortfall),[14] another linear amount,[15] or no specified penalty amount.[16] There is no transparent logic guiding the application of penalties across countries.
Although some countries softened language surrounding U.S. government funding reductions in some cases (such as by including that funding reductions should take place after good faith discussions), the inclusion of the penalties at all raises questions about the predictability of funding under the Trump administration’s strategy.
Many of the MOUs also include “performance incentives” for meeting process and outbreak response metrics.[17][18]
The performance incentive included for some countries refers to an unknown funding pool, which will be used to calculate the size of payments to eligible countries. The MOUs indicate in most cases that the performance incentive will be available in 2027 or 2028 and corresponds to a distribution proportional to the size of a country’s population relative to the population of all eligible countries.
In most cases, the MOUs also provide a cap for each co-signatory country’s potential performance incentive based on a pre-set per person dollar amount ($1,[19] $2,[20] $5,[21] or $10[22] per person per year) or an absolute maximum of $75 million.[23] Two agreements do not note an incentive cap.[24] Rwanda’s MOU is unique among the 17 agreements analyzed. It does not include specific eligibility years, indicates that performance incentives are connected to the achievement of a smaller set of metrics (HIV and malaria metrics), and states that funds may be used in the Government of Rwanda’s strategic investment in the health sector, including Kigali Health City and AI application for healthcare (instead of generally stating that funds may be applied to any health-related costs under the MOU). The two strategic investment areas mentioned are also those where U.S. commercial partnerships and investment interest are indicated.[25]
While the size of the potential pool of money is not stated, the performance incentives suggest potentially tens of millions of dollars allocated for this purpose.
Agreements vary across several other domains
In addition to funding, co-financing penalties, and performance incentives, agreements vary across several other domains.[26]Additional observations from the country agreements include:
- In the available MOUs, the U.S. reduces support for the health workforce, leaving seven countries in the African region with fewer health workers by 2030, taking into account both US and partner government commitments. These are Burundi, Cameroon, Eswatini, Kenya, Lesotho, Madagascar and Malawi.[27] This is a major concern since the African region currently has only 46% of the health workers it needs, and by 2030 it is projected to face a health workforce shortage of 5.85 million workers.[28] In a number of countries, the reduced number of funded positions for community health workers (CHWs) is quite substantial, as for instance in Malawi, where 3,436 positions are eliminated (72%), and Madagascar, where 8,200 positions are eliminated (39%). In Burundi all U.S. funded CHW positions are terminated, with no plan for the government to absorb them. Cameroon has a severe shortage of CHWs, as well as nurses, yet funded positions for both cadres are reduced by 20%.[29]
- The majority[30] of available MOUs reduce the duration of controversial data sharing agreements and specimen sharing agreements (providing the U.S. government access to a specified set of data systems related to the functioning of the MOU and to viral specimens and related data collected as part of public health surveillance, respectively), though they also retain an acknowledgement in some form that failure to fulfill sharing agreement terms could result in reductions in planned funding. These agreements are contentious because they may act as a condition of ongoing funding and the U.S. government originally sought expansive access to data for up to 25 years (well beyond the three to five year period of funding under the MOUs). These agreements stand apart from the MOU itself, and most are not available in wider circulation, making a full assessment of final terms impossible in most cases. In some cases, the data and specimen sharing terms have contributed to stalled negotiations, endangering health funding.[31]
- Most MOUs retained in some form a provision from the State Department’s template related to the acceptance of U.S. FDA approvals for medical countermeasures (g., vaccines and therapeutics) used in outbreak response;[32] far fewer MOUs included the broader separate commitment to modify national regulations to allow for recognition of U.S. FDA approvals for all medical products.[33] Accelerated approval during outbreaks can support swift deployment of medical countermeasures newly identified for the health emergency. A requirement of regulatory recognition for all U.S. FDA-approved products effectively ensures that U.S. companies have market access, which may be the intention of the broader provision, given that the State Department’s guidance document that accompanied the template MOU clarified that this provision should be included for countries with “large domestic markets” or “for which there are other strategic reasons.” The U.S. government’s recent attempts to compel countries to automatically accept U.S. regulatory approvals, as shown in the MOU template and in similar provisions in recent trade agreements,[34] raise concerns about U.S. pressure on countries’ regulatory autonomy and the inclusion of such terms in agreements for renewed health funding.
