On April 7, 2026, Rigetti Computing announced that its Cepheus-1-108Q quantum processor is now available to customers worldwide through Amazon Braket and its own Quantum Cloud Services platform. This is not a prototype press release. Jobs can be queued on this system today. The Berkeley company has deployed twelve interconnected 9-qubit chiplets into a commercial environment—the largest modular superconducting quantum system by qubit count currently accessible via public cloud. RGTI gained 4.77% on the news. That is the market's way of saying: this is real, it is deployed, and it is not catastrophically disappointing.
The superconducting quantum computing space is now three-tier. At the top: IBM and Google, which have 127-qubit and 99-qubit systems respectively but keep them proprietary or within controlled research partnerships. In the middle: Rigetti, IonQ (trapped ion), and Atom Computing (neutral atom), all fighting to be the first open-cloud platform at scale. At the bottom: the rest of the industry, shipping 20–50 qubit systems or announcing them in PowerPoints. Rigetti's move to 108 qubits on Braket changes the middle-tier calculus. It is now the largest publicly accessible gate-based system on any cloud platform. Amazon's integration matters here: Braket is AWS's way of making quantum computing infrastructure as commodified as compute, storage, and networking. Rigetti just became part of that commodity layer. That positioning is worth something, and the market reflected it. The stock move, though, was modest. Reason: the market knows the real test is not qubit count. It is fidelity, and whether fidelity at 108 qubits can actually run deep circuits.
Cepheus-1-108Q is performing at 99.1% median two-qubit gate fidelity with a gate execution time of roughly 60 nanoseconds and 99.9% single-qubit fidelity. The company also reported that a prototype version achieved 99.9% two-qubit fidelity at 28 nanoseconds using a proprietary adiabatic CZ gate scheme. The architecture innovation here is not trivial: Rigetti moved from 36 to 108 qubits by tripling chiplets while refining the coupler design to mitigate interactions that become pronounced above 100 qubits. The company's Alternating-Bias Assisted Annealing technique targets qubit frequency and reduces fabrication defects. This is engineering maturity, not magic. The speed claim—that Rigetti gates are 1,000–10,000 times faster than trapped-ion and neutral-atom systems—is real and matters for circuit depth. A 60 nanosecond gate versus 10 microseconds means you can execute 166 gates in the time a trapped-ion system executes one. That has implications for decoherence during computation. But here is the constraint: 99.1% fidelity is the entry point, not the destination. Fault-tolerant quantum error correction requires roughly 99.9% logical qubit fidelity using surface codes, which in turn requires 99.5% physical qubit fidelity as a minimum. Rigetti has committed to reaching 99.5% median two-qubit fidelity later in 2026. That is the announcement everyone should be waiting for.
Why now? Three forces: one, capital pressure. IQM, Rigetti's closest structural competitor (full-stack, superconducting, cloud-accessible), just secured a €50 million financing package from BlackRock-managed funds and announced a SPAC merger with Real Asset Acquisition Corp at a $1.8 billion pre-money valuation, closing around June 2026. That valuation will become Rigetti's public-market comparator within weeks. Rigetti is already public at $RGTI; its market cap is lower. The benchmark sets expectations. Two, AWS is moving from curiosity to commitment. Braket started as a research partnership. Braket with Cepheus-1-108Q as its flagship superconducting option is infrastructure. AWS does not deploy infrastructure that does not have clear use cases. The addition of a 100+ qubit system signals that AWS sees quantum workloads (materials simulation, optimization, drug discovery) migrating from academic labs to enterprise pipelines. Three, geopolitics. The UK has committed £2 billion to quantum computing and explicitly flagged that it needs private-sector partners to hit a "TeraQuOp"—one trillion quantum operations—by 2035. Rigetti announced a £100 million UK investment to deploy a 1,000-qubit system within four years, aligned directly with that target. The timing of the 108Q announcement matters because it is Rigetti saying: we can scale. The UK investment is Rigetti saying: we will scale in places where government money is allocated. Both statements are capital signals.
Who wins and who loses? Rigetti wins access. It is now the default superconducting option on Braket, which means it is the entry point for enterprises and governments exploring quantum without committing to on-premises hardware. AWS customers in optimization, materials science, and simulation will queue jobs on Cepheus-1-108Q by default. That is margin. Rigetti's on-premises business—the 9–108 qubit systems sold to research labs and quantum centers—likely accelerates because Cepheus-1-108Q proves the chiplet architecture works at scale. The company manufactures in-house at Fab-1, which means unit economics improve with volume. Trapped-ion systems (IonQ) and neutral-atom systems (Atom Computing, neutral-atom startups) lose the "we have more qubits" argument. They have to argue fidelity, coherence time, and error correction roadmap instead. That is a harder sell because it requires engineering depth, not marketing. IBM and Google do not lose anything—they do not compete for cloud access; they build their own stacks. But they are watching Rigetti's fidelity progress because if Rigetti reaches 99.5% and proves it can run useful circuits, the architecture conversation shifts. Superconducting has been written off by some of the neutral-atom community as too noisy and slow to scale. Rigetti with 108 qubits and a clear path to fault tolerance would invalidate that narrative.
Our read: Rigetti has crossed from announcement-dependent to execution-dependent. The 108Q deployment is real and competitive. The technical progress is incremental but genuine—moving from 36 qubits to 108 by solving coupler interactions is systems work, not overnight invention. The gate speed advantage is material and durable. But Rigetti is now on a three-year timer to prove quantum advantage, and it has just told the world that explicitly. That is not confidence. That is commitment because the market demanded it. IQM's public listing will set a valuation floor, and the market will price Rigetti against whatever IQM's trading multiples are. The 99.5% fidelity milestone later in 2026 is the next inflection point. If Rigetti hits it and proves it can run circuits that matter—drug binding simulations, materials properties, optimization for real enterprise problems—then the thesis holds and fidelity becomes the new race. If the 99.5% slips or the circuits still do not work, then Rigetti is a 108-qubit system with good marketing and a hung fidelity curve. Rigetti's bet is that materials and fabrication work will unlock coherence at scale. That is a testable engineering claim. Watch whether the company actually delivers on it. The announcement is honest. The execution is what matters now.