Hitachi has reached 99.1% gate fidelity on ordinary silicon, the same material in every smartphone, with no exotic isotopes required. The method is a microwave control scheme that extended qubit stability by a factor of 280. From fiscal 2027, researchers should be able to run experiments on it through the cloud.

The Phase Control Method, Explained

In February 2026, Hitachi announced a new qubit control scheme developed jointly with the Institute of Science Tokyo (Science Tokyo), including a group led by Associate Professor Tetsuo Kodera of Science Tokyo's School of Engineering. The results were published in npj Quantum Information.

The basics: quantum computers run on qubits, the quantum counterpart of the 0s and 1s in an ordinary computer. Unlike ordinary bits, a qubit can be both 0 and 1 at once, a property called superposition, which is what gives it an advantage on certain problems.

The catch is fragility. Electromagnetic noise, vibration, even trace impurities in the material can destroy the quantum state. That is the main obstacle between the lab and a practical machine.

Hitachi's team protects the qubit with continuous microwave irradiation combined with precise phase control, extending the Concatenated Continuous Drive (CCD) technique the company had been developing. Imagine balancing a ball on a platform that is constantly being bumped. Rather than eliminating every bump, this approach keeps the ball moving so the bumps cancel each other out.

The method works in two layers. Continuous microwave exposure puts the qubit into a noise-resistant "dressed state." Modulating the phase of those microwaves over time then produces a more stable "doubly dressed state." The doubling averages out noise, so errors accumulate more slowly, and it works against control-system noise from fluctuations in microwave intensity as well as against the noise from the material itself.

The Numbers: 280x Longer Coherence, 99.1% Accuracy

Coherence time, how long a qubit holds its quantum state, went from 0.14 microseconds to 40.7 microseconds under Ramsey measurement, roughly a 280-fold improvement. That is how much longer the qubit can work before losing its quantum properties.

Single-qubit gate fidelity, how accurately basic operations are performed, improved from 95% to 99.1%. Practical quantum computing is generally held to require fidelity above 99%, so crossing that line matters.

The spin rotation Q-factor, a measure of operational stability, rose from 2.2 to 25.0, confirming that high-quality quantum operations are possible even in noisy conditions.

Why Ordinary Silicon Matters

Hitachi achieved these results using standard silicon, the material every semiconductor factory in the world already handles.

Most quantum computers today use exotic approaches. IBM and Google build systems with superconducting circuits that must be cooled to near absolute zero (about -273°C or -460°F). IonQ uses trapped ions, individual atoms held in place by laser beams. These approaches work, but they require specialized, expensive equipment that's difficult to scale.

Silicon-based quantum computers could in principle be built in the same fabs that produce the chips in your phone. The obstacle is that natural silicon contains trace amounts of silicon-29 (²⁹Si), an isotope with nuclear spin that generates magnetic noise and erodes qubit coherence. The conventional answer, ultra-purified silicon with ²⁹Si stripped out, is expensive and hard to produce at scale. This demonstration ran on a natural Si-MOS qubit instead.

Hitachi took the other route: instead of purifying the material, engineer the control to survive the noise. It is closer to noise-canceling headphones than to soundproofing the building. Solving it in control rather than in materials is what opens the door to scaling on existing semiconductor infrastructure.

The Roadmap: Cloud Access by 2027

Together with Science Tokyo, Japan's RIKEN research institute and the Belgian semiconductor research center imec, Hitachi has laid out a roadmap:

  • Fiscal 2027: A silicon quantum computer prototype accessible via cloud, letting researchers operate qubits remotely
  • Fiscal 2028: A 100-qubit prototype implementing quantum error-correcting codes
  • Fiscal 2030: 1,000 qubits. Fault-tolerant operation is expected to require on the order of one million physical qubits

The cloud launch matters because it puts real hardware in outside hands. The model resembles IBM Quantum, which lets external researchers reach its superconducting systems online. The target users are researchers running experimental workloads, not general-purpose computing.

Japan's Unique Position in the Quantum Race

The quantum computing landscape is dominated by a fierce rivalry between the United States and China, each pouring tens of billions of dollars into the field.

The U.S. signed the National Quantum Initiative Reauthorization Act in January 2026, securing federal funding through 2034. IBM, Google, Microsoft and IonQ are each pursuing different approaches: Google's Willow chip demonstrated a key error-correction milestone with 105 qubits, and IBM's Nighthawk processor moved toward cloud data center integration in 2025.

China has designated quantum technology a national strategic priority in its 2026-2030 Five-Year Plan. It operates a quantum communication network spanning more than 10,000 kilometers, and a team at USTC led by Pan Jianwei has built the Jiuzhang series of photonic quantum computers.

Japan is not trying to outspend the superpowers. Its strategy rests on three things.

First, a bet on silicon. Superconducting qubits dominate the headlines, but the silicon route plugs into the existing semiconductor manufacturing ecosystem, where Japan has decades of depth. The wager is compatibility with industry rather than raw qubit count.

Second, coordination between academia, industry and government. RIKEN acts as the hub connecting Hitachi's engineering, university research and international partners such as imec. The government's Moonshot Research and Development Program is funding quantum work through 2030.

Third, open international collaboration. With U.S.-China tech rivalry deepening, Japan can partner across geopolitical lines. The planned cloud opening is part of that.

What This Means for You

Quantum computers will not replace your laptop. They are aimed at drug discovery, materials science, financial modeling and logistics optimization, problems that scale badly on classical machines.

What this result shows is that the accuracy can come from the control scheme rather than from exotic materials. If reliable qubits can be built from silicon that is already mass-produced, the path to a practical machine gets shorter and cheaper.

The race is not only about who builds the biggest machine first. It is about which approach reaches manufacturing.

How is quantum computing developing in your country? Have you heard about silicon-based approaches before? Share your thoughts in the comments, we'd love to hear your perspective!

Update: On July 22, 2026, Hitachi announced it had been selected, with Intel's Japan arm, as a planned recipient under NEDO's Post-5G Information and Communications Systems Infrastructure Enhancement R&D programme. The project runs to March 2029 and includes joint research with Japan's National Institute of Advanced Industrial Science and Technology (AIST). Qubit chips will be prototyped on Intel's 18A process to assess device-to-device performance variation. Cloud access to the experimental hardware, via AIST's G-QuAT, is targeted for fiscal 2027 and no later than March 2028; the 100-qubit prototype with error-correcting codes is set for fiscal 2028 and the 1,000-qubit platform for fiscal 2030. (As of August 2026)

References