- The new HIV pre-exposure prophylaxis (PrEP) drug lenacapavir is referenced in eight of the MOUs available to date.[35] Two agreements include specific funding amounts for the purchase of lenacapavir in 2026.[36] The inclusion of lenacapavir in these agreements aligns with several partner governments’ goals to roll out the drug, as well as the Trump administration’s aim to promote products made by American companies and its commitment, in partnership with the Global Fund, to reach three million people with lenacapavir over the next three years.[37] However, current supply commitments fall well below the potential need for HIV prevention medication, which is estimated at as many as 20 million people.[38] At the same time, plans for the rollout of lenacapavir are occurring against the backdrop of the U.S. government’s health cuts. Among facilities receiving PEPFAR support between 2024 and 2025, the number of people starting PrEP decreased by 33%, the number of healthcare workers delivering HIV services declined by 24%, and HIV testing declined by 18%.[39] Robust service delivery, outreach, and testing programs are necessary for the successful implementation of PrEP interventions.
The MOUs include rapid decreases in planned U.S. government global health support over the next five years and agreement terms that are not primarily guided by health, including conditionalities that may worsen funding cuts. Congress has ordered new global health funding. Failing to spend these funds based on health needs will only lead to preventable harm.
[1] Jennifer Kates, Adam Wexler, Anna Rouw, & Stephanie Oum, KFF (Apr. 17, 2025), https://www.kff.org/global-health-policy/analysis-of-usaids-active-and-terminated-awards-list-how-many-are-global-health/
[2] Botswana, Burundi, Cameroon, Côte d’Ivoire, Eswatini, Ethiopia, Kenya, Lesotho, Liberia, Madagascar, Malawi, Mozambique, Nigeria, Rwanda, Sierra Leone (MOUs not previously made public appear in bold).
[3] Public Citizen has made all public texts available here.
[4] The 17th agreement is South Sudan’s draft text. South Sudan’s agreement was not included in the FOIA suit but has been posted online by others.
[5] The Democratic Republic of the Congo.
[6] White House Releases FY26 Budget Request, KFF (May 2, 2025), https://www.kff.org/global-health-policy/white-house-releases-fy26-budget-request/
[7] FY 2026 National Security, Department of State and Related Programs (NSRP) Global Health Funding in the Consolidated Appropriations Act, KFF (Feb. 4, 2026), https://www.kff.org/global-health-policy/global-health-funding-in-the-fy-2026-national-security-department-of-state-and-related-programs-nsrp-conference-bill-explanatory-statement/; U.S. Global Health Budget Tracker, KFF, https://www.kff.org/interactive/u-s-global-health-budget-tracker/
[8] Bureau of Global Health Security and Diplomacy, Advancing Global Health, Simpler.grants.gov, https://simpler.grants.gov/opportunity/ff23ddd9-0a2e-470b-8881-7d7dc74aed03
[9] This is particularly concerning for HIV funds that may be spent within a period of up to five years, potentially masking the extent to which appropriated funds are spent in a timely manner. See K. J. Seung & Vincent Lin, U.S. Global Health Spending Watch (June 15, 2026), https://globalhealthwatch.org/?blog=between-two-systems (showing slower obligations for FY2025 five-year funding than in previous years, when much of the spending occurred in the second year after money was appropriated. This warns that money is not moving at scale, even though it is not under immediate threat of expiry).
[10] Botswana and South Sudan have three-year MOUs. All other countries have five-year MOUs, running through 2030.
[11] Based on 2023 World Bank health finance data.
[12] Clinton Health Access Initiative (CHAI), HIV Services Are Still Falling: A Year into the Funding Crisis, New Data Shows Little Recovery (June 8, 2026), https://www.clintonhealthaccess.org/wp-content/uploads/2026/06/HIV-Market-Impact-Memo-June2026.pdf; Kalipso Chalkidou, Financing Health: Where is the money coming from?, World Health Organization (June 2026).
