The average price of diesel fuel in the US shot up from about $3.69 earlier this year to a high of $6.53 per gallon on Wednesday, September 23. Beyond the sticker shock at the pump, higher-priced diesel means the cost of transporting just about anything by heavy-duty truck, rail, or marine transport will rise, deepening the affordability crunch affecting American households. 

Politico reported that the Trump administration was planning to ban diesel exports for 90 days, while a White House official denounced the report as false. Although we understand the urge to address diesel prices, banning exports will do little to reduce prices in the short run and could backfire spectacularly in the long run. We support the view expressed by Secretary of Energy Chris Wright, who told a New York audience on Wednesday that banning diesel exports is a blunt tool that won’t work.

Read the rest of this post →

At the end of August 1976, in the lingering patriotism of America’s bicentennial celebrations, Congress passed the National Emergencies Act, a bill meant to rein in past delegations of emergency power and ensure that the legislature maintains control over the use of such authorities going forward. President Gerald Ford, who had turned back Ronald Reagan’s challenge at the Republican National Convention in Kansas City earlier that month, signed the NEA into law on September 14.

The law was the culmination of a years-long process initiated to study the implications of terminating a 1950 national emergency proclamation, issued by President Truman during the Korean War, that was still in force and in use throughout the Vietnam War. The special committee soon realized that there were, in fact, four extant national emergencies, the oldest dating back to 1933. Dismayed that a majority of Americans had “lived their entire lives under emergency rule,” the committee was clear-eyed about where blame belonged: “This dangerous state of affairs is a direct result of Congress’ failure to establish effective means for the handling of emergencies and its willingness to defer to Executive branch leadership.” 

Read the rest of this post →

Last week, Williams—a natural gas company—hosted a ribbon-cutting ceremony for a power plant called Socrates South that just began operating in New Albany, Ohio. Notably, this natural gas plant is fully disconnected from the public power grid, supplying the data centers of a Meta affiliate, Sidecat, without affecting anyone else’s electricity. 

Speed to Power

The New Albany site is part of a growing trend of data centers acquiring electric service more rapidly than traditional timelines of six, eight, or even ten years. The Socrates South plant came online in just 17 months, with the first filings made with the Ohio Power Siting Board in March of 2025. 

This is a big victory for the local community, the developer Williams, and their customer Sidecat, but also the Ohio lawmakers behind the legislation that made it possible: last year’s HB 15. That bill provided the regulatory certainty that a fully off-grid generation plant could avoid the long queue for grid connection and bring “speed to power” to customers who don’t want to wait. It’s aligned with a framework we’ve been advocating, called consumer-regulated electricity (CRE), which creates space for new, privately financed electric utilities to supply non-residential customers, such as data centers, without interconnecting with the existing grid.

Read the rest of this post →

This week, Reuters and Politico reported that the White House plans to expand the Ratepayer Protection Pledge by encouraging utilities, data center developers, and Republican governors to join the effort. The pledge rests on a simple principle: The technology companies driving the AI boom should pay the full cost of the electricity infrastructure they require, rather than shifting those costs onto American families.

That’s the right objective. The next step is finding a policy that actually achieves that.

Public concern over data centers has grown rapidly, driven less by opposition to artificial intelligence than by fears of higher electricity bills. New York Gov. Kathy Hochul recently responded by imposing a statewide moratorium on new data center construction. Other states are considering similar restrictions. Whether those concerns are justified or not, the political reality is clear: If policymakers cannot protect consumers from rising costs, they will eventually protect them by slowing development instead.

America needs a better option. The Trump administration should champion reforms that accomplish both goals at once: accelerate AI infrastructure while ensuring American families never have to subsidize development. Consumer-Regulated Electricity (CRE) offers exactly that solution by allowing developers to build and operate entirely private electricity networks outside the traditional regulated grid.

President Trump’s 2025 financial disclosure landed last week showing more than $2 billion in income, roughly triple 2024’s figure. Cryptocurrency alone accounted for $1.4 billion, most from meme-coin royalties and token sales tied to the Trump family’s World Liberty Financial. Read against the deals cataloged below, the filing paints a sordid picture of self-dealing within the upper echelons of the White House—or at the very least, the appearance of such.

