Trump administration adds six chipmakers to U.S. stock portfolio in $874 million CHIPS Act expansion

The Commerce Department has signed seven nonbinding letters of intent offering up to $874 million in CHIPS and Science Act incentives.

Six companies would become new additions to the government’s expanding corporate portfolio, while GlobalFoundries already has a separate proposed federal equity arrangement.

Every final award would require the government to receive a minority, noncontrolling stake.

The agreements remain subject to additional diligence and approval, meaning the funding amounts and investment terms could still change before the deals are completed.

This is not simply another round of semiconductor grants. We are watching a major change in how Washington supports strategic industries. Instead of providing public money without ownership, the Trump administration is increasingly positioning the government as financier, shareholder, customer and regulator at the same time.

Seven semiconductor deals target the AI compute supply chain

The largest proposed award is up to $300 million for GlobalFoundries. The funding would accelerate domestic research into co-packaged optics, a technology that places photonic connections closer to artificial intelligence processors.

The goal is to move enormous volumes of data faster while reducing the energy consumed by advanced computing systems. Commerce officials said the investment could accelerate development of the technology by two to three years.

Kepler would receive up to $245 million to develop high-performance AI memory using three-dimensional and ferroelectric technologies.

Memory has become one of the most important bottlenecks in AI infrastructure because even the most powerful processors cannot work efficiently when data cannot be supplied quickly enough.

Multibeam Corporation is in line for up to $140 million for advanced packaging systems that assemble, stack and connect multiple chips.

Packaging has moved from a supporting role to a central part of semiconductor performance as manufacturers combine specialized processors, memory and communications components into tightly integrated systems.

Extropic would receive up to $75 million to develop thermodynamic sampling units. These experimental processors use natural thermal fluctuations to solve probability-heavy problems in optimization, simulation and AI while consuming less energy than conventional computing approaches.

Thintronics could receive up to $50 million for ultra-low-loss dielectric materials used between layers of semiconductor connections. Such materials are designed to reduce signal loss as chips become denser and computing systems move larger amounts of information.

OBSIDIA Semiconductors is slated for up to $34 million to develop systems capable of identifying counterfeit or malicious electronic components without damaging them. Aeluma would receive up to $30 million for larger semiconductor substrates used to manufacture photodetectors and lasers for AI photonic connections.

Trump administration changes how CHIPS Act money is distributed

The CHIPS and Science Act was signed into law in 2022 to expand domestic semiconductor manufacturing, research and supply-chain resilience. Under the previous administration, the program largely relied on grants, loans and tax incentives.

The Trump administration has changed that model. Federal assistance is increasingly being tied to equity, warrants or ownership-like rights designed to give taxpayers a financial return when a publicly supported company succeeds.

Commerce Secretary Howard Lutnick said the latest investments would strengthen domestic capabilities, create high-paying jobs and preserve American leadership in the semiconductor industry.

Supporters can point to a straightforward argument. When public money absorbs part of a company’s financial risk, the public should also share in the potential upside. Equity could recover more value than a traditional grant, especially when the government supports a company before its technology reaches commercial scale.

The difficulty is that semiconductor research remains expensive, uncertain and slow. A promising laboratory breakthrough may take years to become a dependable commercial product. Some technologies may never achieve widespread adoption.

The value of the government’s stakes will therefore depend on technical milestones, market demand, management decisions and the final terms negotiated with each company.

Federal corporate portfolio may now reach 30 companies

The six proposed additions would bring the Trump administration’s corporate portfolio to 30 companies, according to a tally compiled by Cato Institute analyst Tad DeHaven.

That count includes completed investments, proposed equity arrangements and the federal golden share in U.S. Steel, which provides Washington with specified veto rights rather than a conventional financial interest.

The portfolio stretches beyond traditional semiconductor manufacturing. It includes businesses connected to quantum computing, critical minerals, steel and nuclear energy.

In May, the administration announced approximately $2 billion in proposed funding for nine quantum computing ventures.

The package included an IBM-related project and a separate award of up to $375 million for GlobalFoundries, with federal equity attached to the planned investments.

Intel remains the most visible example of the strategy. In August 2025, the government converted federal funding commitments into 433.3 million nonvoting Intel shares priced at $20.47 each. The transaction gave Washington a stake of roughly 10 percent without a seat on Intel’s board.

Taxpayer upside comes with government oversight concerns

The new approach raises questions that extend beyond whether a particular stock rises or falls.

The government can influence portfolio companies through regulation, federal procurement, trade restrictions, export controls and future subsidies. Its decisions can also affect competing businesses that do not receive federal investment.

DeHaven argues that minority ownership does not eliminate potential conflicts because Washington may simultaneously regulate a market and hold assets whose value depends on government policy.

He also noted that public announcements frequently provide limited details about company valuations, shareholder rights, oversight procedures and plans for eventually selling the stakes.

Those concerns do not erase the national-security case for semiconductor investment. Advanced chips are essential to AI systems, weapons, communications, medical research, industrial automation and financial infrastructure.

The latest proposed awards focus on vulnerable sections of the compute supply chain, including memory, packaging, photonics, substrates and component authentication.

The strategy therefore reaches beyond the construction of expensive fabrication plants and targets the technologies that connect, protect and improve advanced processors.

The central question is no longer whether Washington will intervene in strategic technology markets. It already does.

The question is whether the United States can build a transparent and disciplined investment framework that protects taxpayers, avoids favoritism and helps promising technologies cross the difficult gap between research and commercial production.

With six more companies preparing to enter the portfolio, the federal government’s role in corporate America is becoming broader, more direct and harder to dismiss as temporary.

The $874 million CHIPS Act expansion could strengthen the domestic AI hardware industry. Still, it will also become a defining test of how far Washington can move from subsidizing private companies to owning pieces of them.