PsiQuantum's Billion-Dollar Photons Are Chasing a Million-Qubit Machine

The most expensive machine that doesn't exist yet is a box of light. It will cost billions, require a dedicated power plant, and, if it works, will perform calculations for chemistry and materials science that would take a classical supercomputer longer than the age of the universe. PsiQuantum, a Palo Alto startup founded in 2016, is building that machine. Its bet is not on the superconducting loops favored by Google and IBM, but on particles of light,photons,guided by silicon chips made in a commercial semiconductor foundry. The goal is a million-qubit, fault-tolerant system, a number the company says is the minimum for commercial utility. They are not alone in the photonic race, but with over $3.2 billion in disclosed funding and a new CEO from Advanced Micro Devices, they are certainly the best-funded [psiquantum.com].

The photonic wedge

Every quantum computing architecture is a trade-off. Superconducting qubits are fast but error-prone and must operate at temperatures near absolute zero. Photonic qubits, which use particles of light, are more stable at room temperature and can travel over optical fibers, but they are notoriously difficult to manipulate and entangle at scale. PsiQuantum's founders,quantum physicists Jeremy O'Brien, Terry Rudolph, Pete Shadbolt, and Mark Thompson,bet that the scaling advantages of photonics would ultimately win. Their core insight was industrial: they could manufacture their photonic qubits using the same silicon photonics platform and tier-1 foundries, like GlobalFoundries, that produce conventional chips [psiquantum.com] [multiples.vc]. This means they aim to build tens of thousands of wafer-scale quantum chips, assembling them into cryogenic cabinets that hold hundreds of chips at a time. The promise is a path to a million physical qubits, a scale they believe is necessary for the error correction required to run useful algorithms for problems like designing better batteries or more efficient fertilizers.

The software before the hardware

While the hardware is years from completion, PsiQuantum is already shipping software. Its platform, called Construct, is a cloud-based environment for designing and testing fault-tolerant quantum algorithms. It includes visual circuit designers, Python libraries, and tools to estimate the computational resources an algorithm will need. The company is not waiting for its own hardware to find users; it has partnered with Brookhaven National Laboratory, where researchers are using Construct to explore algorithms for future fault-tolerant systems [bnl.gov]. This serves a dual purpose. It builds a developer ecosystem and gathers data on real-world use cases, while also providing a tangible product for enterprise and government partners who are funding the hardware development. Named partners already include Airbus, Lockheed Martin, Mercedes-Benz, and Mitsubishi Chemical, all of whom are exploring specific industrial applications [technologyreview.com].

The capital and the calendar

PsiQuantum's funding history reads less like a venture capital timeline and more like a national infrastructure project. The company has secured capital from a mix of sovereign wealth funds, asset managers, and multiple governments.

Series E (Sep 2025) | 1000 | M USD
Series C (Jan 2026) | 620 | M USD
Round (May 2026) | 1500 | M USD
U.S. Dept. of Commerce Grant (May 2026) | 100 | M USD
DARPA Grant (Jul 2026) | 125 | M USD
Australian Gov't Grant (Apr 2024) | 620 | M USD

This capital is tied to specific, ambitious deployment plans. The company has committed to building its first utility-scale, fault-tolerant quantum computer in Brisbane, Australia, with a target of 2027 [psiquantum.com]. A separate $100 million agreement with the U.S. Department of Commerce is intended to support a facility in Chicago, expected to create at least 150 jobs [psiquantum.com, May 2026] [psiquantum.com, July 2024]. The scale of funding implies staggering capital expenditure, but it also reflects the confidence of backers like BlackRock, Temasek, and Baillie Gifford in a winner-take-most outcome for the first company to reach utility scale.

