Pomegra
CoreWeave Q2 revenue hit $2.58B, up 112% YoY, on GPU cluster demand from AI labs; Cerebras trades at $34.85 as its wafer-chip IPO gain consolidates.
- CoreWeave Q2 2026 revenue rose 112% year-over-year to $2.58B; full-year guidance raised to $12.4B–$13.2B.
- Contracted backlog reached $104 billion, up 246% YoY, locking in multi-year enterprise and AI-lab compute commitments.
- Cerebras Systems trades near $34.85 after surging from a $25 IPO price to nearly $50, as investors recalibrate the wafer-chip AI compute thesis.
Lead
CoreWeave (CRWV) delivered second-quarter 2026 results that exceeded consensus estimates, posting revenue of $2.58 billion on 112% year-over-year growth driven by sustained spending from large-language-model developers and enterprise customers on its coreweave gpu cluster infrastructure. Management simultaneously raised full-year revenue guidance to $12.4 billion–$13.2 billion, up from a prior range of $10.9 billion–$11.7 billion. In a parallel development, Cerebras Systems (CBRS) — maker of the cerebras wafer chip — is trading near $34.85 following a volatile debut that carried shares from their $25 offering price to a peak approaching $50 before broader AI compute valuations pulled back.What Happened — CoreWeave
CoreWeave's second-quarter adjusted EBITDA reached $1.5 billion, a 59% margin that more than doubled year-over-year, reflecting the operating leverage embedded in its hyperscale GPU cluster model. Management raised full-year capital expenditure guidance to $35 billion–$39 billion from $31 billion–$35 billion previously, signaling accelerating commitments to new data-center capacity.
The contracted revenue backlog — a forward measure of locked-in demand — climbed 246% year-over-year to $104 billion. More than $25 billion in net-new commitments were recorded in the opening weeks of the third quarter alone. The company targets 1.85 gigawatts of active power capacity by end-2026 and more than 8 gigawatts by 2030.
Net losses remain elevated at $626 million for the quarter against a net debt position of approximately $32.9 billion, a leverage ratio of 7.4 times EBITDA that reflects the capital intensity of scaling GPU cluster infrastructure at hyperscale speed. GAAP profitability is not expected before 2028.
What Happened — Cerebras
Cerebras Systems completed its initial public offering earlier this year at $25 per share, with shares rallying sharply toward $50 as investors sought direct exposure to the AI silicon market outside of Nvidia (NVDA). The stock has since retraced to the $34–$36 band as the broader AI infrastructure trade digests elevated valuations.
The company's Wafer Scale Engine integrates an entire silicon wafer as a single processor, positioning Cerebras as an architectural challenger to GPU-based compute for inference-intensive workloads. A multi-year compute agreement with OpenAI and a cloud partnership with Amazon Web Services (AMZN) provide meaningful revenue visibility and validate the commercial case for the cerebras wafer chip beyond early adopters.
Market Reaction
CRWV shares moved higher in after-hours trading following the earnings release. The raised guidance and backlog expansion outweighed continued scrutiny of the company's leverage position. CBRS has held the $34–$36 range on elevated volume as institutional investors weigh whether the post-IPO correction represents a structural repricing or a consolidation before the next earnings catalyst.
Strategic Context
Both companies represent distinct expressions of the same macro thesis: the AI compute buildout is advancing faster and lasting longer than prior technology infrastructure cycles. CoreWeave operates on the infrastructure layer, monetizing GPU cluster density through long-term leases to AI labs and cloud-native enterprises. Cerebras operates at the silicon layer, betting that wafer-scale architecture can deliver the performance-per-watt profile demanded by the next generation of large-model inference.
The divergence in their models — CoreWeave as a capital-intensive GPU cloud, Cerebras as a chip-first company moving into AI cloud services — illustrates how the market is stratifying between hardware-as-a-service providers and silicon-level challengers to established GPU incumbents.
Outlook
CoreWeave's raised guidance and $104 billion backlog materially reduce near-term demand uncertainty, though the company's debt load and accelerating capital program require sustained pricing discipline and long-term customer retention. Cerebras must demonstrate that its wafer-scale approach can scale commercially beyond anchor partnerships; a sustained hold above $34 would signal that institutional investors are treating the IPO correction as an entry point rather than a warning sign. Together, the two companies serve as a real-time indicator of how aggressively enterprise and AI-lab customers are committing to next-generation compute infrastructure heading into 2027.
