NEC is reported to have ended, as of late March 2026, the development of its own quantum computing machines, according to an investigation published by Japanese outlet Diamond Online. The decision is particularly notable because NEC researchers were behind one of the papers that helped open the way for today’s superconducting qubits, back in 1999. The company has not formally confirmed a withdrawal from quantum computing and says it will continue exploring practical applications and industrialization of these technologies.
Key points on NEC’s exit from quantum hardware in 20 seconds
- NEC reportedly halted development of its own quantum machines in March 2026.
- The company pioneered superconducting qubits with a paper published in Nature in 1999.
- Some of the program’s researchers reportedly moved to Fujitsu.
- NEC will keep working on applications, proofs of concept, and possible industrial uses.
- Fujitsu keeps investing in superconducting hardware and has just unveiled a prototype based on diamond spins.
The distinction matters. The available information doesn’t support the claim that NEC is abandoning quantum computing altogether. What it appears to be giving up is the most expensive and riskiest part of this race: physically designing and building quantum computers.
NEC can continue developing software, algorithms, use cases, and client projects, in addition to using hardware built by third parties.
Diamond Online also points to a particularly notable move within Japan’s tech industry: numerous researchers linked to NEC’s project reportedly moved to Fujitsu, including staff who played a prominent role in the program.
The company hasn’t publicly detailed the scope of those moves.
NEC helped kick off the superconducting qubit race in 1999
The history turns the decision into something more than the cancellation of another experimental project.
On April 29, 1999, Yasunobu Nakamura, Yu. A. Pashkin, and Jaw-Shen Tsai published a paper in Nature titled Coherent control of macroscopic quantum states in a single-Cooper-pair box.
Nakamura and Tsai were at NEC’s fundamental research labs at the time.
The experiment demonstrated coherent control of macroscopic quantum states in a small superconducting structure based on a Cooper-pair box.
That device could behave as a two-level quantum system and, therefore, as a qubit, the basic unit of information in a quantum computer.
The paper became an early reference point for superconducting quantum computing.
The historical significance is easy to appreciate from 2026. Far more advanced variations of superconducting circuits are currently one of the leading technologies used to build quantum processors.
In the years that followed, NEC kept working in this direction.
The company researched both gate-based quantum computing and quantum annealing, an approach especially geared toward certain optimization problems.
In 2022 it announced a cell based on parametron-type superconducting qubits for a quantum annealing architecture. Two years later, it took part alongside Japan’s National Institute of Advanced Industrial Science and Technology (AIST), Yokohama National University, and Tohoku University in demonstrating a superconducting circuit designed to control multiple qubits.
NEC kept publicly presenting these advances as steps toward larger-scale quantum systems.
That’s why the reported shutdown of hardware development barely two years later marks a significant shift.
Building a commercial quantum computer is still a very expensive bet
Diamond Online attributes the withdrawal to an issue that’s becoming increasingly clear around quantum computing: the time it takes to turn scientific breakthroughs into a profitable business.
Manufacturing a quantum processor is only part of the problem.
Superconducting systems need to operate at temperatures extremely close to absolute zero. That requires dilution refrigerators, specialized electronics, control systems, calibration, and considerable infrastructure surrounding the processor itself.
Then comes the harder problem: errors.
Qubits are extremely sensitive to noise and easily lose their quantum state. Building computers capable of running useful calculations over extended periods requires improving fidelity and developing quantum error correction.
That introduces an important distinction between physical qubits and logical qubits.
A fault-tolerant quantum computer may need numerous physical qubits to create a single, sufficiently reliable logical qubit. That’s why a machine announced with thousands of physical qubits doesn’t necessarily equate to thousands of qubits usable for complex algorithms.
The race demands heavy investment over years without a market yet comparable to classical computing or artificial intelligence.
NEC appears to have decided that taking on that hardware investment directly no longer fits its priorities.
But giving up manufacturing doesn’t mean giving up the future commercial opportunity either.
If quantum computing reaches meaningful enterprise applications, NEC could offer integration, software, consulting, or sector-specific solutions using other manufacturers’ computers.
It’s a strategy similar to the one followed in other areas of infrastructure: you don’t need to build the processor to build services around it.
Fujitsu advances while gaining talent from NEC
The possible transfer of researchers is especially interesting because Fujitsu is heading in the opposite direction.
In April 2025, Fujitsu and RIKEN unveiled a 256-qubit superconducting quantum computer, quadrupling the 64 qubits of the system the two organizations had previously presented.
The company also launched a program to develop larger-scale superconducting systems and announced in August 2025 the formal development of a computer exceeding 10,000 physical qubits, aiming to complete it during fiscal year 2030.