Leadership for the build phase

In February 2026, founder Jeremy O'Brien transitioned from CEO to Executive Chairman, and Victor Peng, the former president of Advanced Micro Devices, was named interim, then permanent, CEO [psiquantum.com]. This is a classic deeptech playbook move: the scientific founder steps back as the company enters its capital-intensive construction phase, and a seasoned hardware industry executive with deep supply-chain and manufacturing experience takes the helm. Peng's four decades in semiconductors at AMD and Xilinx are directly relevant to the challenge of building millions of qubits in a foundry. The rest of the founding team remains deeply involved in their scientific and engineering specialties, with Terry Rudolph leading theory, Mark Thompson on engineering, and Pete Shadbolt as chief scientific officer focused on scaling [technologyreview.com].

Where the light could scatter

The risks here are not subtle. They are the size of continents and the patience of pension funds. The entire venture is a marathon sprint to a technical milestone,a fault-tolerant, million-qubit machine,that has never been achieved. The competitive landscape is fierce, with well-funded efforts from companies like IBM and Google pursuing superconducting qubits, and other photonic specialists like Canada's Xanadu also in the race. PsiQuantum's bet on photonics is a high-stakes technical path. Furthermore, the company's valuation, reported at $10.5 billion as of May 2026, is a promise on a distant future revenue stream [sacra.com]. Any significant delay or technical setback could test the patience of even the most long-term-oriented sovereign wealth fund. The company's answer to these risks is embedded in its structure: massive capitalization to outspend and outlast competitors, partnerships that anchor demand, and a manufacturing-led approach designed for volume.

The next twelve months

The coming year will be less about flashy qubit count announcements and more about poured concrete and supply-chain logistics. Watch for progress on the ground in Brisbane and Chicago. The hiring of roles like Director of U.S. Government Program Management and Senior Photonics Engineers signals the shift to execution [careers.redpoint.com]. Technically, the focus will be on integrating more chips into their cryogenic systems and advancing the performance of their Construct software with partners like Brookhaven. The company must also begin to demonstrate that its modular, foundry-made approach can reliably produce the high-quality photonic components needed for error correction. Success won't be measured in user growth, but in milestones like wafer yield and the stability of their increasingly complex assemblies.

PsiQuantum's bet is that the unit economics of quantum computing ultimately favor the architecture you can manufacture, not just the one you can demonstrate in a lab. If a single useful quantum calculation for a new catalyst requires a million error-corrected qubits, the cost per reliable qubit becomes the only metric that matters. By that logic, using a GlobalFoundries fab is not just convenient, it's the whole point. The incumbent they must beat isn't another startup, it's the entire classical supercomputing industry for specific, valuable problems. If they can build their box of light, the energy saved on failed simulations alone could justify the build cost. If they cannot, it will be the most expensive physics experiment in history.

Sources

1. [psiquantum.com] PsiQuantum - Building the World's First Useful Quantum Computer | https://www.psiquantum.com/
2. [technologyreview.com, July 2026] PsiQuantum has a plan to make a massive quantum computer out of light | https://www.technologyreview.com/2026/07/14/1140356/psiquantum-plan-massive-quantum-computer-out-of-light/
3. [multiples.vc] PsiQuantum | https://multiples.vc/private-comps/psiquantum
4. [psiquantum.com, May 2026] PsiQuantum Signs $100 Million Letter of Intent with the U.S. Department of Commerce | https://www.psiquantum.com/news-import/us-department-of-commerce
5. [psiquantum.com, July 2024] PsiQuantum To Build First US-Based Utility-Scale Quantum Computer in Chicago, Illinois | https://www.psiquantum.com/news-import/illinois-iqmp
6. [bnl.gov] Brookhaven National Laboratory collaboration news | https://bnl.gov/newsroom/news.php?a=123142
7. [sacra.com] PsiQuantum | https://sacra.com/c/psiquantum/
8. [careers.redpoint.com] PsiQuantum job postings | https://careers.redpoint.com/companies/psiquantum/jobs/91764851-solution-business-development-lead
9. [thequantuminsider.com, 2026-07-22] DARPA grant announcement | https://thequantuminsider.com/2026/07/22/psiquantum-darpa-125m/
10. [businesswire.com, 20240429080449] Australian government grant announcement | https://www.businesswire.com/news/home/20240429080449/en/