Bill Gates is not objecting to government support for strategic technology. He is warning that government ownership changes the contest itself.
Bill Gates has put his finger on the uncomfortable part of America’s new industrial policy: the state is no longer only writing checks, setting rules, and buying technology. It is starting to own pieces of the companies competing for those checks, rules, and contracts.
Times of India summarized Gates’ CNBC remarks on June 13, saying the Microsoft co-founder warned that U.S. equity stakes in companies such as Intel and IBM could tilt decisions toward firms the government already owns, rather than firms with the best technology. That is not a small complaint from someone who dislikes public spending. Gates has long argued for government funding in hard science, clean energy, and basic research. His objection is more precise. Once Washington becomes a shareholder, the rules of the game become harder for everyone else to read.
The Intel deal is the cleanest example because the numbers are public. Intel said in August 2025 that the U.S. government would invest $8.9 billion in its common stock, buying 433.3 million shares at $20.47 each for a 9.9% stake. The money came from $5.7 billion in unpaid CHIPS Act grants and $3.2 billion tied to the Secure Enclave program for Defense Department chips. Intel also said the stake was passive, with no board seat or governance rights.
Passive ownership sounds reassuring until the government is also the largest customer, the regulator, the grant maker, and the national security referee. Intel itself saw the problem. In a securities filing reported by Reuters and The Washington Post after the deal, the company warned that government ownership could hurt international sales, create litigation risk, and complicate future grants. That is the dull sentence in the filing that matters. The risk is not only political theater. It is operational.
Washington’s move into equity has already moved beyond one troubled chipmaker. Axios reported in May that the Trump administration agreed to take equity stakes in nine quantum computing companies as part of a $2 billion CHIPS Act push. Half of that money was set aside for IBM to build a quantum chip foundry called Anderon, while venture-backed companies including Quantinuum, PsiQuantum, Atom Computing, and Diraq were also listed for funding.
That makes Gates’ concern more serious than a passing criticism of one transaction. The pattern is spreading across the hardware stack that AI depends on: semiconductors, quantum chips, advanced manufacturing, and the domestic supply chains behind them. These are not app companies that can be rebuilt in a garage over a weekend. A fab, a quantum facility, or a secure chip supply chain takes years of capital, permits, engineers, and customers willing to commit before the technology is fully proven.
Government support may be unavoidable in some of these markets. Taiwan’s position in leading-edge manufacturing, China’s industrial subsidies, and the cost of new fabs have made a purely hands-off approach look thin. The question is whether support has to become ownership, and whether ownership can stay neutral once contracts and procurement decisions begin to flow through the same departments that hold the stock.
That is where startups should pay attention. A young AI hardware company does not only compete with Intel on engineering talent or chip performance. It may now compete with Intel’s balance sheet, Intel’s grant history, and a federal shareholder that has a political interest in seeing its investment succeed. Even if no official ever says that out loud, founders and investors will price the risk.
Startups now face a different customer
The hardest part for startups is not that the government is spending money. Many would welcome more serious public purchasing in AI infrastructure, defense technology, chips, energy, and quantum computing. The harder part is competing for a customer that may already have an economic interest in a rival.
Procurement is already slow. Add federal equity stakes, and every award can start to look suspect. If Intel wins a major government AI infrastructure contract, competitors will ask whether its technology won or whether Washington was protecting its position. If IBM wins quantum work after a federal-backed foundry plan, rivals will ask the same question. That suspicion can damage the winners too, because customers outside the United States may wonder how independent the companies really are.
There is also a capital markets problem. Venture investors can handle tough competition. They are less comfortable with competition whose rules may change after a CNBC interview, a White House post, or a grant conversion. Gates’ point about predictability matters here. Chip manufacturing and AI infrastructure are long-cycle bets. A founder cannot raise money on a 10-year hardware plan if the market believes the government can abruptly turn one incumbent into a state-backed national champion.
The irony is that this policy is meant to strengthen American technology. It might, in some cases. Intel needed capital for a turnaround, and the United States has a clear strategic interest in domestic chip production. IBM has real quantum research depth, not a paper plan. But a country can support strategic technology while still making clear how contracts are awarded, how conflicts are managed, and whether better engineering can beat government-backed incumbency.
Gates is useful in this debate because he is not pretending the old system was perfect. Microsoft itself grew up around federal customers, antitrust scrutiny, and the long shadow of IBM. He knows large technology markets are never clean contests. His warning is that America is making them murkier at the exact moment AI infrastructure needs more entrants, more technical discipline, and more trust in the scoreboard.
The next test will not be whether Washington takes another stake. Recent moves suggest it probably will. The test is whether a startup with a better chip, a better quantum device, or a cheaper AI infrastructure design can still win when the government already owns part of the company standing across the table.
Also read: Disney is turning AI coding into an operating discipline • UAL's Nerve Lab puts children's screen time under a sharper lens • AI agents are becoming guards for EV charging networks