Tungsten and Kazakhstan

Let’s start with tungsten. Today, China controls roughly 80 percent of the world’s supply and has spent the past year and a half tightening export controls. This has caused the metal’s cost to grow from around $340 per ton to well past $3,000. The US hasn’t run a commercial tungsten mine since 2015. Given tungsten’s importance to a number of security-related products, including weapons systems, American policymakers are actively looking to break China’s chokehold.

Over the summer of 2025, the US Export–Import Bank and US International Development Finance Corporation issued letters of interest—non-binding but meaningful signals—for up to $1.6 billion in potential financing for a Kazakh tungsten venture led by Cove Kaz Capital: $900 million from EXIM and $700 million from DFC, against total development costs pegged at roughly $1.1 billion. The joint venture itself, pairing Cove Kaz with Kazakhstan’s national mining company Tau-Ken Samruk, was unveiled in November 2025 at the C5+1 summit by President Trump and Kazakh President Kassym-Jomart Tokayev. President Trump personally worked out the deal with Kazakhstan President Kassym-Jomart Tokayev; Commerce Secretary Lutnick was also heavily involved. The developer has since gone back to the well, asking the Pentagon for another $400 million.

The president’s two eldest sons found their way into the deal through a side door. According to Financial Times reporting, Donald Trump Jr. and Eric Trump entered the picture in August 2025 through American Ventures, an investment vehicle holding a stake in Skyline Builders, a Nasdaq-listed firm, and they added to those positions on October 28, 2025. On October 31—weeks after Lutnick met Tokayev in New York with President Trump joining by phone—investors participating through a financing structure involving Dominari Securities, a Trump Tower-based firm partly owned by the brothers, acquired a 20 percent stake in the entity tied to the tungsten project. The US-Kazakhstan agreement was signed six days later. Skyline and Cove Kaz have since agreed to combine into Kaz Resources Inc., expected to trade on Nasdaq as KAZR. The Trump brothers’ precise financial interests have not been fully disclosed.

Cantor Fitzgerald, run by Secretary Lutnick’s two sons, helped raise $210 million for one of the lead investors working with Dominari on the deal—the kind of assignment that typically produces millions in fees.

The New York Times reported that it counted $18.6 billion in federal backing for 60 critical minerals projects since President Trump took office, with the Trump and Lutnick families having some ties to at least 14 of them.

The new disclosure shows this wasn’t a one-off. Trump reported capital gains of $100,000 to $1 million on shares of MP Materials, the rare-earth miner whose stock jumped roughly 50 percent the day the Pentagon announced it was taking a 15 percent equity stake. And Donald Trump Jr.‘s venture fund, 1789 Capital, bought into Vulcan Elements in August 2025—three months before the Pentagon handed Vulcan a $620 million loan that helped push its valuation from around $200 million to nearly $2 billion. Different metal, different shell company, same shape.

Semiconductors and the United Arab Emirates

The UAE side of the ledger runs on a similar mechanism, just with chips instead of ore. As the Wall Street Journal detailed earlier this year, four days before Trump’s second inauguration, an Abu Dhabi vehicle backed by Sheikh Tahnoon bin Zayed Al Nahyan—the UAE’s national security adviser, director of the country’s largest sovereign wealth fund, and brother of the UAE’s president—bought 49 percent of World Liberty Financial for $500 million. Eric Trump signed the deal, and roughly $187 million landed in Trump-controlled accounts before anyone outside the deal knew it had happened. The CEO of World Liberty Financial is Zach Witkoff, son of Steve Witkoff, President Trump’s special envoy to the Middle East and key member of the administration’s foreign policy team.

Meanwhile, White House AI czar David Sacks pushed hard to loosen Biden-era chip restrictions on the UAE despite his fund’s ties to Tahnoon’s own sovereign wealth vehicle. The one senior official resisting that push, National Security Council Senior Director David Feith, was fired within weeks. Days before it would have taken effect, the administration killed the Biden export rule; by May, Washington had agreed to sell the UAE 500,000 advanced Nvidia chips a year, with one-fifth earmarked for Tahnoon’s own AI company. That same month, another Tahnoon vehicle ran $2 billion through World Liberty’s stablecoin into Binance—the exchange whose founder Trump pardoned five months later.

