You've probably heard the hype. Quantum computing is always "five years away," a carrot dangled by researchers and venture capitalists to keep the funding flowing. But if you’ve been watching the Berkeley-based pioneer Rigetti Computing lately, that timeline just got a bit more specific—and a little more complicated.
Honestly, it's a weird time for the company.
Just a few days ago, on January 13, 2026, Rigetti dropped a bit of a bombshell. They’re pushing back the general availability of their flagship 108-qubit system, known as Cepheus-1-108Q. It was supposed to be out by now. Instead, we’re looking at the end of Q1 2026.
Why? Because quantum physics is hard. Really hard.
The Reality of Rigetti Quantum Computing Progress Right Now
Most people think progress in this field is just about "more qubits." It’s not. It’s about how well those qubits behave. Rigetti’s CEO, Dr. Subodh Kulkarni, basically admitted that while the 108-qubit system is hitting a median two-qubit gate fidelity of 99%, that’s just not good enough for what they want to achieve. They are aiming for 99.5%.
That 0.5% difference might sound like nitpicking. It isn't. In the quantum world, that’s the difference between a machine that can actually solve a chemistry problem and one that just spits out expensive noise.
Breaking Down the Current Hardware Stack
If you look under the hood of what Rigetti has deployed as of early 2026, you see a company obsessed with "chiplets."
📖 Related: Global Science and Technology NZ: Why Kiwi Innovation is Punching Way Above Its Weight
- Novera QPU: This is their 9-qubit workhorse. It’s what they sell to labs like Fermilab and the Air Force Research Lab (AFRL). It’s small, but it’s stable, hitting 99.7% fidelity.
- Ankaa-3: This 84-qubit system was the big news of late 2024 and 2025. It achieved that magic 99.5% fidelity mark using a specialized gate called fSim.
- Cepheus-1-36Q: This was a massive milestone in 2025. It’s a 36-qubit machine built by tiling four 9-qubit chiplets together. It proved that Rigetti’s modular architecture actually works.
This modular approach is Rigetti's whole "thing." Instead of trying to build one giant, impossible-to-manufacture chip, they make small, high-quality ones and link them together. Think of it like LEGO sets for supercomputers.
Why the 108-Qubit Delay is a Double-Edged Sword
There's no sugarcoating it: delays suck. Especially when you're a public company (NASDAQ: RGTI) that is burning through cash. Rigetti reported an operating loss of about $20.5 million in just the third quarter of 2025. They have the money—roughly $600 million in the bank as of November 2025—but the clock is ticking.
But here’s the nuance most people miss.
Rigetti is choosing to delay because they are moving away from "hero experiments"—one-off results that look good in a press release but can't be replicated—and toward actual reliability. The delay is specifically to fix "tunable coupler complexities." These are the components that let qubits talk to each other. If those couplers aren't perfect, the whole 108-qubit array becomes a mess of errors.
💡 You might also like: The Mark 1 British tank and why it was a beautiful, mechanical disaster
The NVIDIA Factor
One reason to stay optimistic about rigetti quantum computing progress is their deep integration with the classical world. In late 2025, they announced support for NVIDIA NVQLink.
This is huge.
Basically, it allows an NVIDIA GPU supercomputer to talk to a Rigetti QPU with incredibly low latency. If we ever want to see "quantum advantage"—where a quantum computer beats a classical one at a useful task—it’s going to happen through this kind of hybrid setup. You use the GPU for the heavy lifting and the QPU for the specific, weird quantum math.
What's Next on the Roadmap?
If they hit the new Q1 2026 target for the 108-qubit machine, the rest of the roadmap looks aggressive.
- Late 2026: A 150+ qubit system with 99.7% fidelity.
- End of 2027: The "holy grail" of 1,000+ qubits.
Is it realistic? Maybe. The jump from 100 to 1,000 qubits is a vertical cliff. But Rigetti is betting that their proprietary "Alternating-Bias Assisted Annealing" (ABAA) technique—which helps them precisely target qubit frequencies during manufacturing—will allow them to scale without the yields falling off a cliff.
Actionable Insights for the Quantum-Curious
If you’re tracking this space, don't just look at the stock price or the qubit count. Those are vanity metrics.
Watch the Gate Fidelity: If Rigetti can’t push that 108-qubit system past 99.5% fidelity by March 2026, they’re in trouble. That is the number that determines if the machine is a tool or a toy.
Keep an eye on On-Premise Sales: Rigetti is unique because they actually sell hardware. They recently secured $5.7 million in orders for two Novera systems for an Asian tech manufacturer and a California startup. This "on-prem" model is a massive differentiator from competitors like IonQ or IBM, who mostly stay in the cloud.
📖 Related: Why the Home Theater System DVD Setup Still Dominates Real Living Rooms
Monitor the NVIDIA Partnership: The success of hybrid algorithms using NVQLink will be the first real-world proof of value. Look for white papers or benchmarks involving "Quantum Machine Learning" (QML) in finance or materials science.
The next three months are sort of a make-or-break window for Rigetti. They’ve moved the goalposts, but if they actually kick the ball through this time, the "quantum winter" might finally be over.
Next Steps for Tracking Progress:
- Check the Rigetti Investor Relations page in late March 2026 for the "General Availability" announcement of Cepheus-1-108Q.
- Monitor the reauthorization of the National Quantum Initiative Act in the U.S. Congress, as this will dictate how much federal R&D money flows into Rigetti’s partnerships with labs like Oak Ridge and AFRL.
- Audit the 2025 Q4 earnings report (usually released in March) to see if the "non-cash charges" that led to that scary $201 million net loss in 2024 have finally stabilized.