Microsoft Quantum Computing Gains Put Crypto on a Shorter Clock

Microsoft quantum computing gains now span three separate bets: Majorana 2’s topological qubit parity lifetime, toric-code error-correction work from Atom Computing, a neutral-atom hardware company, and electron-on-helium readout work from EeroQ, a Chicago startup. In Microsoft’s Majorana 2 announcement, the company said its new chip reaches a mean qubit lifetime of 20 seconds and moves its scalable-computer target to 2029.

The crypto angle arrives through signatures. Bitcoin’s developer guide for transactions says Bitcoin uses the Elliptic Curve Digital Signature Algorithm (ECDSA, a public-key signature method) with the secp256k1 curve. A future fault-tolerant quantum computer running Shor’s algorithm would attack that class of math. The June results leave that machine in the future and put migration planning back on engineering calendars.

The Twenty-Second Result Sets the Pace

Microsoft’s topological route depends on parity, the evenness or oddness of electrons in a topological wire. The new Majorana 2 technical preprint, submitted to arXiv on June 2, reports a characteristic parity switching time of about 20 seconds in an indium arsenide-lead tetron device, with some instances reaching minute-scale duration.

That timing has to be read next to the operation speed. The same abstract says typical qubit operation times are on the order of microseconds. Microsoft is arguing that the lifetime window is now wide enough for measurement-based control to become an engineering problem, with enough headroom to run operations before parity flips damage the computation.

The table also shows why a single qubit count no longer tells much. Microsoft is changing materials, Atom Computing is trying to keep logical information alive after repeated checks, and EeroQ is trying to read fragile electron states fast enough to make its hardware useful.

Lead Replaced Aluminum in the Qubit Stack

The materials change is simple to describe and hard to execute. Microsoft replaced the aluminum superconductor used in Majorana 1 with lead and updated the semiconductor active region to a combination of indium arsenide and indium arsenide antimonide. Lead is part of the company’s pitch because it creates a larger topological gap, the energy separation that helps protect the state from environmental noise.

Microsoft also tied the device work to Microsoft Discovery, its agentic artificial intelligence platform for scientific research. The company said its quantum team used those tools to manage workflows, automate measurements, optimize fabrication and find flaws in process data. That Build-era push fits a broader Microsoft campaign to own more of its artificial intelligence stack, a theme Oton covered in its report on Microsoft’s homegrown AI models.

• 1,000x reliability gain – Microsoft says Majorana 2’s materials stack gives qubits a 1,000-fold improvement over the prior generation.
• 20-second mean lifetime – the Majorana 2 paper reports about 20 seconds as the characteristic parity switching time.
• Microsecond operations – Microsoft says topological operations run on a microsecond scale, leaving a large gap between operation time and measured parity lifetime.

None of that settles the physics argument around topological qubits. It does give Microsoft a dated, measurable claim: a preprint, a materials stack and a public target year.

Atom Computing Attacks the Error Problem

Atom Computing’s June 3 update comes from a different hardware world. Its systems use neutral atoms trapped and rearranged by lasers, with quantum information stored in atomic states. In Atom Computing’s toric-code announcement, the company said its neutral-atom system reduced errors as larger numbers of qubits were used and became the first neutral-atom system to show that result with toric-code error correction.

Practical quantum error correction can be achieved with our neutral-atom technology.

Ben Bloom, Atom Computing’s founder and chief executive, said that in the company’s June 3 release. The claim builds on prior Microsoft and Atom work. An earlier neutral-atom processor paper reported 24 logical qubits encoded into 48 atoms and a Bernstein-Vazirani algorithm running with up to 28 logical qubits encoded into 112 atoms.

That is why error correction has moved from theory talk into hardware scorekeeping. Oton’s earlier piece on quantum error correction crossing into hardware covered the same shift through Google’s Willow processor. Atom Computing is now pushing the neutral-atom branch of that race with Microsoft standing nearby through Azure Quantum and the Magne system being installed for QuNorth, a Nordic quantum initiative funded by EIFO and the Novo Nordisk Foundation.

EeroQ Works the Measurement Layer

EeroQ is earlier in the machine-building chain. The company uses electrons floating above liquid helium, with the qubit encoded in the electron’s spin or motion. Its bet is that a clean helium surface can isolate electrons from many of the defects found in solid materials while still using chips built with complementary metal-oxide semiconductor (CMOS, the standard manufacturing family behind conventional silicon electronics).

The company’s February 2025 resonator work is useful here because measurement keeps showing up as the bottleneck. In EeroQ’s microwave resonator update, the company said it fabricated titanium nitride resonators for its CMOS-based platform and that the microwave responses were predicted within 2 percent by its calculations. It also said one resonator could read many separate electron qubits because electrons can be moved around its chips.

That belongs beside Microsoft and Atom Computing for a practical reason. Quantum machines need long-lived states, repeated correction and fast readout. A system that gains one layer and stalls on another still stalls. EeroQ’s hardware is a reminder that the race is being fought across the whole stack, from materials to measurement electronics.

Crypto Security Has Standards Ready

The U.S. National Institute of Standards and Technology (NIST, the federal standards lab) already moved before these hardware announcements. In August 2024, NIST’s post-quantum standards finalized Federal Information Processing Standards (FIPS, U.S. government computer-security standards) 203, 204 and 205. The signature standards matter to blockchains because ownership is proved through signatures, not account names.

FIPS 203, 204 and 205 do not give Bitcoin or Ethereum a ready-made switch. Public blockchains have old wallets, smart contracts, exchanges, custody systems, bridges and hardware devices that cannot be upgraded like a cloud service. They also have a governance problem: changing signature schemes can affect address formats, transaction size, wallet recovery, fee markets and backward compatibility.

For protocol teams and wallet makers, the inventory work is specific:

• Exposed public keys – addresses or accounts that reveal public keys on-chain need to be identified, because Shor’s algorithm attacks the public-key math once a powerful fault-tolerant machine exists.
• Signature dependencies – consensus rules, wallet firmware, custody policies, bridges and smart contracts need lists of every elliptic-curve signature assumption they rely on.
• Upgrade paths – projects need migration plans that cover dormant wallets, exchange-held funds and contracts whose owners may no longer be reachable.

Public machines remain far from deriving blockchain private keys from public keys. Migration work still takes years, and the standards now exist before the code-breaking demonstration.

The Public Test Moves to Timelines

Microsoft’s 2029 target is the date that will travel furthest outside physics circles. The company says its timeline for a scalable quantum computer has been cut in half, and it continues to work through the Defense Advanced Research Projects Agency (DARPA, the Pentagon’s research arm) benchmarking process. The Majorana paper is still a preprint. The public record has a measured parity lifetime, not a finished fault-tolerant computer.

Atom Computing has a company-announced toric-code result and prior arXiv work with Microsoft. EeroQ has a readout component that fits the measurement bottleneck. Taken together, the three paths make quantum computing feel less like a single moonshot and more like a set of engineering queues, each with its own failure mode.

For crypto teams, the next dated document is a migration plan.