Most of Washington's $215 million quantum prize pays only after teams reach 100 logical qubits

Add up the three prize pools and $200 million of the $215 million, about 93%, goes only to teams that have already shown an error-corrected machine of 100, 150 or 200 logical qubits ^[17][18]. If all of the remaining $15 million went to early-milestone awards at the $1.5 million cap, it would pay for ten of them ^[19]. Logical qubits use error correction to make calculations more reliable ^[16]. The entry requirement is a machine with at least 100 of them that can perform hundreds of millions of fault-tolerant operations, according to crypto.news ^[9].

For a startup, the contest pays after the work is done. Before a team clears that bar, federal money comes in awards of up to $1.5 million ^[3], so the years spent building toward 100 logical qubits still run on private capital, or rather on investors willing to bet that the $100 million pool ^[10] will still be funded when they get there. That part is open. The department said the awards depend partly on future congressional appropriations ^[11]. The government is not paying for the build-out up front. The counter-case is that a federal prize with a stated dollar figure gives venture investors a milestone they can underwrite, and could pull private rounds forward. I think that effect is real but small while the money still waits on Congress ^[11].

Nothing in the package sets a date for the threat to bitcoin keys. The announcement does not describe the quantum contest as a cryptocurrency security program, crypto.news reported ^[4]. Europol, in two reports on Oct. 7, said a sufficiently powerful machine could derive private keys from public keys that are already visible and move coins without the owner's permission ^[14]. The agency did not claim existing machines can do this, and it called the timing uncertain ^[15]. Three EU financial watchdogs issued a similar warning on Sept. 23 ^[7], and Cointelegraph reported that no quantum computer capable of the attack exists ^[8]. Neither source says how many logical qubits an attack on bitcoin signatures would take, so the contest's 100-qubit bar cannot be placed against Europol's warning from this record.

Two outcomes would change that. If a team claims the 100-qubit pool soon, developers get a federal judgement that a machine of that size exists ^[10], and the Energy Department's separate $45 million laboratory program for testing and verifying quantum hardware and software gives the government its own way to check ^[13]. If the applicants are far short of the bar, the pools sit unclaimed and developers learn nothing new. The view here, that this package leaves crypto's clock where Europol left it, is wrong if the first outcome comes quickly and the EU authorities tighten their timing in response.

The bigger number in the package is compute for agencies. Eleven companies pledged $2.4 billion in AI tools and compute credits for research at 15 federal agencies ^[5]. Nvidia's $1 billion, AMD's $500 million, OpenAI's $200 million and $150 million apiece from Anthropic and Google add up to $2 billion, so the other six companies account for $400 million between them ^[6][20]. Against that, the quantum contest is under 4% of the more than $6 billion headline ^[22].