By Jason Nelson
7 min read
When scientists at Caltech flicked on their new neutral-atom quantum array in September, the quantum machine broke a threshold many scientists thought was years away. For the first time, researchers successfully trapped 6,100 atomic qubits in a single system and maintained coherence in a way that pushed quantum hardware past the “toy demo” stage.
What happened in that lab meant large-scale, error-corrected quantum hardware is no longer a distant aspiration but a credible possibility. And for digital currencies like Bitcoin, whose security depends on cryptography assumed safe for decades, it signals that the quietly accelerating threat posed by quantum computers is now edging into view.
The threat is not imminent—but the window for adaptation is finite. That’s why, at Emerge, we consider quantum computing’s advance—and crypto’s lack of readiness—our Tech Trend of the Year.
“We can now see a pathway to large error-corrected quantum computers. The building blocks are in place,” principal investigator Manuel Endres said in a statement.
For years, the standard comfort for cryptographers has been that quantum computers remained too noisy, too fragile, and too immature to matter to crypto. In 2025, that stance weakened. Roadmaps tightened. Error-correction improved. And several labs produced results that made fault-tolerant machines feel like a question of when, not if.
So-called “neutral-atom systems” use electrically neutral atoms as qubits, trapping single atoms in fixed positions with lasers so each one can store and manipulate quantum information. “Coherence” measures how long those qubits remain in a usable quantum state before noise destroys it. Both became central in 2025 as the field shifted from lab demonstrations to architectures designed to scale.
Understanding the gains of 2025 requires understanding what has held quantum systems back. Qubits (quantum bits) lose their quantum state easily, and scaling them often amplifies that instability. This year, several systems behaved differently.
Google, IBM, and Caltech each reported advances in 2025 that narrowed the timeline for fault-tolerant quantum machines. Google’s 105-qubit Willow processor showed steep error-rate reductions as it scaled, and in October, the company said its Quantum Echoes benchmark ran roughly 13,000 times faster than leading supercomputers. The results indicated that stable logical qubits might be achievable with far fewer physical qubits than the thousand-to-one ratios long assumed.
IBM advanced the picture from another angle. Its “Cat” family processors demonstrated 120-qubit entanglement and extended coherence, and its Starling roadmap, released in June, targeted 200 error-corrected qubits by 2029 with support for 100 million quantum gates. A separate effort with AMD showed that standard FPGA hardware could run error-correction logic ten times faster than required, bringing real-time correction closer to practical use.
Caltech added scale in September through what researchers described as the world’s largest neutral-atom system, trapping 6,100 cesium atoms as qubits, demonstrating coherence for 13 seconds with 99.98% operational accuracy. Together, the results pointed to a broader shift: qubit quality, control, and scaling efficiency improved at the same time, tightening expectations for when usable logical qubits—and with them credible threats to Bitcoin’s signature scheme—could arrive.
Erik Garcell, director of quantum enterprise development at Classiq, said the more consequential shift is the changing ratio between physical and logical qubits. “It’s trending toward a few hundred to one,” he told Decrypt, a sharp improvement from earlier estimates requiring thousands. “Much of the industry’s attention in 2025 shifted toward error correction.”
Qubits collapse under environmental interference, limiting how long they can remain coherent. That’s where error correction comes in. Error correction works by duplicating a qubit’s state across many physical qubits, giving the system enough redundancy to spot when noise knocks one off course and automatically correct it. Without it, qubits fall apart too quickly to do meaningful computation.
Across the field, researchers said the same thing: the machines aren’t just growing; they’re behaving.
While Bitcoin isn’t threatened by the machines that exist today, what changed in 2025 was the tone of the conversation about tomorrow.
Jameson Lopp, who co-founded Casa in 2018 to provide tools that allow people to store and protect their own Bitcoin, said the risk remains far away.
“Whether or not the network can be ready in time ultimately comes down to how quickly advancements happen in quantum computing,” Lopp told Decrypt. “We are orders of magnitude away from having a cryptographically relevant quantum computer. There need to be multiple major breakthroughs before it’s really a threat to Bitcoin.”
Even so, Bitcoin must contend with a constraint that other blockchains like Ethereum or Zcash don’t: coordination. Migrating to a quantum-safe signature scheme would require simultaneous movement from miners, wallet developers, exchanges, and millions of users.
“I really don’t see that whole process happening in less than a five-year time frame,” Lopp said. “Once you have millions and millions of individual actors, asking them to coordinate to make a change becomes effectively impossible.”
Quantum risk is often imagined as a sudden moment when the machines become dangerous. Researchers say the reality will look more gradual.
Ethan Heilman, a research fellow at MIT’s Digital Currency Initiative and co-author of Bitcoin’s BIP-360 post-quantum proposal, said improvements accumulate over time. “We’ll see gradients as it gets stronger and stronger,” he told Decrypt.
He works from a long horizon. Bitcoin is already being treated as a multigenerational asset by many of its users. “If people treat Bitcoin as a savings account—something they can lock away for a century and expect their children to recover—then the protocol should be built to withstand that timeline,” he said.
Heilman expects Bitcoin to adapt. But he noted that markets react to stagnation earlier than they react to risk. “The degree to which Bitcoin does not address that threat could cause downward pressure on the price,” he said.
The field, he said, cares less about dates than about the direction of progress.
“We’ll see steady progress, but going from a coal-powered train to the Concorde in a year seems very unlikely to me,” he said. “I think it will happen, but I think that we will see stages.”
Alex Shih, head of product at Q-CTRL, said quantum risk becomes meaningful only once machines can run large, error-corrected algorithms.
“If there is a large enough quantum computer resource, yes, in theory, it could break today’s RSA encryption,” he told Decrypt. “But getting to that point is still years away. Optimistically, maybe the mid-2030s.”
Early fault-tolerant machines won’t immediately endanger existing cryptography. They will broaden the kinds of algorithms quantum computers can realistically attempt as reliability improves.
Shih pointed to fragmentation as a challenge slowing the field. “Interoperability is still a major point of friction,” he said. “Every vendor releases different specifications and frameworks, and it is left to the end user to make everything work together.”
Even with those hurdles, 2025 clarified momentum. IBM hit its roadmap milestones. Google’s scaling behavior matched expectations. Caltech delivered stability at a size the field had never reached.
Together, these results gave researchers a clearer sense of how the next decade may unfold.
Quantum computing didn’t threaten Bitcoin this year, but it removed ambiguity.
Researchers spoke with more confidence about timelines. Developers in other industries began adjusting long-term plans. Bitcoin’s ecosystem—which rarely revisits its cryptographic foundations without outside pressure—approached the discussion with new seriousness in 2025.
By the end of the year, the debate wasn’t about whether quantum would matter. It was about when its impact became unavoidable.
Decrypt-a-cookie
This website or its third-party tools use cookies. Cookie policy By clicking the accept button, you agree to the use of cookies.