Overview
This is a solo explainer and interview episode from science communicator Dr Ben Miles, built around Microsoft's announcement of the Majorana 2, a topological quantum computing chip. Miles visited Microsoft's quantum lab in Copenhagen and interviewed Chetan Nayak, Director of Quantum Hardware at Microsoft, to examine what changed technically, whether the results hold up, and what the announcement means for the broader timeline to a practical quantum computer.
Bottom Line
The episode covers genuinely complex physics and does so carefully, building from first principles before addressing the new results. Listeners without a background in quantum computing will need to follow along closely, but the explanations are structured and methodical. It is most valuable for people tracking quantum computing developments who want more than a press release, though it does acknowledge the limits of what Microsoft has publicly demonstrated.
Key Themes
- How topological qubits differ from conventional qubits
- The physics of Majorana zero modes and the tetron architecture
- Why switching from aluminium to lead mattered
- What Majorana 2 improved over Majorana 1
- The remaining engineering gap between today's chip and a useful quantum computer
- Ongoing scientific skepticism about Microsoft's claims
What Was Discussed
The qubit stability problem. Quantum computers require qubits to maintain their state long enough to perform useful calculations. Any interaction with the environment — heat, vibration, radiation — can cause a qubit to collapse, a process called decoherence. Most approaches struggle to keep qubits stable for more than a few hundred milliseconds, even inside heavily shielded dilution refrigerators cooled to near absolute zero.
The topological approach. Microsoft's strategy is to encode quantum information across the global structure of a material rather than at a single point. This is achieved using superconducting semiconductor nanowires cooled to 50 millikelvin, where quantum effects produce so-called Majorana zero modes — quasi-particles that exist simultaneously at both ends of the wire. Two wires are joined in an H-shaped configuration called a tetron, encoding the qubit in a property of the whole system (parity) rather than any local detail. This makes accidental corruption far less likely, because the environment would need to disturb both ends of the wire simultaneously.
Majorana 1 to Majorana 2. Majorana 1 demonstrated that topological qubits could be built, but qubit lifetimes of 1–10 milliseconds allowed only around 10,000 operations — far short of the millions to billions needed for fault-tolerant computation. Microsoft identified two remaining failure modes: quasi-particle poisoning (stray electrons disrupting parity) and Majorana hybridization (a small coupling between the zero modes that causes errors). Both trace back to the superconducting gap — the energy required to break apart Cooper pairs.
The aluminium-to-lead switch. Majorana 1 used aluminium, which has a superconducting gap of around 300 microelectronvolts — low enough that infrared photons emitted by the refrigerator walls could break Cooper pairs. Microsoft replaced the aluminium layer with lead, which has a gap of around 1,300 microelectronvolts, roughly four times higher. The material swap increased qubit lifetime from 12 milliseconds to over 20 seconds, a roughly 1,000-fold improvement that allows approximately 20 million operations before an error is likely.
What remains and what's contested. The remaining challenge is scaling: making thousands of qubits behave this way in an integrated array. Microsoft has revised its timeline to a useful quantum computer to 2029, partly because algorithmic improvements have reduced the estimated qubit count needed to run significant computations (from around 10 million to closer to 600,000 for breaking RSA-2048 encryption, according to Nayak). Miles notes, however, that Microsoft has not yet publicly demonstrated superposition control, gate operations, or algorithm execution on the Majorana 2. The company says this data exists and will be released in stages.
Notable Points
Microsoft's progress outpaces the field's standard benchmark. Schoelkopf's Law — the quantum equivalent of Moore's Law — holds that qubit coherence time roughly doubles each year. A 1,000-fold improvement in 12 months represents approximately a decade's worth of progress by that measure, if the claim holds up.
The lead deposition was technically difficult for reasons beyond materials science. Lead is highly contaminating in fabrication environments and adheres to surfaces where it is not wanted. Getting it to bond correctly at the atomic level without polluting equipment or disturbing the semiconductor beneath took years of failed attempts.
There is unresolved skepticism in the physics community. Microsoft retracted an earlier paper after criticism that its measurements of Majorana zero modes were not sufficiently definitive. Miles notes that similar concerns are being raised about the Majorana 2 announcement, specifically that the evidence published so far does not fully substantiate the claims being made.
Nayak argues that topological qubits have a clearer improvement path than competing approaches. Because the superconducting gap can in principle be increased further through material changes, the team believes they understand which levers to pull — an advantage he suggests other qubit architectures do not have.
Worth Listening If…
- You follow quantum computing and want a technically grounded account of what Microsoft's Majorana 2 announcement actually involves
- You're interested in how materials science and quantum physics intersect in cutting-edge hardware engineering
- You want to understand the gap between a promising lab result and a commercially useful quantum computer
Skip If…
- You're looking for a balanced multi-perspective debate; this is largely built around Microsoft's own framing, with skepticism noted but not explored in depth
- You have no prior exposure to quantum computing concepts — the explanations are clear, but the episode assumes the topic is worth your attention before making that case