[13] Ethiopia, Kenya, Mozambique, Cameroon, and Malawi.
[14] Uganda and Côte d’Ivoire.
[15] Nigeria, according to section 5.1.1.
[16] Rwanda, Liberia, Lesotho, Eswatini, Sierra Leone, Madagascar, Burundi, Botswana, and South Sudan.
[17] Agreements for Liberia, Botswana, Burundi, Madagascar, Sierra Leone, and South Sudan do not include references to performance incentives.
[18] Many MOUs also include potential penalties for failing to meet metrics. This is often included as a general statement that the U.S. government plans to reduce funding if countries fail to maintain outcome and process metric baselines or achieve outbreak response targets.
[19] Uganda ($45,900,000 cap, based on population size stated in Uganda’s MOU).
[20] Mozambique ($69,263,532 cap), Cameroon ($61,281,634), Côte d’Ivoire ($62,331,308), Malawi ($45,600,000). Mozambique’s MOU does not include its population size; value for cap is based on 2024 World Bank population data.
[21] Lesotho ($10,582,135 cap).
[22] Eswatini ($12,605,420 cap).
[23] Ethiopia and Nigeria.
[24] Kenya and Rwanda.
[25] Article 2.6.4, Future Strategic Opportunities in the U.S.-Rwanda Partnership Beyond the MOU, Multispecialty Hospital and AI Application for Healthcare.
[26] Several domains are documented in an earlier Public Citizen report that observed diverging terms across agreements when compared with the template MOU and plotted them in a negotiation tracker. The tracker has been updated with the new agreements.
[27] Eswatini’s agreement shows fewer lab workers, doctors, nurses, CSWs and CHWs by 2030; while it states CHW tasks will be shifted to nurses and that other CHWs will be absorbed by NGOs, nurses have a distinct scope of work and no NGO is a party to the agreement. Kenya’s agreement shows the number of frontline healthcare workers declining by 53% in 2028 when U.S. support for this workforce ends; the agreement includes a commitment by Kenya to funding continuity for many cadres, but there is no commitment to absorb the large number of U.S. funded CHWs. MOUs signed by Botswana and Rwanda do not include specific numerical commitments per cadre.
[28] State of the health workforce in Africa 2026: plan, train and retain. Brazzaville: WHO Regional Office for Africa (2026), https://www.afro.who.int/publications/state-health-workforce-africa-2026-plan-train-and-retain
[29] It is worth noting that, despite significant reductions in U.S. funding, some governments commit to increasing the number of health workers by 2030, namely Côte d’Ivoire, Liberia, Mozambique, Sierra Leone and Uganda, based on the available MOUs.
[30] Of the 17 available MOUs, only the Madagascar and Rwanda agreements retain the 25-year data sharing agreement duration from the template MOU (though without access to the data sharing agreements themselves, it is difficult to know if this duration may be altered in the final text); other MOUs reduced the duration to between five and ten years, sometimes with the possibility of renewal. All MOUs that include information on the duration of the specimen sharing agreement reduce it to between five and ten years, sometimes with the possibility of renewal. For some countries, the agreement durations cannot be discerned from the MOU text. Additionally, some MOUs reference a data sharing agreement but remove references to a specimen sharing agreement (Sierra Leone, Botswana, Liberia). South Sudan’s draft agreement does not reference either a specimen sharing or data sharing agreement.
[31] For example, in Zimbabwe and Ghana negotiations reportedly stalled over the terms. In Zambia, negotiations also stalled; Zambia’s foreign affairs minister said negotiations broke down due to “unacceptable” data sharing terms “in violation of our citizens’ right to privacy” and “the insistence on preferential treatment of U.S. companies over Zambia’s critical minerals.”
[32] Six MOUs incorporated the template provision (Madagascar, Rwanda, Mozambique, Liberia, Kenya, Ethiopia), six modified the provision (Sierra Leone, Malawi, Botswana, Côte d’Ivoire, Cameroon, Nigeria), and five removed the provision (Uganda, Lesotho, Eswatini, Burundi, South Sudan).
