The AI Safety Menu

All laws

Build chips at home

Bigger subsidies for U.S. chip factories. More grants and tax credits to build advanced chip factories in the U.S. instead of Taiwan.

Grows lead by~3 daysof America's lead over China, over 3 years
No p(doom) effectNonefrom a 5% starting estimate
Holds up○ Training○ Lab's own use○ Public release
Enact it on the menuSee the findings

What it does

The government expands grants and tax credits for building advanced chip factories in the U.S., so America depends less on Taiwan for the chips AI runs on.

Applies to: Companies building advanced chip factories in the U.S.

Where things stand

Large CHIPS grants and loans and a 35% investment tax credit already exist, now with equity and renegotiation conditions. Taiwan still makes the vast majority of the most advanced chips.

Why it gains ~3 days

New chip factories take three to four years to build in the U.S., so a year in there are almost no new chips. The gain within a year is about a day. The payoff comes later, as insurance against a blockade of Taiwan.

Biggest unknown: Whether extra subsidies speed up factories that were coming anyway, and how much advanced packaging stays in Taiwan.

Its effect on p(doom)

Not a safety law. It makes America's AI chip supply less dependent on Taiwan.

It changes where chips are made, not how AI is built. It could lower the risk from a war over Taiwan, or entrench the race. No clear effect on p(doom).

The strongest case that it costs more

New factories take three to five years, so extra subsidies change nothing within a year, and much of the money goes to companies that would have invested anyway. Cheaper, more plentiful chips could also speed up the race rather than make it safer.

The debate

For

Against

Sources

Rough starting points, not precise forecasts. Lead costs assume China doesn't depend on U.S. models, the case least favorable to safety laws, and count 3 years. On the menu you can change every assumption and put in your own numbers. Last priced 2026-09-26.