Cerebras Systems
Cerebras IPO'd May 14 (raised $5.55B); its first full public quarter — Q2 2026 (Aug 12) — showed cloud/inference revenue up ~287% YoY and core revenue doubling to $210M, but the market sold the stock down 12-17% on a weak Q3 guide and a confirmed customer-concentration problem: $25.4B of remaining performance obligations, the majority from a single $20B OpenAI master agreement (in which OpenAI also holds equity — a related-party wrinkle). The GAAP net loss of $450.5M is mostly non-cash IPO-vesting stock comp, so it overstates operating deterioration; the real signal is that concentration rotated from UAE (G42/MBZUAI, 86% of 2025 revenue) toward OpenAI rather than diversifying away. Wafer-Scale Engine 3 architecture remains genuinely differentiated. Ticker CBRS on Nasdaq.
Scenarios
$5.55B IPO capital provides runway to land hyperscaler contracts beyond UAE. WSE-3's latency advantage for inference is architecturally validated; market embraced the story at $95B close market cap.
86% revenue concentration in two UAE-linked related-party entities is the same risk that derailed the original 2024 IPO attempt. Any geopolitical disruption to UAE AI investment or G42/MBZUAI budget cuts would be immediately existential. Nvidia CUDA moat still dominates training workloads.
Key Factors to Watch
- ●IPO May 14: $185/share, $5.55B raised, close at $311 (+68%), ~$95B closing market cap; ticker CBRS
- ●2025 revenue $510M (+76% YoY), net income $237.8M — first profitable year
- ●G42 (24%) + MBZUAI (62%) = 86% of 2025 revenue; related parties per prospectus — no real diversification
- ●MBZUAI alone = 77.9% of accounts receivable as of Dec 31, 2025 — credit concentration risk
- ●WSE-3 architecture unique; $10B OpenAI contract provides non-UAE anchor for future growth
Score History
| Date | Score | Direction | Note |
|---|---|---|---|
| 2026-08-16 | 7.5 | Positive | Score 7.8→7.55 (dr 4→5). First full public quarter (Q2 2026, Aug 12) confirmed the pre-IPO bear case with a hard number: $25.4B of remaining performance obligations, the majority from one $20B OpenAI master agreement (OpenAI also an equity holder — related-party risk). Cloud revenue +287% YoY, but the market sold a beat-adjacent print down 12-17% on concentration plus a weak Q3 guide. Concentration rotated from UAE to OpenAI, it did not diversify; dr up 1. Held positive. |
| 2026-05-16 | 7.8 | Positive | W20 IPO — score 8.0→7.8 (md 8→7). Prospectus confirms G42+MBZUAI=86% revenue (related parties), unchanged from 85% G42-only in 2024. Market cap ~$95B; ticker CBRS on Nasdaq. |
| 2026-03-08 | 8.0 | Very Positive | Score 8.1→8.0 (formula reweight: sb 0.25→0.15, are 0.20→0.25, md 0.20→0.25, dr 0.20→0.25, aam 0.15→0.10) |
| 2026-03-08 | 8.1 | Very Positive | Score 7.9→8.1 (md 7→8). External research cross-ref: $10B OpenAI deal + unique wafer-scale architecture validate moat durability |
| 2026-03-08 | 7.9 | Positive | Initial assessment from batch 8 research |
Software Peers
Last researched: 2026-08-16
This is research and analysis, not financial advice. Scores reflect AI impact potential, not investment recommendations.