US researchers say they have built a laser system precise and powerful enough to control vast grids of atoms, a step they argue could ultimately support 100,000-qubit quantum computers. The advance targets one of the hardest engineering problems in the field: how to scale from today’s fragile prototypes to machines large and reliable enough to tackle real-world chemistry, logistics, and security problems.
Instead of adding a few more qubits to existing chips, the team is rethinking the light that holds and manipulates those qubits in the first place. By reshaping laser beams into thousands of tightly focused “tweezers” for individual atoms, they are trying to turn a laboratory curiosity into an architecture that can plausibly stretch to the 100,000-qubit range without collapsing under its own complexity.
Why a new laser could change the qubit numbers game
The promise of this laser technology is not just raw power, it is control. To reach 100,000-qubit machines, engineers need to address and measure each qubit with exquisite accuracy while keeping the entire system stable. The US group behind the new work argues that their design can scale the same basic optical pattern that works for a few hundred atoms up to arrays that, in principle, could host 100,000-qubit processors, a claim that directly targets the field’s most stubborn bottleneck and is framed as a realistic path rather than a distant fantasy in their description of 100,000-qubit hardware.
What makes this credible is the way the laser is structured. Instead of steering a single beam around a chip, the system splits and shapes light into a dense grid of optical traps, each one capable of holding a neutral atom in place as a qubit. The researchers emphasize that the same optical platform that works for a few thousand traps can be extended to far larger grids, because the complexity is pushed into a carefully engineered optical element rather than into thousands of separate control lines. In practical terms, that means the leap from thousands to tens of thousands of qubits becomes an exercise in optical design and power management, not a complete reinvention of the hardware stack.
Neutral-atom arrays and the metasurface leap
The laser breakthrough sits squarely inside the fast-moving world of neutral-atom quantum computing, where individual atoms are trapped and moved using light instead of being etched into solid-state chips. In their latest work, the team reports that they used a metasurface optical tweezer platform to trap atoms into a variety of patterns, demonstrating that the same hardware can generate regular grids, more exotic geometries, and reconfigurable layouts. That flexibility is crucial, because different quantum algorithms and error-correction schemes demand different connectivity patterns, and the group stresses that, For the purposes of realistic scaling, the ability to dial in new atom arrangements without rebuilding the entire optical bench is as important as the raw qubit count.
Neutral-atom arrays have been gaining momentum because they combine long coherence times with the ability to rearrange atoms on the fly. The metasurface approach pushes that trend further by compressing what used to be a table full of lenses and mirrors into a compact, engineered surface that imprints a complex pattern onto an incoming beam. In practice, that means a single chip-like element can create thousands of optical tweezers with fixed relative positions and intensities, dramatically simplifying alignment and stability. The researchers argue that this kind of integration is what turns neutral-atom platforms from delicate physics experiments into something that can be manufactured, replicated, and eventually deployed in data centers.
From lab demo to 100,000-qubit roadmaps
Even if the new laser can support massive atom arrays, the field still needs a roadmap that connects those arrays to useful machines. Large players have already sketched out what that might look like. In a widely discussed vision video, Our long-term planning from IBM describes a “quantum-centric supercomputer” powered by 100000 qubits by 2033, positioning that scale as the threshold where quantum systems begin to tackle the world’s most challenging problems. The new US laser work does not guarantee that timeline, but it directly addresses one of the core engineering assumptions behind it: that there will be a practical way to control that many qubits in a single, coherent device.
In that sense, the laser is less a standalone breakthrough and more a missing piece in a broader ecosystem that stretches from academic labs to corporate roadmaps. If neutral-atom platforms can use metasurface-based lasers to reach tens of thousands of qubits per processor, then the 100000-qubit target becomes a question of how many such processors can be networked together and how effectively error correction can be layered on top. The IBM vision of a quantum-centric supercomputer assumes that classical and quantum resources will be tightly integrated, and a scalable optical control system is one of the few plausible ways to feed that many qubits into such a hybrid architecture without drowning in wiring and cryogenics.
Racing to make the biggest atom arrays
The US laser development also lands in the middle of a quiet arms race over who can build the largest and most controllable neutral-atom arrays. Earlier work highlighted how Columbia Researchers Know to Make the Biggest Arrays Yet, underscoring that the race is not just about qubit quality but about sheer scale. Those Columbia efforts focus on pushing neutral-atom platforms to host far more qubits than today’s superconducting chips, and they rely heavily on sophisticated optical control to keep thousands of atoms pinned in place and individually addressable.
On the West Coast, Caltech has been building what it describes as the world’s largest neutral-atom quantum systems, emphasizing that One of the unique advantages of these platforms is physical reconfigurability, where atoms can be rearranged during a computation using mobile optical traps. That capability pairs naturally with the new metasurface laser approach, which can generate dense initial arrays that are then reshaped in real time. Together, these strands of research suggest a future in which quantum processors are not static chips but living, reconfigurable atom grids, sculpted and steered by programmable light.
The optics industry quietly prepares the ground
Behind the scenes, a specialized optics industry is emerging to support exactly this kind of high-end quantum hardware. Companies that once focused on laboratory instruments are now building integrated systems that can monitor, shape, and stabilize powerful beams with minimal human intervention. One example is a firm that notes that Today it is pioneering instruments that consolidate multiple optical lab devices into a single platform, pushing high-power laser characterization to unprecedented levels. For quantum computing, that kind of consolidation is not a luxury, it is a prerequisite for turning fragile lab setups into robust products.
As quantum teams push toward 100,000-qubit ambitions, they will depend on this broader optical ecosystem to deliver stable, manufacturable components. The new US laser design shows what is possible when metasurfaces, neutral-atom physics, and industrial-grade optics converge around a single goal. If that convergence holds, the field could move from debating whether 100,000-qubit machines are possible to arguing over who gets there first and what problems those machines should tackle once the light finally switches on.
More from Morning Overview