[33] Two MOUs accepted the provision as proposed (Rwanda, Liberia), Six modified the provision (Sierra Leone, Botswana, Cameroon, Uganda, Nigeria, Kenya), and nine removed the provision (Ethiopia, Mozambique, Lesotho, Côte d’Ivoire, Eswatini, Burundi, Madagascar, Malawi, South Sudan).
[34] For example, annex III, article 2.4(5), U.S.-Malaysia Agreement on Reciprocal Trade (Oct. 26, 2025), https://www.whitehouse.gov/wp-content/uploads/2025/10/MALAYSIA-ANNEX_APPENDIX.pdf
[35] Sierra Leone, Botswana, Eswatini, Cameroon, Ethiopia, Rwanda, Lesotho, Kenya.
[36] Botswana’s agreement indicates that commodities funding under the agreement will shift to Botswana over the course of the agreement, apart from lenacapavir, for which “the U.S. Government plans to continue procurement over the duration of this MOU.” In 2026, the agreement notes that $550,000 will purchase lenacapavir and $300,000 will support its rollout. Sierra Leone’s agreement indicates $200,000 from the U.S. government for lenacapavir in 2026. A third agreement (Kenya’s) includes $20,000,000 for antiretroviral therapy and lenacapavir. Eswatini’s agreement does not include a specific funding amount, though it does state that the country will target women ages 15–34.
[37] AVAC, Source of Lenacapavir for PrEP Supply to Early Adopter Countries, https://avac.org/resource/infographic/lenacapavir-supply/(showing current PEPFAR and Global Fund supply commitments by country; there are plans to add additional supplies in the coming months).
[38] UNAIDS, Global HIV response falters as reemergence looms (July 27, 2026), https://www.unaids.org/en/resources/presscentre/pressreleaseandstatementarchive/2026/july/20260727_PR_UNAIDS_special_report_AIDS2026
[39] Brian Honermann, Anna Grimsrud, Elise Lankiewicz, Jennifer Sherwood & Greg Millett, The Impact of the United States Foreign Aid Freeze on HIV Service Delivery in PEPFAR-Supported Countries: A Facility-Level Analysis of 2024–2025 Programme Data, Journal of the International AIDS Society (2026), https://onlinelibrary.wiley.com/doi/full/10.1002/jia2.70182
You might be interested in
Stay Updated on Public Citizen
Follow Public Citizen
Support Our Work
U.S. Bilateral Health Agreements: Texts and Related Information
Following the dismantling of USAID and large-scale disruption to U.S. foreign assistance programs, the Trump administration is restructuring health aid according to its “America First Global Health Strategy.” Under the strategy, the Department of State is negotiating multi-year bilateral health funding agreements (memoranda of understanding, or, MOUs) with dozens of countries. The administration is negotiating these agreements on tight timelines, with limited transparency and minimal civil society engagement, despite their significant impact on how health systems and funding operate for years to come.
The State Department posted the main texts of several signed agreements, as well as some of the related “specimen sharing agreements” (but not the similarly related “data sharing agreements”) to its website.
PDFs of the agreements published by the Department of State are available in the table below. The table additionally includes agreements not published by the U.S. but posted by other sources.
An asterisk (*) indicates that the U.S. government made the text or other information public via its Case Act reporting page.
A caret (^) indicates that the U.S. government made the text public pursuant to Public Citizen’s Freedom of Information Act (FOIA) requests.
