# Why Did NEC Stop Building a Quantum Computer?

NEC halted development of its quantum computer on September 5, 2026, concluding that the commercialization timeline is too long to generate an acceptable return on investment. The decision ends an internal hardware program that stretched back to 1999, when NEC became the first company in the world to demonstrate qubit operation using a superconducting solid-state device. That milestone — achieved by Yasunobu Nakamura, who later became director of RIKEN's Center for Quantum Computing — gave NEC an early lead in superconducting quantum hardware that the company ultimately could not convert into a commercial product.

The exit covers both of NEC's active hardware tracks: its gate-based superconducting effort under Japan's Moonshot program (which targets [fault-tolerant quantum computing](https://quantumintel.tech/glossary/fault-tolerant-quantum-computing) by 2050) and a distinctive quantum-annealing architecture based on Josephson Parametric Oscillator (JPO) circuits. The JPO annealer, which offered longer [coherence time](https://quantumintel.tech/glossary/coherence-time) than conventional annealers and required fewer qubits to represent equivalent Ising-model problems, never progressed beyond an 8-qubit cloud-access research system launched with Tohoku University in June 2023. The target had been to scale that system to 100 or more qubits before offering practical services — a milestone NEC will not reach.

The announcement arrives as Japan simultaneously plans to send 30,000 young scientists abroad to study AI and quantum research, a national-level commitment that stands in direct contrast to this private-sector retreat.

---

## A Pioneer Exits at a Difficult Moment for Annealing

NEC's hardware strategy carried a structural vulnerability that the source material identifies plainly: the broader industry has moved away from pure annealing machines. [D-Wave Systems](https://quantumintel.tech/companies/d-wave-systems) dominates that segment with an established installed base and is itself now building gate-based machines aimed at [logical qubit](https://quantumintel.tech/glossary/logical-qubit) operation and fault tolerance. Competing against D-Wave's entrenched customer relationships with an 8-qubit prototype was always a difficult commercial proposition.

The JPO architecture had genuine technical merit on paper. By using superconducting parametron circuits, NEC's annealer offered:

- Longer coherence times than flux-qubit-based conventional annealers
- More compact Ising-model representations, reducing the raw qubit count needed for a given problem
- A design philosophy oriented toward scalability rather than brute-force qubit increases

But the engineering obstacles proved formidable. Scaling superconducting annealing systems requires wiring and measurement hardware that grows with qubit count, and the cooling capacity of [dilution refrigerator](https://quantumintel.tech/glossary/dilution-refrigerator) systems operating near 10 mK imposes hard physical constraints. NEC was investigating compact control electronics capable of operating inside the cryostat — a legitimate research direction — but that work was not sufficiently advanced to unlock a path to 100+ qubits on a commercially viable timeline.

On the gate-based side, NEC served as project manager for a superconducting-circuit effort under Japan's Moonshot program. The program's 2050 fault-tolerance target is, by any measure, a long investment horizon. For a technology group weighing capital allocation against nearer-term product revenues, that timeline is difficult to defend internally.

---

## The RIKEN–Fujitsu Thread NEC Could Not Sever

The historical lineage here matters for understanding the competitive dynamics. Nakamura's 1999 superconducting charge qubit demonstration at NEC is the foundational experiment for essentially all subsequent superconducting quantum computing work. When Nakamura moved to RIKEN, that expertise migrated with him. Fujitsu's superconducting processors are built at the RIKEN RQC–Fujitsu Collaboration Center — the institutional successor to NEC's original research environment.

This means NEC was, in effect, competing against the institutional offspring of its own pioneering research. Fujitsu has articulated public ambitions to become an AI and quantum competitor at the national-champion level, and its machines benefit from the RIKEN ecosystem that NEC's own researchers helped build. That is not a comfortable competitive position for any corporate R&D program to occupy.

The source material notes that the quantum news landscape produced very little in the way of NEC product releases or technical announcements in recent years — a signal that the program was stagnating well before the formal halt.

