Cerebras Systems vs. Rigetti Computing: Is an AI or Quantum Computing Stock the Better Buy in 2026?

Cerebras generated approximately $510 million of FY2025 revenue, up 76% year over year, and nearly $238 million of net income, versus Rigetti's $7.1 million of revenue, down 34%, and a $216 million net loss. Cerebras is favored on its AI-inference exposure, despite negative $393 million free cash flow, a negative 0.5x debt-to-equity ratio, and elevated 222x forward P/E and 71.2x P/S multiples. Analyst estimates project Cerebras sales rising to about $888 million in FY2026, $3 billion in 2027, and $7.5 billion in 2028, while Rigetti remains a more speculative quantum-computing bet despite a potential federal award worth up to $100 million over three years.

Analysis

CBRS is being valued as a share-taker in inference rather than as a hardware vendor, leaving little tolerance for the operational realities of a proprietary wafer-scale architecture: manufacturing yield, customer deployment cadence, and service/support costs. The gap between reported earnings and deeply negative free cash flow is the key diligence issue; unless working-capital releases, customer prepayments, and capex commitments normalize, equity value will be increasingly contingent on external financing despite apparent liquidity. Over the next 1-3 months, investor focus should shift from top-line estimates to backlog quality, customer concentration, gross-margin progression, and cash conversion.

The more consequential competitive threat to CBRS is not simply NVDA's current product cycle, but hyperscalers' incentive to internalize inference through AWS Trainium/Inferentia, Google TPUs, and Microsoft-developed silicon. These platforms can bundle compute, networking, storage, and credits, making a standalone accelerator's performance advantage insufficient unless it demonstrates materially lower total cost per token at production scale. That dynamic favors AMZN, GOOG, and MSFT strategically even if CBRS wins selected high-performance workloads; it also caps the duration of any scarcity premium assigned to specialist AI hardware.

RGTI should be treated as a long-duration funding-option, not a revenue multiple story. Government support may extend its runway and validate its technical roadmap, but equity participation and future capital needs create meaningful dilution risk before commercial quantum demand is measurable; IBM and GOOG can sustain much longer development cycles from existing cash flows. Consensus is likely underestimating that the next positive catalyst is technical—error-correction or fidelity milestones—rather than sales, while also overestimating the immediate monetization of cloud partnerships.

Key Decisions for Investors

• Do not initiate a directional CBRS long solely on forward revenue estimates. Establish a diligence alert around the next earnings release: consider a tactical long only if operating cash burn materially narrows while gross margin and disclosed backlog convert; invalidate the setup if cash use remains elevated or guidance relies on a small number of counterparties.
• Express AI-infrastructure exposure through a pair: long NVDA or AMZN versus short CBRS in equal beta-adjusted dollars over a 3-6 month horizon. The trade captures CBRS execution/multiple risk against incumbents' ability to monetize AI demand; stop out if CBRS demonstrates sustained cash conversion and independently verified broad customer diversification.
• Avoid RGTI as a core long at current speculative valuation. For a high-risk sleeve, use a small, defined-risk call structure dated 9-15 months only around announced processor/error-correction milestones; the thesis is falsified by roadmap slippage, equity issuance on unfavorable terms, or a failure to secure follow-on government funding.
• Monitor AMZN, GOOG, and MSFT capex commentary and custom-silicon utilization over the next two quarters. Evidence that internal accelerators are being deployed beyond captive workloads would be a negative read-through for CBRS's addressable inference margin pool, even if aggregate AI compute spending remains strong.