Riverlane turned 10 last week. Ten years ago, quantum computing was a tiny field. You could have fit most of the people working on it into a single lecture hall. Physics, maths, computer science, academia, nearly everyone came from a similar field.
The big question back then was whether useful quantum computers would ever exist at all. I thought they would. And I bet everything on it.
Back then, the science was extraordinary, and the hardware was starting to improve. If that continued, the next question became obvious: what can you actually run on these machines? And that’s the question Riverlane first tried to answer.
The early days: a quantum algorithm company
Today, Riverlane is the world leader in quantum error correction. But ten years ago, it was founded as a quantum algorithm company.
Why? Well, when much of the industry was racing to make hardware more capable, I saw a different opportunity: make the algorithms more efficient.
The premise was simple. On one hand, there’s the capability of the computer: how many operations it can perform before errors overwhelm the calculation. On the other hand, there are the demands of the program: how many operations a useful algorithm actually needs. The two must meet in the middle.
Our first collaborators were two brilliant master’s students, Oscar Higgott and Daochen Wang, working part-time on some of the earliest ideas behind the company. That was Riverlane at the start: small, scrappy, ambitious, and trying to find its feet in a field where the potential felt endless.
Early funding came through grants, consultancy work and a fair amount of personal risk. Some grant funding has a strange rule of its own: you often only receive the money after you’ve already spent it. At one point, I put a credit card down as security just to access one. It wasn’t always comfortable, but it taught me how to build a company from very little.
A pivot to error correction
By 2019, we’d closed our seed round. That felt like a major threshold. We had real capital in the bank, we took a lease on an office, and for the first time Riverlane felt less like an idea held together by sheer will and more like a proper company.
Within months, our thinking evolved. Algorithms mattered. They still do. Better algorithms are essential if we want quantum computers to solve useful problems efficiently. But I realised something bigger: without error correction, quantum computing would not scale.
Qubits are extraordinary, but they’re also fragile. You can’t run a useful computation by relying on a single physical qubit behaving perfectly. You need to combine many physical qubits into reliable logical qubits, then perform logical operations with error rates low enough to sustain a long computation.
So, Riverlane expanded. We stopped thinking as a pure software company and started building the error correction technologies needed to make quantum computers reliable. In practice, that meant moving closer to hardware: becoming something much more like a semiconductor company, building a product that sits next to qubits and decodes errors in real time.
Industry learnings
Clockwise from top left, Riverlane and Rigetti, Riverlane, IQM and Zurich Instruments, one of our senior quantum scientists, Nick Johnson, working on Deltaflow, Riverlane and PASQAL
The most important thing we’ve learned as a company is that useful quantum computing will only be built by specialists working together. We need hardware and software to scale.
What was once a narrow field now includes chip designers, control engineers, software engineers, physicists, mathematicians, product teams, manufacturing specialists, and people who’ve spent their careers building complex classical systems. That breadth is necessary. Quantum computing has become a systems engineering challenge.
We’ve also learned the field rarely moves in a straight line. Ten years ago, it was easy to expect some qubit modalities to hit a wall and fall away. Instead, many very different approaches have continued to improve. That’s made the ecosystem more interesting and made it even more important to build technology that works across hardware platforms.
Partnerships matter for the same reason. No single organisation can build the whole future of quantum computing alone. The hard problems sit across the stack: qubits, control, error correction, compilers, algorithms, cryogenics, fabrication, packaging, software, benchmarking, applications. Progress depends on good interfaces and shared language as much as it depends on individual breakthroughs.
The public conversation has moved on too. Bigger systems still matter, but useful quantum computing will be defined by reliability and logical performance. We need better metrics: ones that tell us whether we’re actually getting closer to machines capable of running valuable computations.
There’s a human lesson here as well. What I’m most proud of isn’t a single technical milestone. It’s watching brilliant people grow at Riverlane, take on hard problems, and build careers around something that genuinely matters. That’s the part I feel most strongly when I look back.
The future for quantum
The next decade of quantum computing will be defined by error-corrected systems.
That means reliable logical qubits, real-time decoding, integration with control systems, better benchmarks, more mature supply chains, and the ability to connect quantum machines into the wider compute stack. It also means procurement, standards, talent and collaboration.
Quantum will develop as part of a much bigger computing landscape, one that includes AI, high-performance computing and specialised classical hardware. The question increasingly isn’t how many qubits you have. It’s where quantum gives you a genuinely new capability and how you make that capability available to the scientists, engineers, companies and governments who need it.
I started Riverlane with a belief that useful quantum computers would exist and a question about what could run on them. Ten years on, that question has sharpened: how do we engineer quantum computers reliable enough to be useful?
The answer: we need better hardware, better algorithms, better error correction, better systems engineering, and the brightest minds in the world working on hard problems.
After ten years, it still feels like the beginning. Riverlane’s purpose is to help make useful quantum computing happen sooner. That was the bet at the start, and it’s still the work ahead.