Latest / Quickly Quantum / September 7, 2026 — NEC Walks Away From Quantum Hardware
Transcript
- Twenty-seven years. That's how long NEC has been building quantum computer hardware — and it turns out the company that started it all quietly walked away back in March, with the news only surfacing this week. Today on Quickly Quantum: NEC halts development of its quantum computer, and the question that actually matters — if a twenty-seven-year pioneer quits building the machines right as governments pour billions into hardware bets, what does that say about honest timelines for this field? Before that, in the headlines: a photonics company just posted the highest single-photon flux ever reported, and the G7 is telling the world, again, to get moving on quantum-proof encryption. Welcome back to Quickly Quantum, your daily brief on the quantum frontier. It's Monday, September 7, 2026. Let's get into it. First up — a genuine speed record. Sparrow Quantum, working with Ruhr University Bochum, built a source that fires single photons — the individual particles of light some quantum computers and quantum networks use to carry information — at one gigahertz, delivering more than 500 million usable photons per second into a single optical fiber. That's the highest single-photon flux reported to date. And here's why the number matters beyond bragging rights: photonic quantum computing has always been a numbers game where losses multiply — a source good enough for one photon can be hopeless for ten. This one keeps more than half its light usable in fiber without any filtering tricks to make the numbers look better, and the company says it's enough supply to make experiments with ten to twenty photons practical rather than aspirational. That's the company's own characterization — Sparrow Quantum put this out as a press release, and independent replication hasn't happened yet. Still, if the fiber-efficiency number holds up under outside scrutiny, that's a real bottleneck loosening, not just a marketing photo op. Now, sticking with the theme of urgency versus reality — the G7's cybersecurity working group and the U.S. agency CISA published a joint advisory on September 3rd, telling organizations to start migrating to post-quantum cryptography now, not later. The core worry is something called harvest-now-decrypt-later — adversaries are already scooping up encrypted data today, banking on decrypting it once a quantum computer capable of breaking today's encryption actually exists. This follows a 2024 G7 warning aimed specifically at the financial sector, so the alarm itself isn't new — what's new is coverage catching up to a document that's already four days old by the time you're hearing this. And here's the tension that's about to become the whole show: that same government urgency sits right next to a company that just quietly gave up on building the hardware the threat depends on. Our main story today: the twenty-seven-year exit. NEC — the company whose researchers demonstrated the world's first superconducting solid-state qubit back in 1999, the basic building block used in many quantum computers today — has discontinued development of an actual physical quantum computing system. Nikkei Asia reports the shutdown happened at the end of March, with the news only becoming public through the outlet's report this week, and the reasoning is blunt: NEC apparently concluded that bringing the technology to market would take too long to produce an adequate return on investment. Now, before you file this as NEC quitting quantum entirely — it isn't. The company is continuing quantum annealing, an approach tuned for optimization problems like delivery routes or investment portfolios, and it's expanding quantum-inspired computing, which mimics some quantum problem-solving methods on regular computers you can deploy today without any exotic hardware at all. What's actually ending is the harder bet — building an actual quantum machine. And the record NEC is walking away from is a real one. Alongside Japan's AIST, the company built an eight-qubit superconducting annealing prototype in 2023 with Tohoku University, using an architecture from Austria's ParityQC designed to cut down on noise sensitivity as the machine scaled up. It was billed as Japan's first domestically produced quantum annealing machine that outside researchers could access over the internet. But it stayed exactly what it sounds like — an early research device, not a commercially useful computer. Practical systems would need far more qubits, plus real advances in control, reliability, and manufacturing NEC apparently decided weren't worth chasing anymore. Nikkei Asia's own framing of the news, posted on X, cuts right to it: quote, "Japan's NEC scraps effort to build a quantum computer. Tech group sees development taking too long to earn return on investment," end quote. That's the whole calculus in two sentences — not a scientific failure, a financial one. Now here's what makes the timing land differently than it might have a year ago. This is happening the same week the G7 is telling the entire world to urgently prepare for cryptographically relevant quantum computers — machines powerful enough to break today's encryption. And it's happening while Illinois, PsiQuantum, IBM, and others are sinking billions into hardware campuses betting the opposite direction. NEC's exit narrows the field of large Japanese and multinational players actually building quantum hardware — meanwhile Fujitsu just announced in August a partnership with Monash University and Australia's science agency CSIRO to keep pushing its own hardware research forward. So this isn't the industry agreeing on anything. It's one major player's ROI math landing on a very different answer than its neighbors. So what do we actually make of this? The honest read is that NEC's exit is narrower than the headline makes it sound — this is a retreat from gate-based hardware, the kind IBM and Google are racing on, not a full quantum exit. NEC keeps its annealing work and its quantum-inspired services running on conventional computers, and its annealing bet was always a smaller wager than the superconducting gate-model race everyone else is running. One company's return-on-investment calculus doesn't automatically forecast the whole field's fault-tolerant timeline. But I don't think you get to wave this off entirely either. This is the same company that gave the field its first superconducting qubit demonstration in 1999 — twenty-seven years of institutional expertise, and the people closest to the physics looked at their own roadmap and decided the payoff was too far out to keep funding. That's not nothing. And it lands the exact week governments are telling banks and critical infrastructure to move now because the threat is closer than comfortable. You can't have it both ways forever — either the hardware timeline is close enough to justify billion-dollar national security warnings, or it's far enough out that a twenty-seven-year player correctly decided to walk. Both of those can't be fully true at once, and right now the industry is being asked to believe both. The open question worth sitting with: is NEC idiosyncratic — its own risk appetite, its own budget cycle — or a leading indicator that other early corporate movers quietly follow over the next couple of years? Fujitsu, IBM, PsiQuantum and the U.S. government are all still writing checks like the answer is close. If a second or third major name makes NEC's call within the next year, that's a very different story than one company's spreadsheet. Time for the Hype Check. I'm putting this one at a 6. The facts here are solid — a corroborated retreat from real, documented hardware, reported by Nikkei Asia and confirmed by NEC's own history. But the framing risk is real too: it's easy to read "pioneer quits quantum" as a verdict on the whole field when it's really one company's math, on one narrower slice of the technology, at one point in its funding cycle. Real news, real caveats, and a genuinely open question about whether it's a signal or a data point. If NEC's calculation is right, the billions still flowing into Illinois and PsiQuantum's campuses are betting against a company that actually built the first qubit — and somebody's going to be proven wrong on the timeline within the next few years. Follow Quickly Quantum wherever you listen so you catch it when we find out which side that turns out to be. This has been Quickly Quantum, an AI-voiced podcast, created and built by a real human using today's cutting-edge technology. Nothing you heard on this show is financial advice. I'm Brian Lampert, and I'll catch you all tomorrow — take care!