Fujitsu intends to combine that hardware with error-correction technologies and its hybrid quantum computing platform.
But the Japanese company isn’t betting solely on superconductors, either.
On that same day, September 8, 2026, Fujitsu announced a prototype quantum computer based on diamond spins, developed together with Delft University of Technology and QuTech.
The system uses tin-vacancy (SnV) centers integrated with photonic circuits.
One feature of this approach is the ability to optically connect different quantum modules, which could make larger-scale modular architectures easier to build.
Fujitsu says the prototype can operate at -271.6°C, compared with roughly -273.13°C in conventional superconducting systems.
The company plans to develop a multi-module prototype of this technology in 2027 and to also explore its integration with its superconducting architecture.
Its new roadmap targets 250 logical qubits during fiscal year 2030 and 1,000 logical qubits by 2035.
The arrival of researchers from NEC, if confirmed along the lines reported by Diamond Online, would add experience to an organization that’s simultaneously expanding several quantum hardware tracks.
Japan is concentrating its quantum bet among fewer players
NEC’s decision can also be read within the broader reorganization Japan’s tech industry is going through.
For decades, large conglomerates like NEC, Fujitsu, Hitachi, and Toshiba kept huge in-house research departments and manufactured an extremely wide range of technologies.
NEC knows well what it means to exit capital-intensive businesses.
The company led global semiconductor sales worldwide back in 1990. Over the following decade it lost ground as Intel grew around the PC microprocessor market and Asia’s industry transformed the memory business.
NEC gradually spun off its semiconductor activities.
In 1999 it combined its DRAM business with Hitachi’s to form NEC-Hitachi Memory, later renamed Elpida Memory. In 2002 it spun off most of its remaining chip operations into NEC Electronics.
That company eventually merged with Renesas Technology to form Renesas Electronics.
The comparison with quantum computing has to be handled carefully, since these are very different markets and technological moments.
But there’s a common thread: developing frontier hardware demands growing amounts of capital and long payback periods.
NEC currently focuses much of its activity on IT services, enterprise systems, telecommunications, artificial intelligence, security, aerospace, and defense.
Keeping a proprietary quantum processor program going for possibly another decade means competing for resources with those areas.
NEC isn’t disappearing from quantum computing
NEC’s own response forces a caveat on the “withdrawal” headline.
Faced with Diamond Online’s questions, the company avoided directly confirming that it had abandoned quantum computer development.
NEC said it continues evaluating practical applications and the industrialization of quantum technologies, in addition to running proofs of concept with clients.
That leaves room for a considerable presence in the sector.
Quantum computing covers far more than manufacturing qubits: algorithms, compilers, optimization, integration with supercomputers, cryptography, simulation, enterprise software, and hybrid platforms are all areas where a tech company can work without owning its own processor.
There’s also the option of consuming quantum capacity through external services.
The shift, then, appears to be about where NEC places the risk.
For more than two decades it tried to take a direct part in building the machines.
Now it seems willing to wait for other manufacturers to solve part of the physical problems, while keeping the door open to using the technology once commercially interesting applications exist.
The move is also a reminder that the quantum race won’t necessarily unfold like generative AI.
AI models have quickly found products capable of generating revenue while the infrastructure keeps growing. In quantum computing, there’s still a considerable gap between demonstrating experimental improvements and having fault-tolerant machines able to economically justify their enormous complexity.
NEC was one of the companies that helped kick off that race in 1999.
Twenty-seven years later, it appears to have decided it no longer needs to build the computer to remain part of it.
Frequently Asked Questions
Has NEC completely abandoned quantum computing?
No. The reported information points to NEC abandoning the development of its own quantum hardware. NEC says it will continue evaluating practical applications, industrialization, and proof-of-concept projects related to quantum technologies.
What did NEC achieve in quantum computing in 1999?
NEC researchers demonstrated coherent control of quantum states in a superconducting Cooper-pair box. The paper, published in Nature, became one of the early milestones in the development of superconducting qubits.
Have NEC’s researchers moved to Fujitsu?
Diamond Online reports that numerous researchers linked to NEC’s quantum development have joined Fujitsu. No complete list of affected employees has been published.
What is Fujitsu developing in quantum computing?
Fujitsu is running superconducting hardware projects and has just unveiled a prototype based on diamond spins. Its roadmap targets superconducting systems with more than 10,000 physical qubits and later goals expressed in logical qubits.
Fujitsu’s diamond-spin work builds on its existing partnership with QuTech: see our earlier coverage of Fujitsu and QuTech achieving high-precision quantum gates with less than 0.1% error.
via: diamond.jp