| Country | Agreement Date | Main Text | Data Sharing Agreement | Specimen Sharing Agreement |
|---|---|---|---|---|
| Kenya | Dec. 4, 2025 | PDF*^ | PDF (Health Policy Watch) | |
| Rwanda | Dec. 5, 2025 | PDF^ | PDF* | |
| Liberia | Dec. 9, 2025 | PDF^ | ||
| Lesotho | Dec. 10, 2025 | PDF^ | PDF* | |
| Uganda | Dec. 10, 2025 | PDF* | Reported to Congress (text not disclosed)* | |
| Eswatini | Dec. 12, 2025 | PDF^ | PDF* | |
| Mozambique | Dec. 15, 2025 | PDF*^ | PDF* | |
| Cameroon | Dec. 16, 2025 | PDF^ | PDF* | |
| Nigeria | Dec. 20, 2025 | PDF*^ | PDF* | |
| Sierra Leone | Dec. 22, 2025 | PDF^ | ||
| Botswana | Dec. 23, 2025 | PDF^ | ||
| Ethiopia | Dec. 23, 2025 | PDF*^ | Reported to Congress (text not disclosed)* | |
| Madagascar | Dec. 22, 2025 or Dec. 23, 2025 | PDF^ | ||
| Côte d'Ivoire | Dec. 30, 2025 | PDF^ | PDF* | |
| Malawi | Jan. 14, 2026 | PDF*^ | PDF* | |
| Burundi | Feb. 6, 2026 | PDF^ | ||
| Burkina Faso | Feb. 25, 2026 | |||
| Panama | Feb. 25, 2026 | |||
| Democratic Republic of the Congo | Feb. 26, 2026 | PDF* | ||
| Dominican Republic | Feb. 26, 2026 | |||
| Niger | Feb. 26, 2026 | |||
| El Salvador | Feb. 27, 2026 | |||
| Guatemala | Feb. 27, 2026 | |||
| Guinea | Feb. 27, 2026 | |||
| Honduras | Mar. 13, 2026 | |||
| Senegal | Mar. 13, 2026 | |||
| Angola | Mar. 19, 2026 | |||
| Cambodia | April 2, 2026 | |||
| Tajikistan | April 6, 2026 | |||
| Philippines | April 7, 2026 | |||
| Bolivia | April 17, 2026 | |||
| Papua New Guinea | April 22, 2026 | |||
| South Sudan | June 25, 2026 | PDF (draft) (Think Global Health) | ||
| Tanzania | July 1, 2026 | |||
| Vietnam | Sept. 11, 2026 | |||
| Zambia | Oct. 8, 2026 | PDF (Health GAP) |
Source of MOU texts posted via the Case Act reporting page: here. Source of specimen sharing agreements and related information posted via Case Act reporting pages: here and here.
Source of country agreement dates: KFF or State Department press releases.
Public Citizen filed a lawsuit against the U.S. State Department for failing to produce records responsive to Freedom of Information Act requests for the health agreements and related information. Read the press release related to the suit here. The Washington Post covered the suit in its reporting on the Trump Administration’s secretive approach to global health assistance.
Additional documents related to the MOUs and the implementation of the America First Global Health Strategy (AFGHS)
Template agreements:
- Template MOU (retrieved from PEPFAR Watch)
- Template MOU guidance (retrieved from PEPFAR Watch)
- Model Specimen Sharing Agreement (retrieved from Health Policy Watch)
- Model Data Sharing Agreement (retrieved from Emily Bass, Substack)
Draft agreements:
- Zambia draft MOU (retrieved from Zambia’s Draft Memorandum of Understanding With the U.S. Government: What Do We Know?, Health GAP, March 16, 2026)
- Zimbabwe draft MOU (retrieved from From Standoff to Shutdown: Trump’s Diplomatic Failures Would Trigger Collapse of Healthcare Systems and Services in Zimbabwe, Health GAP, Sept. 29, 2026)
Implementation planning:
- MOU Implementation Plan Template (retrieved from Emily Bass, Substack, Jan 2026)
- MOU Implementation Plan Companion Tool (retrieved from Emily Bass, Substack, Jan 2026)
- MOU Implementation Process Overview (retrieved from Emily Bass, Substack, Jan 2026)
- AFGHS implementation documents (retrieved from Emily Bass, Substack, May 2026; also discussed here).
- Section 1 of Uganda Implementation Plan Dated April 2026 (retrieved from Emily Bass, Substack, June 2026)
- Guidance for FY2027 AFGHS Bilateral Operational Planning (retrieved from Emily Bass, Substack, Sept. 16, 2026)