Most quantum startups sell a promise: give us money and years, and eventually we will hand you a machine that changes everything. The awkward question is always whether the design actually works at full scale. On 22 July, the US defence research agency DARPA effectively said to PsiQuantum, right then, prove it, and attached $125 million to the request.
It is PsiQuantum’s most valuable US government contract to date, an expanded agreement under DARPA’s Quantum Benchmarking Initiative, or QBI. The money funds verification and validation work on the company’s utility-scale hardware design, its custom photonic manufacturing processes and the cryogenic plumbing that keeps everything cold enough to function. In plain terms, DARPA is not paying PsiQuantum to dream bigger. It is paying to have PsiQuantum’s existing claims poked, prodded and independently checked.
What makes PsiQuantum the odd one out
Most of the famous quantum names cool metal qubits to near absolute zero and control them with microwaves. PsiQuantum went a different route: photonic quantum computing, which uses single particles of light as qubits and runs them through chips built on the same kind of silicon-photonics production lines that already make telecom components. The pitch is manufacturability. If your qubits are made of light on standard-ish chips, you can, in theory, print millions of them in a real factory rather than hand-building a few hundred.
That is the bet, and it is a big one. PsiQuantum is one of only two companies still standing in Stage C, the final and most demanding phase of QBI, where the whole point is to confirm a company can actually build, deploy and operate a useful quantum computer as designed. The new tasking builds on a $31.8 million Stage C award from September 2025, so this is a serious step up, not a top-up.
The angle: governments are now underwriting the risk
Zoom out and a pattern shows. The expanded DARPA deal follows a May 2026 letter of intent with the US Department of Commerce for around $100 million in proposed CHIPS Act incentives, and PsiQuantum is building large Quantum Compute Centres in Brisbane and Chicago with government backing on both sides of the world. The state is increasingly the anchor customer and the de-risking balance sheet for quantum, because the timelines are too long and the odds too uncertain for ordinary commercial money to carry alone.
QBI exists to answer one blunt question: is it even possible to build an industrially useful quantum computer by 2033? That framing is refreshingly honest. Nobody involved is pretending the answer is a settled yes. The healthy reading of this $125 million is that it is not a prize for winning. It is payment for submitting to rigorous, sceptical scrutiny, which is exactly what a field drowning in hype needs more of. The unhealthy reading is that even the best-funded photonic player still has to be independently checked before anyone believes the machine will scale.
Either way, the era of taking quantum roadmaps on faith is ending. Someone with a budget and no incentive to flatter is now marking the homework.
Did you know: PsiQuantum’s qubits are individual particles of light, which means a big chunk of its “quantum computer” is built on the same silicon-photonics manufacturing tech already used to move data through the internet’s fibre-optic backbone.
Related on Top Tool Stack: Google’s self-correcting quantum chip · IBM’s self-running quantum chip