These private profits for Trump administration officials and their family members sit downstream of specific US international trade and investment practices, including efforts to mitigate China’s chokehold on tungsten, other rare earths, and semiconductor export controls. None of these policy goals is necessarily illegitimate on its own terms, but the means of pursuing those goals are transforming ordinary policy into family enrichment schemes.

Update: This post has been revised to clarify the timing of the EXIM and DFC Letters of Interest, the structure of the Trump family’s financial exposure to the Kazakhstan tungsten deal and Cantor Fitzgerald’s role in the transaction.

Last week, the Trump administration announced it is offering $17.5 billion in financing to build five new two-reactor nuclear power plants featuring Westinghouse Electric’s AP1000 large reactor. The offer comes on the heels of last November’s Trump announcement of an $80 billion deal with Westinghouse intended to jump-start a fleet of new US reactors.

Government doling out money (and other breaks) to nuclear power is old hat. But the two announcements, taken together, put a new spin on the practice of politicians picking winners and losers in the marketplace: Under the November deal, Uncle Sam could get a 20 percent equity stake in Westinghouse if its nuclear business booms in the coming years. Writing at the time, I predicted that it probably won’t happen: Nuke makes lousy business sense, as energy economist Steve Thomas recently detailed in Cato’s policy magazine Regulation, and it’s doubtful energy companies would take on such risk. Apparently, my prediction’s been right so far, so the Trump administration is offering a sweetener to get things going. 

Read the rest of this post →

A new anti-data-center push seems to be uniting the left and the right, with critics ranging from Sen. Bernie Sanders and Rep. Frank Pallone to Sen. Josh Hawley and Rep. Nancy Mace calling for some form of federal limitation on data center construction. As with any expansion, there are legitimate policy conversations to have about resource use and other underlying issues. However, a moratorium on the construction of a particular kind of facility based on arbitrary ceilings when demand is growing is a misguided policy sledgehammer that could not only derail critical innovation but also fail to resolve the underlying policy issues.

Data centers aren’t new, but the rising backlash and policy proposals to ban construction come at a critical time for innovation. The need for data centers should not be seen as an either/​or but rather as an opportunity to consider better solutions to policy questions around resource use and other related issues facing the communities where they are built. If we aren’t careful, bad policies like data center construction moratoriums and outdated energy policy could backfire — not only on progress in artificial intelligence, but also on a range of current and future technologies, including cloud computing and quantum computing.

The Rise of Data Center Construction Moratorium Policy and Its Problems

The most concrete federal effort to block data center construction is the AI Data Center Moratorium Act. Proposed by Senator Bernie Sanders on March 25, 2026, this Act would halt construction of new or existing data centers until a comprehensive federal AI regulatory framework is enacted. In the months since, the Overton window has shifted, with both the public and policymakers seemingly expressing a desire to pause construction of critical infrastructure for current and emerging technology demands.

The federal government is not alone. As of the end of June 2026, 116 municipalities have imposed local moratoriums. At the state level, California, Minnesota, Montana, and Kansas have all passed legislation that either incentivizes or disincentivizes data center construction in their states. These are in addition to the countless AI bills proposed in state legislatures across the country.

Unsurprisingly, states and localities have been and will continue to both restrict and enable datacenter construction for the foreseeable future.

Responding to Critics’ Common Concerns

Senator Sanders justified his bill by arguing two main points common among data center critics. His first claim is that AI is advancing too quickly and that Congress is moving too slowly. Because of this, Congress needs a data center moratorium to give it time to catch up with the technology. His second claim is that data centers pose environmental and economic risks to their locality by raising utility rates and consumption.

