Microsoft Quantum

What it is

Microsoft Majorana is the label for one of the most distinctive strategies in quantum computing, and one of the most contested. In February 2025 the company announced Majorana 1, a topological processor, with a second generation, Majorana 2, following. The premise is that a topological qubit can carry error resistance in its own physics, so that software correction has less work to do. If it can be made to work at scale, it promises a shortcut past the enormous overhead that other approaches spend on catching and fixing mistakes.

The topological idea

The physics is genuinely different from the rest of the field. A topological qubit is meant to store quantum information in a way that is spread out and hard to disturb, using exotic states associated with Majorana particles, so that small local disturbances do not corrupt the data. That is the appeal of the Microsoft Majorana approach: build stability into the hardware, and rely less on the layers of correction that dominate superconducting and trapped-ion designs. The difficulty is equally real. The technology remains at prototype stage, and Microsoft has not yet demonstrated error correction at scale on it. Majorana 1 is a statement of direction more than a finished machine.

A second, safer route

Microsoft is not waiting on that research alone. It has partnered with Atom Computing on Magne, a neutral-atom system targeting 50 logical qubits for deployment in 2027, giving it a second, more conventional route to useful hardware. The company was also selected under the US2QC Phase 3 programme in February 2025, alongside PsiQuantum, and has been chosen by DARPA for advanced quantum benchmarking. Those selections read as external validation that the topological bet, however unproven, is taken seriously by the agencies funding the field. Government money in quantum computing tends to follow approaches judged to have a plausible path to a genuine advantage, which is part of why the topological work continues to be funded despite the absence of a proven device.

The Azure reseller model

Around all of this sits a commercial layer that is easy to overlook. Azure Quantum, Microsoft’s cloud service, resells hardware from IonQ, Quantinuum and Atom Computing. So while the Microsoft Majorana programme researches a substrate that may not pay off for years, the company already sells access to other people’s quantum computers today. That is a shrewd position: revenue and customer relationships now, with the option of a leap ahead later if the topological work comes good.

Claim, proof and what to watch

The distinction between claim and proof is important here. That topological qubits exist and can be controlled well enough to build a processor is the central assumption of the Microsoft Majorana story, and it has drawn scrutiny precisely because the underlying physics is so hard to verify. A qubit that resists errors by its nature would be a considerable prize, which is exactly why the evidence needs to be strong. For now the honest summary is that Majorana 1 and Majorana 2 are early prototypes on an unusual and high-risk path, without demonstrated error correction at scale.

Where the Microsoft Majorana effort sits in the wider field is as the boldest of the substrate bets. Superconducting and trapped-ion machines are further along and better understood, yet they carry heavy correction overheads. If the topological approach delivers, it could change the economics of the whole sector by making each qubit far more reliable from the outset. If it does not, Microsoft still has Azure Quantum reselling working hardware and a neutral-atom partnership in progress. Readers tracking the competing designs can follow them through our quantum explainer hub; the result to watch here is whether a future Majorana processor ever shows error correction working at scale, the single demonstration that would move the topological idea from promise to proof.