August 1st, 2025. On an otherwise unremarkable Thursday, Fujitsu quietly dropped an announcement that could eventually make your favorite simulation-heavy supercomputers look… quaint. They’ve begun development on a superconducting quantum computer designed to scale past 10,000 physical qubits—a project slated to mature by fiscal year 2030.
Yes, ten thousand. Not a typo. Not theoretical. Not “maybe one day.” It’s on the roadmap.
Opening the Door to Fault-Tolerant Quantum Reality
Let’s break it down: physical qubits are the building blocks—fragile, error-prone, like glass marbles on a windy day. To do anything useful, we need logical qubits—stable, resilient bundles of physical qubits stitched together with quantum error correction.
Fujitsu’s endgame?
By 2030: 250 logical qubits.
By 2035: 1,000 logical qubits.
And beyond that? Well, that’s when it gets spooky-good.
Their secret weapon is called STAR architecture—a design created in partnership with Osaka University. It’s a sort of quantum origami that folds complexity into elegance. Compared to traditional blueprints, it dramatically cuts down the number of physical qubits needed per logical one—much like swapping a twelve-wheeler for a sleek electric bike and still hauling more.
Why This Matters: More Than Just Big Numbers
The real promise isn’t just about quantity—it’s about capability. Imagine simulating a molecular system so complex that even your beefy supercomputer just groans and times out after five years. Fujitsu says their projected 60,000-qubit system (built on STAR) could finish that job in about 10 hours.
From materials science to pharmaceuticals to weather modeling, the kinds of challenges this machine could crack open are the ones that change industries and lives.
So What’s the Hard Part?
Oh, almost everything.
Here’s what Fujitsu’s up against:
- High-precision qubit manufacturing: Getting those tiny Josephson junctions quiet enough to behave like good quantum citizens.
- Chip-to-chip interconnects: Think of it as building an interstate for qubits—low resistance, high speed, and cryogenically stable.
- Thermal packaging: Keeping it cold, compact, and chaos-free.
- Error correction: Developing algorithms that don’t just mop up errors but predict where they’ll strike next.
Every one of these is a PhD thesis on its own. Fujitsu is juggling all of them at once.
The Japan Effect: Funding, Friends & Forward Motion
This isn’t a solo moonshot. Japan’s NEDO program is backing the mission as part of its post‑5G infrastructure strategy. Fujitsu’s working shoulder-to-shoulder with RIKEN and AIST, two research giants with serious quantum credentials.
Some key milestones so far:
- 64-qubit system (2023): The first foothold.
- 256-qubit system (April 2025): Already running, already accessible via cloud to researchers worldwide.
And get this—the 256 system scaled up from the earlier model without a total redesign. That’s a massive architectural win. They modularized it, built “unit cells” of four qubits each, and fine-tuned the thermal packaging. Efficiency with elegance. A very Fujitsu move.
The Diamond Side Quest: Beyond Superconducting Qubits
While superconductors remain the flagship, Fujitsu isn’t putting all its Schrödinger eggs in one basket.
They’re working with QuTech and Delft University on something rather beautiful: diamond spin qubits. These are tiny nitrogen-vacancy centers in synthetic diamonds—qubits you can shine.
What do you think?
Stay Connected
Follow our journey and be part of the conversation:
🔗 Find us on LinkedIn
📬 Join our mailing list
📺 Subscribe to our YouTube channel
