Rigetti Computing

Learning Objectives

• Understand what Rigetti Computing builds and why its chiplet architecture is a distinctive bet on scaling quantum hardware.
• Learn what the Cepheus processor achieved and where Rigetti's roadmap is heading through the end of 2026.
• See how Rigetti connects to AI work through NVIDIA NVQLink and its Algorithmiq partnership, and how you can access its systems today.

What Is Rigetti Computing?

Rigetti Computing is a quantum hardware company founded in 2013 by Chad Rigetti, a former IBM quantum researcher. It is headquartered in Berkeley, California, and trades publicly on the Nasdaq under the ticker RGTI, with a market capitalization that has ranged roughly from $6 billion to $8 billion and tends to be volatile. The company holds a strong balance sheet for its stage: about $571 million in cash and no debt.

Rigetti builds quantum processors using superconducting qubits, the same broad family of technology used by several other leading quantum efforts. What sets Rigetti apart is not the qubit type but how it assembles a large processor. Rather than fabricating one big monolithic chip, Rigetti tiles together several smaller qubit chips into a single working processor. This is the core engineering idea behind the company, and the bet it is making on how quantum machines will eventually scale to useful sizes.

The Chiplet Approach

A central challenge in quantum computing is yield: as you try to fabricate larger and larger single chips, the odds that every qubit on the chip works correctly fall sharply. One bad qubit can compromise a whole processor. Rigetti's answer is the chiplet architecture, which sidesteps this problem by building from smaller, more reliable pieces.

If the approach holds up, it offers a more manufacturable path to growth: improve the chiplets, then tile more of them together. The open question the field has watched closely is whether a processor stitched together this way can hold its performance, especially the fidelity of operations that span the boundaries between chiplets.

Cepheus and the Roadmap

Rigetti's flagship system is Cepheus-1-108Q, referred to as Cepheus. It is a 108-qubit processor built from twelve nine-qubit chiplets, with a median two-qubit gate fidelity of about 99.1 percent. Cepheus became generally available on April 7, 2026, and is notable as the first system with more than 100 qubits to validate the chiplet scaling approach in practice. (An important clarification: the older figure of roughly 99.5 percent fidelity belongs to Rigetti's earlier 36-qubit Ankaa-3 system, not to the 108-qubit Cepheus.)

The roadmap continues from there. Rigetti has targeted a system with more than 150 qubits at about 99.7 percent median two-qubit gate fidelity by the end of 2026, which would represent both more qubits and meaningfully higher operation quality.

The AI Angle

Quantum processors do not replace classical computers; they work alongside them. Rigetti leans into this hybrid model. It integrates NVIDIA NVQLink, a connection layer for hybrid quantum-classical work, and reports roughly a 3-times hybrid speedup from that integration. This lets a quantum processor and classical AI hardware coordinate more tightly on a shared problem.

Rigetti has also moved its hardware toward concrete applications in machine learning. A February 2026 partnership with Algorithmiq applies quantum machine learning to financial fraud detection, an area where pattern recognition over complex, high-dimensional data is valuable and where hybrid approaches are being actively explored.

How You Access It

Rigetti offers commercial access through its own cloud service and through major cloud marketplaces, and it also sells a quantum processing unit you can run on-premises.

Strengths

• Distinctive scaling strategy. The chiplet architecture offers a more manufacturable path to larger processors by building from small, high-yield, individually testable pieces.
• Validated milestone. Cepheus made Rigetti the first to demonstrate chiplet scaling above 100 qubits, moving the approach from theory to a working, generally available system.
• Strong balance sheet. About $571 million in cash and no debt gives the company runway to pursue a multi-year hardware roadmap.
• Multiple access paths. Availability through QCS, Amazon Braket, and Microsoft Azure, plus the on-premises Novera, makes the hardware reachable for many kinds of users.
• Hybrid AI integration. NVIDIA NVQLink support and the Algorithmiq fraud-detection partnership tie the hardware to practical quantum-classical and machine learning workflows.

Limitations & Considerations

• Early-stage technology. Even at 108 qubits and about 99.1 percent two-qubit fidelity, today's systems are far from fault-tolerant, general-purpose quantum computers; practical advantage on real-world problems is still emerging.
• Roadmap risk. Higher targets such as 150-plus qubits at about 99.7 percent fidelity are goals, not guarantees, and quantum hardware timelines often slip.
• Cross-chiplet performance. The chiplet approach must continue to prove that operations spanning chiplet boundaries hold their fidelity as more chiplets are tiled together.
• Market volatility. As a publicly traded quantum company, Rigetti's valuation swings widely, which can complicate planning for organizations evaluating long-term commitments.

Best Use Cases

Getting Started

1. Decide whether you want cloud access or on-premises hardware. For most users, cloud access is the natural starting point.
2. Choose an access path that fits your existing tools: Rigetti Quantum Cloud Services, Amazon Braket, or Microsoft Azure.
3. Create an account on the platform you selected and locate Rigetti's quantum processors in its available devices.
4. Start with small, well-understood quantum programs to learn the system's behavior and current fidelity before scaling up.
5. If you are exploring hybrid workflows, look into the NVIDIA NVQLink integration to coordinate quantum and classical compute.
6. For dedicated, in-house hardware, evaluate the Novera QPU as a purchasable on-premises option.

Key Takeaways

• Rigetti Computing is a public, Berkeley-based quantum hardware company that builds superconducting processors and trades on the Nasdaq as RGTI.
• Its defining bet is the chiplet architecture — assembling one large processor from several small qubit chips to improve manufacturability and scaling.
• Cepheus, a 108-qubit system built from twelve nine-qubit chiplets at about 99.1 percent two-qubit fidelity, became generally available in April 2026 and validated chiplet scaling above 100 qubits; the roadmap targets 150-plus qubits at about 99.7 percent by the end of 2026.
• The AI angle comes from NVIDIA NVQLink hybrid integration (roughly a 3-times hybrid speedup) and an Algorithmiq partnership applying quantum machine learning to fraud detection.
• You can access Rigetti through QCS, Amazon Braket, and Microsoft Azure, or buy the on-premises Novera QPU.