The data on data centers, however, do not bear out these claims. Data centers are needed to meet increasing demands beyond just AI, including the volume of connected devices and data Americans use regularly. Our capacity needs have expanded with AI, but demand had been growing even before AI’s exponential leaps. Data centers are critical to AI and the progress it enables across a variety of areas. AI products power not only the direct-to-consumer chatbots many have come to rely on to improve access to information but also an array of products that can improve health, enhance cybersecurity, or support responses to natural disasters. All of this is to say AI’s benefits are far from merely ephemeral.

When it comes to concerns about energy usage or utility rates, there are better solutions than the moratorium sledgehammer, and the data indicate such claims are vastly overblown. Multiple studies have found that while energy rates have increased, it’s far from clear that data centers are to blame. In fact, one study even found a connection between lower retail energy rates and data centers.

The reality is that energy policy keeps the grid and our energy supply outdated, which can increase costs. A better solution is to pursue energy policy reform, as our colleague Travis Fisher has written, to enable greater energy abundance and innovation.

Similarly, water usage in some regions of the country is a legitimate debate. But it is a problem that is far broader than just data centers. In aggregate, golf courses use significantly more water than data centers but rarely face the same kind of backlash.

Why The Backlash Could Backfire

The reality is that technological advancement is overwhelmingly a force for good. Some communities don’t want a new data center, while others may embrace the millions in tax revenue and hundreds of high-paying jobs that come with it. Individual communities are best positioned to know the specific trade-offs they face with any large investment, as they have been for decades.

Data centers date back about 70 years to when IBM was building them to support the cutting-edge technology of the day: mainframe computers. The type of data centers we think of today can trace back over 25 years. Cities and regions like Loudon County, Virginia, have worked with communities and companies to find a balance that enables infrastructure investments that provide economic opportunities and development while responding to community-specific concerns. More recently, however, these debates have spiraled from the typical debates over new construction or development to something much more heated.

The backlash against datacenters and AI has, at times, mushroomed into irrationality and violence. On April 10, 2026, a Molotov cocktail was thrown at Sam Altman’s home. Two days later, the same home was struck by gunfire from a vehicle. In 2025, a 25-year-old was arrested and charged with an act of terrorism for his plan to burn down xAI’s Memphis Colossus datacenter. These are paired with shootings at the homes of local councilmembers, physical confrontations at town meetings, and death threats against county commissioners.

This is particularly concerning, given the often-overlooked benefits of data centers. One study estimated that data center construction could create 4.7 million construction and skilled trade jobs in the United States. Many skilled trade jobs last beyond the initial construction, given the increased need for HVAC technicians and electricians, among other jobs, to support the center’s operation.

For users, there is a growing demand for less latency and more storage. The downside is that without the necessary infrastructure to support the demands of modern startups, it grows more difficult to start the next great company in a garage. This could also mean that those needing this level of access to develop cutting-edge AI applications must migrate to regions that provide the necessary data center infrastructure to support it.

This is in part why a federal moratorium would be particularly concerning. Companies and talent need to know they will not face unnecessary roadblocks from the government, particularly for capital-intensive investments. Given the global nature of AI competition, preventing such a critical buildout could hinder the US’s global competitiveness.

Conclusion

The debate over data centers has now jumped from a local issue to a national one. While the debate over AI policy should largely be a federal issue given its interstate nature, the issue of data centers is much more local and intrastate.

This should be concerning not only for its impact on infrastructure but also because a federal data center moratorium would represent an extraordinary reach by the federal government into hyper-local decisions about the development of individual plots of land across the country. Giving the federal government power over such a hyper-local issue might not be limited to data center construction and could spill over into an array of issues.

The goal should be to have a market that allows capital investment to increase supply and meet growing demands. Whether it’s moratoriums that stop needed development or artificial inducements that could result in overbuilding or cronyism, state and federal government intrusion makes finding the actual correct balance more difficult for the local communities most directly impacted.

Banning data centers comes at a critical moment for AI. But they will impact much more, including not only the storage and speed we’ve come to expect in the cloud era, but also the advancements needed for the next generation of technology, such as quantum computing. The underlying concerns driving data center bans can be addressed with far better policy approaches that allow innovation while actually solving long-standing issues. Moratoriums on data center construction both fail to solve the concerns about resource usage and delay critical and potentially life-saving progress.