From faith back to science, if the claims hold up
In the long human effort to transcend the limits of classical computation, Microsoft has announced a milestone that may one day be remembered as a turning point: its Majorana 2 chip sustains quantum states for 20 seconds, a thousand times longer than its predecessor, edging the theoretical closer to the practical. The company, which has pursued an unconventional topological approach for two decades, now claims a commercially viable quantum computer could arrive by 2029. Whether this represents a genuine inflection point or an ambitious projection remains contested, but the questions it raises — about the nature of matter, the boundaries of computation, and the pace of human problem-solving — are ones the world is watching with growing urgency.
- Microsoft's Majorana 2 chip keeps qubits alive for 20 seconds on average — a leap so dramatic the company compares it to a phone battery lasting years instead of a single day.
- The announcement carries enormous stakes: quantum computers, if realized at scale, could unlock solutions to problems like microplastics contamination and fertilizer design that classical machines cannot touch.
- Deep skepticism persists among physicists — the supporting paper has not been peer-reviewed, and Microsoft's topological qubit approach has long been viewed by some as closer to scientific faith than settled science.
- The engineering gap remains staggering: Majorana 2 holds just 12 qubits, while a commercially useful quantum machine would demand millions, leaving the 2029 deadline dependent on breakthroughs not yet achieved.
- Microsoft has invited scrutiny by enrolling in Darpa's final quantum validation stage and sharing sensitive data, signaling confidence — but the broader industry and independent scientists are reserving judgment.
Microsoft announced this week a thousand-fold improvement in quantum computing reliability, centered on a new chip called Majorana 2. Where its predecessor's qubits survived only milliseconds, the new chip sustains them for an average of 20 seconds — a difference the company likens to a phone battery that lasts years rather than a single day. On the strength of this, Microsoft says it will deliver the first commercially useful quantum computer by 2029.
The promise of quantum computing has long outpaced its delivery. Qubits are extraordinarily fragile, collapsing under the slightest disturbance — a temperature shift, a vibration, stray electromagnetic noise. For twenty years, Microsoft has pursued a path most competitors avoided: topological quantum computing, rooted in the theoretical quasi-particle predicted by Italian physicist Ettore Majorana in the 1930s. The first Majorana chip arrived in 2025 to a skeptical reception. The second generation improved partly through a material change — replacing aluminium with lead as a superconductor — a substitution conceived by human scientists, not artificial intelligence.
Yet the obstacles ahead are formidable. Majorana 2 contains only 12 qubits; a practical machine would require millions. The paper accompanying the announcement has not been peer-reviewed, and independent physicists, while acknowledging a genuine advance, are calling for greater transparency. Microsoft has responded by enrolling in Darpa's quantum validation program and opening its data — including commercially sensitive material — to the defense research agency.
The broader race remains wide open. No company has yet built a scalable quantum computer that works reliably at commercial scale, and even Demis Hassabis of Google DeepMind has cautioned that the true limits of classical computing may not yet be understood. Whether Microsoft's 2029 target proves visionary or premature, the Majorana 2 announcement has shifted the terms of a debate that will shape how humanity confronts its most intractable problems.
Microsoft announced this week that it has achieved a dramatic leap in quantum computing reliability—a thousand-fold improvement that the company says positions it to deliver the first commercially useful quantum computer by 2029. The breakthrough centers on a new chip called Majorana 2, where the fundamental building blocks of quantum machines, called qubits, now persist for an average of 20 seconds. That may sound modest until you learn that the previous generation lasted only milliseconds. To grasp the scale of the improvement, Microsoft compares it to the difference between a phone that demands charging every single day and one that needs plugging in only once every few years.