---

## What This Signals for Japan's Quantum Strategy

Japan's national quantum posture is now bifurcated in a way that will draw scrutiny. The government is committing resources to send 30,000 scientists abroad for AI and quantum training. The Moonshot program's 2050 fault-tolerance goal remains formally active. Yet one of Japan's most historically significant quantum hardware companies has concluded that the private-sector economics do not work on the relevant timescales.

This is not a uniquely Japanese problem. Corporate quantum hardware exits have become more frequent globally as the gap between [NISQ](https://quantumintel.tech/glossary/nisq)-era capabilities and commercially useful fault-tolerant systems remains stubbornly wide. The difference in Japan's case is the symbolic weight: NEC's 1999 demonstration is in every quantum computing textbook. Its exit is a data point that the industry's serious institutional investors will not ignore.

The contrast with U.S. and Chinese investment — where both government and private capital continue to flow at scale into superconducting, trapped-ion, neutral atom, and photonic platforms — is stark. Japan's quantum ecosystem is not collapsing, but it is concentrating. RIKEN and Fujitsu carry more of the national load now.

For enterprise buyers and quantum software developers, NEC's exit has limited near-term operational impact — the 8-qubit JPO system was a research platform, not a production service. The strategic signal, however, is clear: even companies with a 27-year head start in superconducting quantum hardware can conclude that the commercialization math does not close.

---

## Key Takeaways

- **NEC formally halted quantum computer development on September 5, 2026**, citing an ROI timeline too long to justify continued investment.
- **The exit covers both tracks**: the gate-based Moonshot program and the JPO-based quantum annealing system, which stalled at 8 qubits in a Tohoku University cloud-access partnership launched in June 2023.
- **NEC's 1999 superconducting qubit demonstration** — the world's first with a solid-state device — established the lineage now carried forward by RIKEN and Fujitsu, not NEC itself.
- **The annealing market's maturation** around D-Wave's installed base left NEC's JPO architecture without a clear commercial entry point, particularly as D-Wave itself pivots toward gate-based fault-tolerant systems.
- **Japan's national quantum strategy is now more concentrated**: government ambitions (30,000 scientists abroad, Moonshot 2050) continue, but private-sector hardware development is consolidating.
- **Dilution refrigerator scaling constraints and cryogenic control electronics** were identified as the principal unsolved engineering obstacles before the program was cancelled.

---

## Frequently Asked Questions

**Why did NEC stop quantum computer development?**
NEC concluded on September 5, 2026 that the time required to commercialize quantum hardware is too long to generate an acceptable return on investment. The company had been pursuing both a gate-based superconducting approach under Japan's Moonshot program and a JPO-based quantum annealing system, neither of which had reached a commercially viable scale.

**What was NEC's quantum annealing technology?**
NEC developed a quantum annealer based on superconducting parametron and Josephson Parametric Oscillator (JPO) circuits. The architecture offered longer coherence times than conventional annealers and required fewer qubits for equivalent Ising-model problems. The most advanced system publicly demonstrated was an 8-qubit machine used for cloud-access research with Tohoku University, launched in June 2023.

**What is NEC's historical significance in quantum computing?**
NEC became the first company in the world to demonstrate qubit operation with a superconducting solid-state device in 1999. That experiment, conducted by Yasunobu Nakamura, is foundational to the superconducting quantum computing field. Nakamura later became director of RIKEN's Center for Quantum Computing, and the research lineage now runs through Fujitsu's machines built at the RIKEN RQC–Fujitsu Collaboration Center.

**How does NEC's exit affect Japan's quantum computing program?**
Japan's government-level quantum ambitions remain intact — including a plan to send 30,000 young scientists abroad to study AI and quantum research and the Moonshot program targeting fault-tolerant quantum computing by 2050. NEC's exit concentrates Japan's private-sector quantum hardware activity further around the RIKEN–Fujitsu axis.

**Does NEC's decision affect quantum annealing as a field?**
It reinforces the competitive difficulty of the pure-annealing segment. [D-Wave Systems](https://quantumintel.tech/companies/d-wave-systems) holds an established installed base in quantum annealing and is now building gate-based machines targeting logical qubits and fault tolerance. NEC's JPO annealer, which never moved beyond a small-scale research prototype, could not establish a foothold before the program was cancelled.