Google’s Santa Barbara lab has led the quantum race for a decade, but a new Colorado team pursuing neutral-atom technology has turned the marathon into a sprint.
For almost ten years, Google’s quantum effort centered on superconducting qubits chilled to near absolute zero in a California lab. In 2019, the company announced its machine had achieved “quantum supremacy,” completing a calculation in under four minutes that would have taken a traditional supercomputer 10,000 years. Now, a separate Google-funded lab in Colorado is betting on a fundamentally different approach: neutral atoms trapped by laser light.
The shift has compressed expert timelines. Many researchers now believe a viable quantum computer could arrive before the end of the decade, according to reporting from DNYUZ and The New York Times. In May, after China vowed to pump more than $15.2 billion into quantum research, the Trump administration directed $2 billion toward American start-ups and chip makers.
Two Architectures, One Funder
Google’s Santa Barbara operation builds qubits from superconducting metals cooled to temperatures nearly 460 degrees below zero. The hardware resembles a “Byzantine chandelier” of silver and gold lattice holding exotic metal chips inside massive dilution refrigerators. This approach has historically looked like the surest path forward; Amazon, IBM, and others have followed the same roadmap, stringing together hundreds of qubits to perform increasingly complex calculations.
The Colorado lab pursues neutral atoms. In a Harvard basement laboratory, lasers, lenses, and mirrors bounce beams into a tiny vacuum chamber. “See that little glow?” said Mikhail Lukin, a physics professor at Harvard, pointing to a microscope image. “That is a cloud of atoms.” Each atom acts as a qubit, manipulated by optical tweezers — focused laser beams that can move and arrange individual atoms with precision.
Hartmut Neven, who leads Google’s quantum efforts, acknowledged the internal competition. “Superconducting will get us there, but I wanted to avoid that last moment when some other technology might pass us by, even if it is only temporary,” he said. “Neutral atoms have that chance.” Still, Neven believes superconducting qubits will prove more effective in the long run.
The Error Correction Threshold
Raw qubit count is not the bottleneck. The central challenge is error correction — keeping quantum states coherent long enough to complete a calculation. As far back as December 2024, Google said its superconducting system had surpassed the “error correction threshold,” meaning that adding more qubits exponentially reduced the error rate rather than increasing it.
“We have finally demonstrated all the elements that we need to get there,” Lukin said. “For the first time, the ideas that people had three or four decades ago are now within our line of sight.” Both architectures claim progress on this front, but neither has yet built a fully error-corrected, logically sound machine at scale.
Encryption Deadline Moves Up
The practical stakes are immediate. A working quantum computer would break current encryption standards, which rely on factoring problems that quantum algorithms can theoretically solve. Earlier this year, Google vowed to overhaul its encryption techniques before 2030.
“In 2022 or 2023, I would have told you that the encryption scare was far out,” Neven said. “The timelines have come down.” The acceleration reflects a broader shift: what was once a theoretical physics problem has become an engineering race with national security implications.
Why Neutral Atoms Gained Ground
Neutral-atom systems offer distinct advantages. Atoms are identical by nature — no fabrication variations plague the qubits. They can be packed densely in two- or three-dimensional arrays and rearranged dynamically during computation. The laser control systems operate at room temperature, even if the vacuum chamber requires cooling.
Superconducting qubits, by contrast, require millikelvin refrigeration and suffer from manufacturing inconsistencies. Each chip is slightly different, introducing calibration overhead. But the superconducting ecosystem benefits from years of process development, established supply chains, and integration with classical control electronics.
The physicist Richard Feynman framed the ultimate goal in a 1981 lecture: “Nature isn’t classical, dammit. If you want to make a simulation of nature, you’d better make it quantum mechanical.” Whether the machine that finally does that runs on superconducting circuits or laser-trapped atoms, the finish line has moved from theoretical horizon to engineering target.