ORCA Computing Targets Quantum Advantage for AI Within Two Years
London-based startup builds photonic quantum systems aimed at boosting generative AI without waiting for error correction
For ORCA Computing, the immediate future of quantum computing isn’t about waiting decades for fully error-corrected machines. It’s about delivering value right now, specifically in the rapidly evolving world of generative AI.
According to co-founder and CEO Richard Murray, ORCA’s mission is clear: integrate quantum computing into today’s AI systems to unlock applications that classical computing can’t handle alone.
“In a 24-month time frame, firstly, we'll reach quantum advantage,” Murray said. “Our systems will start being able to perform calculations that are not possible on a classical computer. But more importantly, we'll be doing that within existing AI hardware.”
This goal positions ORCA as one of the few quantum players squarely focused on near-term deliverables. Rather than relying on futuristic models that require decades of refinement, ORCA’s photonic approach is designed to work in real data center environments, accelerating machine learning tasks without requiring massive infrastructure overhauls.
“Quantum is valuable as soon as possible to companies,” Murray emphasized. “So with photonics, with our technology, you can allow that.”
How ORCA's System Works
Founded in 2019 as a spin-out from Oxford University, ORCA Computing builds quantum computers using photonics—specifically, off-the-shelf telecommunications equipment. But what sets ORCA apart is its hybrid use of atomic systems—rubidium-based memories that allow photons, which typically don’t interact, to influence each other.
“Photons are great, but one challenge is that photons have very long lifetimes,” Murray explained. “The problem is that they don't interact with other photons. So the atoms create a way for each photon to interact with other photons.”
This interaction, known as non-linear behavior, is essential for scaling quantum systems. Without it, photonic systems lack the entanglement and connectivity for quantum advantage. ORCA’s atomic integration provides a scalable path forward that many in the field see as technically elegant and commercially promising.
To avoid losing photons as they move through components—a critical risk in quantum computing—ORCA recently acquired a semiconductor company in Austin, Texas. The goal: to develop technology capable of changing the properties of photons without losing them.
“You must not lose photons,” Murray said. “The Austin team has a semiconductor platform that allows you to control the lights without introducing any losses. So the perfect quantum components.”
Semiconductors are a natural fit for Texas, which boasts a rich ecosystem in that field—something the UK currently lacks. The acquisition brings fabrication in-house and positions ORCA to integrate photon control tightly into its systems.
The company’s current offering, PT-2, represents a significant step forward. Launched in late 2024, PT-2 is already used by government and academic clients in the UK, US, and Europe. The system offers quantum-enhanced machine learning capabilities and is built to operate in conventional HPC environments.
“I think everyone is still curious what quantum advantage means,” Murray noted. “And I think that we will deliver a system that is also linked to applications.”
Next Launch: PT-3 and Beyond
The next system is PT-3, ORCA’s forthcoming quantum advantage platform, scheduled for release in early 2026.
“I think the first product that will really make a difference—change the way people talk about quantum computing—will be our PT-3 system,” Murray said. “It is our industrial commercial quantum advantage system… an accelerator of generative AI workloads.”
PT-3 builds on PT-2’s architecture but adds much-needed connectivity between photons, which PT-2 lacks despite sufficient photon volume. That connectivity will come via Austin's newly acquired semiconductor technology, completing the puzzle needed to achieve quantum advantage.
“You need enough photons. And you need to be able to control and connect those photons,” Murray said. “We already have enough photons. But we don't have the connectivity between the photons yet.”
The company’s vision for PT-3 is to serve as a quantum accelerator embedded directly within classical high-performance computing environments—particularly for AI and optimization workloads. Rather than operating in isolation, ORCA’s quantum machines will plug into familiar workflows and toolkits like Nvidia’s CUDA-Q platform.
“It is designed to integrate within classical high-performance computing,” Murray said. “It is an accelerator of generative AI workloads, which is in our view the quickest way to deliver meaningful applications with a quantum computer.”
ORCA is finalizing development and expects PT-3 to enter the market in early 2026.
Global Competition and Talent Needs
Murray shared his views with TechJournal.uk in an interview on the sidelines during the Commercializing Quantum Global 2025 summit, held in London and organized by The Economist on May 13-14. Among competing photonic quantum startups like PsiQuantum and Xanadu, ORCA stands out for its short-term commercial focus.
“Those guys, although they're pursuing the same hardware, are purely focused on error-corrected systems, which I believe will take longer to come to market,” Murray said. “ORCA is similar in hardware to PsiQuantum and Xanadu, but taking a different approach in order to deliver some of these near-term applications.”
He added that, in the long term, quantum computing will not be a zero-sum game. Instead, he sees a future where multiple modalities coexist, each suited to different functions, similar to how modern computers use multiple forms of memory and storage.
“I think you'll see all of the technologies being relevant and useful,” he said. “A future quantum computer may well contain not just one modality, but some different modalities to deliver all the different functions you need.”
On the global front, Murray praised China’s efforts in photonics, noting that ORCA’s work on Gaussian Boson Sampling (GBS) statistics is aligned with China’s Jiuzhang system. “I think it is neck and neck,” he said. “The progress that I see being made is comparable to the progress I see being made in the West.”
Domestically, Murray acknowledged the UK government’s funding commitments but called for more concentrated investment. “There is a lot of funding. There has been a billion pounds committed,” he said. “But it tends to be that you win lots of smaller projects. Maybe the money is spread around a bit more.”
The real constraint lies not in physics talent but in engineering.
“It’s less about the scientific talent. And more now about the engineering and application talent,” Murray said. “The people we find the hardest to recruit are product people who can make reliable optical systems. It is not so much about the quantum physics anymore.”
Commercial Use and Industry Buy-In
ORCA’s focus on application-first development is already paying off. Through programs like the UK’s Quantum Technology Access Programme (QTAP), ORCA has enabled companies including Vodafone, BAE Systems, and Origami Labs to build and deploy early quantum applications.
“ORCA Computing is setting a new standard for quantum computing productivity,” Murray said. “Businesses can transition from exploration to real-world use cases in just a few months, a speed never seen before in the industry.”
The company’s SDK (Software Development Kit) allows developers without deep quantum knowledge to begin experimenting with use cases in optimization and machine learning. As Murray emphasized, the future of quantum computing isn’t locked away in abstract physics—it’s starting to become a practical engineering problem.
With commercial deployments already underway and PT-3 on the horizon, ORCA Computing is charting a pragmatic course—aimed not at theoretical supremacy, but at tangible performance that companies can use today.