Quantum computing has long promised to solve problems that would take conventional computers decades or centuries—everything from designing better fertilizers to eliminating microplastics from the environment. But qubits are notoriously fragile. They exist in a delicate quantum state that collapses at the slightest disturbance: a shift in temperature, a vibration, even stray electromagnetic noise can introduce errors that cascade through calculations. For two decades, Microsoft has pursued an unconventional path called topological quantum computing, an approach based on exploiting the properties of a theoretical quasi-particle predicted by Italian physicist Ettore Majorana in the 1930s. The company's first Majorana chip arrived in 2025, but it drew skepticism from some physicists who questioned whether the research had crossed from science into something closer to faith.
Zulfi Alam, Microsoft's corporate vice president of quantum, stated the company will have a quantum machine by 2029 capable of solving commercially viable problems. Yet that timeline depends on solving a staggering engineering challenge: the current Majorana 2 chip contains only 12 qubits. A practical quantum computer would require millions. The improvement from Majorana 1 to Majorana 2 came partly from a material substitution—replacing aluminium with lead as a superconductor—a change that human scientists, not artificial intelligence, conceived. Microsoft has also enrolled in Darpa's quantum development program, the final validation stage, and says it has shared all its data and workings with the defense research agency, including commercially sensitive material.
The skepticism, however, persists. The paper Microsoft published alongside its announcement has not undergone peer review, the standard process where independent experts scrutinize research before it enters the scientific record. Paul Stevenson, a physics professor at the University of Surrey, acknowledged that Microsoft appears to have made a genuine leap in topological qubit production. If the claims hold up, he said, the company would jump from having no production quantum computer to becoming a serious contender in the race to build fault-tolerant machines. But other scientists wanted more transparency. Jason Zander, Microsoft's executive vice president of quantum and discovery, pushed back on the criticism, saying the company welcomes scientific debate and inviting skeptics to read the published papers and speak with experts given access to detailed information.
The broader context matters here. Quantum computing remains in its infancy. No company has yet built a scalable quantum computer that works reliably at commercial scale. The entire industry is racing toward that goal, knowing that whoever reaches it first will have solved one of the most consequential technical problems of the era. Yet some voices counsel caution about the timeline. Demis Hassabis, co-founder of Google DeepMind, recently noted that we may not yet understand the true limits of classical computing—the machines we use today. That uncertainty hangs over the quantum race: even if Microsoft delivers on its 2029 promise, it remains unclear whether quantum computers will prove as transformative as the hype suggests, or whether conventional machines will continue to surprise us with their capacity to handle problems we thought were beyond their reach.
Citações Notáveis
We will have a quantum machine in 2029 that can solve commercially viable, reasonable problems— Zulfi Alam, Microsoft corporate vice president of quantum
If they succeed, they will leap from being a player with no production quantum computer, to being a serious player in the race to make the next generation of fault-tolerant machines— Paul Stevenson, physics professor at University of Surrey
A Conversa do Hearth Outra perspectiva sobre a história
Why does it matter that qubits last 20 seconds instead of milliseconds? That still sounds impossibly short.
It's about the ratio of time a qubit survives to the time it takes to perform a calculation. If your qubit collapses before you finish the math, the answer is garbage. Twenty seconds gives you a window to do meaningful work. Milliseconds doesn't.
And Microsoft is confident it can scale from 12 qubits to millions?
That's the honest unknown. They've solved one problem—stability—but scaling is a different beast entirely. Each additional qubit introduces new sources of error. It's like saying you've built a reliable engine; now build one with a million pistons.
Why is the topological approach controversial?
Because it relies on a quasi-particle that was purely theoretical until recently. Some physicists think Microsoft is betting the company on something that might not exist in the way they claim. Others say the skepticism is warranted caution.
What happens if they actually deliver in 2029?
Then problems that would take classical computers 30 years to solve might take months. Drug discovery, materials science, optimization problems that affect billions of people. The economic and scientific implications would be enormous.
But there's still doubt about whether quantum computers are even necessary.
Right. We don't know if classical computers have hit a wall or if they'll keep surprising us. Quantum might be the answer to a question we haven't fully asked yet.