Chip Subsidies Are Slow. Private Money Moves Faster.

Making semiconductors is expensive. Building a single new chip factory can take years and cost billions of dollars. On the other hand, the global chip shortage of 2020-22 showed how risky it is to rely on imported chips. In response, the US and European governments launched big programmes to bring more chip production closer to home.

Great idea, but with one big problem. Government money doesn’t move as fast as the market.

The US CHIPS and Science Act set aside $39 billion to encourage chip manufacturing and another $11 billion for research and development. But by April 2026, the US Commerce Department had paid out only $13.1 billion of its direct funding. That is about 42% of the money it had awarded. The Government Accountability Office (GAO), a US government watchdog, also found that some projects were running behind schedule.

The story in Europe is no different. The European Chips Act set an investment target of €43 billion. But a 2026 review by the European Parliament found that most major factories the programme backs have not yet started production. Chip projects take a long time to build and to reach full output.

Subsidies Move at the Speed of Paperwork

Building a chip factory is hard enough without extra layers of approvals and changing programme rules. In the US, projects that receive CHIPS money must meet set targets for construction, production and hiring before the money is released. The GAO found that $30.9 billion had been awarded across 40 projects as of July 2025, but companies expected some of those projects to run until 2033. Other problems include building costs, permits, utility connections and a shortage of skilled workers.

The rules continue to change. In 2025 and 2026, the US government changed several existing CHIPS agreements. By April 2026, 14 awards had been changed, on matters ranging from project deadlines to hiring requirements. One recipient’s expected completion date was pushed back by two years!

Meanwhile, Europe has been trying to make its own system simpler. A 2026 proposal from the European Commission said that permits and design for advanced chip factories in the EU took an average of 7.5 months longer than in the main competing regions. The Commission estimated that each year of delay could add about 5% to a project’s total cost.

Private investors have different considerations. They need to estimate whether a project can make money. That’s a much faster way to decide. 

Private Investment Fills the Gap

The strongest evidence comes from AI infrastructure. In August 2026, Nvidia announced deals with major financial firms, including Blackstone, Apollo, BlackRock, Brookfield, Goldman Sachs and KKR. The aim is to bring in more than $500 billion of outside money over time for AI infrastructure. That money is meant to pay for the chips, data centres and related equipment that Nvidia’s customers need.

This doesn’t mean investors have already spent $500 billion. It’s a planned pool of financing. But it shows how much private money is now available to the chip and computing industry.

Companies are also placing their own bets on chip design. In September 2025, Meta announced plans to buy Rivos, a startup that works with RISC-V, an open-source chip design, to speed up its work on custom AI chips. The deal made simple business sense: making its own chips can mean relying less on outside suppliers, and it may lower costs.

The same motive is pushing chipmakers to spend their own money. Micron announced a planned $200 billion expansion of its US manufacturing in 2025. And in July 2026, Nvidia and South Korea’s SK Group announced an AI infrastructure and memory plan worth more than $500 billion.

Only Speed Matters

Private money has an advantage that subsidy programmes lack. It goes where investors expect to make money.

When demand for computing grows, investors don’t need a government programme to show them the opportunity. They can pay for a data centre, back a chip startup, buy new equipment or take a stake in an established manufacturer.

The chip race is a useful test of two approaches. Governments can commit huge sums, but handing out that money takes rules, approvals and checks. Private money can move as soon as investors see a business opportunity.

Chip technology changes very fast, but factories take years to build. In an industry like that, only the private sector can bring in the speed needed for successful execution. Period.