Brookhaven Lab Taps PsiQuantum's Construct Platform for Fault-Tolerant Quantum Algorithm Research

Brookhaven National Laboratory and PsiQuantum announced this week that Brookhaven researchers will get access to Construct, PsiQuantum's software platform for designing and simulating quantum algorithms meant to run on future fault-tolerant quantum computers. The collaboration is explicitly framed as support for the Department of Energy's Quantum Genesis initiative, which has a genuinely aggressive target: a scientifically relevant, fault-tolerant quantum computing capability by 2028.

For anyone who's watched the quantum industry cycle through hype waves for the past decade, "fault-tolerant" is the key phrase here. Today's quantum hardware is noisy. Qubits decohere, gates introduce errors, and most near-term devices can only run shallow circuits before the noise swamps the signal. Fault-tolerant quantum computing (FTQC) uses error-correcting codes to spread logical information across many physical qubits, letting a machine run long, deep circuits reliably. That's the regime where quantum computers are expected to actually outperform classical machines on hard problems, not just match them on toy benchmarks.

The catch: nobody has fault-tolerant hardware at useful scale yet. So the algorithm and applications work has to happen ahead of the hardware, which is exactly what Construct is built for.

What Construct Actually Does

PsiQuantum describes Construct as the first comprehensive platform for building fault-tolerant quantum algorithms. Practically, that means it gives researchers a toolchain to:

• Design quantum circuits targeting FTQC architectures rather than today's noisy intermediate-scale (NISQ) devices
• Write and simulate quantum algorithms before any physical qubits are involved
• Estimate and optimize resource requirements, i.e., how many logical qubits, how many gates, how much runtime a given algorithm will actually need on real fault-tolerant hardware

That last point matters more than it might sound. Resource estimation is one of the unglamorous but essential parts of quantum algorithm design. If you don't know whether your algorithm needs 100 logical qubits or 100,000, you don't know whether it's a research curiosity or something that could run this decade. PsiQuantum's team spent years building out this tooling internally before opening it up. In May 2026, the company made Construct freely available to all developers, a move that positions it less as a proprietary lock-in tool and more as shared infrastructure for the field, similar to how open-source ML frameworks lowered the barrier to entry for deep learning research.

Brookhaven's Associate Laboratory Director for Discovery Technologies, Gabriella Carini, framed the deal as a template for how Quantum Genesis is supposed to work. "DOE's new Quantum Genesis initiative requires innovative partnerships across our national labs, industry, academia, and government," she said. "This is a great example of one such collaboration." PsiQuantum's VP of Government Relations, Heath Bumgardner, echoed that framing, calling it a demonstration of "how effective public-private collaboration can help accelerate scientific discovery."

That language is standard press-release fare, but the underlying structure is worth paying attention to. Quantum Genesis sits as a foundational piece of DOE's broader Genesis Mission, the department's AI-focused push to expand computational power for scientific discovery. Pairing a photonics-based quantum hardware company with a national lab that has deep bench strength in materials science and high-energy physics is a fairly deliberate bet: get the algorithm design work done now, on freely available tooling, so that whoever eventually has working fault-tolerant hardware (PsiQuantum's own photonic systems, or someone else's) has a backlog of validated, resource-costed algorithms ready to run.

PsiQuantum's pitch on the hardware side leans on photonics, using photons instead of superconducting circuits or trapped ions as qubits. The company argues this approach lets it piggyback on existing high-volume semiconductor manufacturing and cryogenic infrastructure, which in theory helps it scale faster than approaches that need bespoke fabrication for every qubit. Whether that architectural bet pays off is a separate and much longer story. But it's the reason PsiQuantum is positioned as a serious enough player for a DOE national lab to build a formal research collaboration around its software stack rather than just watching from the sidelines.

The application areas being targeted aren't surprising, but they're the right ones: materials science, pharmaceutical research, and cryptography are the usual suspects for where fault-tolerant quantum computing is expected to have real advantages over classical simulation. Quantum chemistry and materials problems in particular involve simulating electron interactions that scale exponentially on classical hardware, exactly the kind of problem quantum computers are theoretically suited for, once you can run circuits deep and long enough without errors piling up.

Key Takeaways

The 2028 target for Quantum Genesis is aggressive by any historical measure in this field, and skepticism is warranted. Fault-tolerant quantum computing at scientifically relevant scale has been "five to ten years away" for most of the last decade, and hardware timelines have a well-earned reputation for slipping.

What's notable about this particular announcement isn't the timeline promise, though, it's the sequencing. Rather than waiting for hardware to mature and then scrambling to write algorithms, DOE and PsiQuantum are pushing algorithm development and resource estimation to happen in parallel, using simulation and design tools that don't require physical fault-tolerant qubits to exist yet. That's a sensible hedge regardless of whether PsiQuantum's photonic hardware or a competitor's approach ends up winning the race to useful scale.

For researchers outside this specific partnership, the bigger signal is that Construct is free and open to any developer, not just Brookhaven. That lowers the barrier for smaller labs and academic groups to start prototyping fault-tolerant algorithms now, rather than waiting until hardware access becomes the bottleneck. Whether that translates into a meaningful head start once fault-tolerant machines actually ship is the question worth revisiting in a couple of years.