Cerebras Stock Is Up 11% Since Its IPO. Is It a Buy?

Cerebras’ AI-chip IPO at $185 on May 14 opened at $350 but now trades around $205, leaving late buyers about 11% underwater versus the offer price. The company cites a $25B backlog and a multi-year $20B OpenAI deal to deploy 750MW of wafer-scale inference systems, with core revenue up 76% to $510M in 2025 and guided to $855–$865M in 2026 (+68%–70%). Despite strong growth potential (projected 2028 revenue of $7.32B and EBITDA expected to turn positive in 2027–2028), gross margins are shrinking as it builds first-party data center capacity, keeping profitability a key overhang and likely sustaining stock volatility.

Analysis

CBRS is a classic "story vs. unit economics" setup: the market is paying for addressable demand, but the real driver of rerating will be whether the company can turn contracted backlog into durable gross profit without leaning on customer-funded capacity. The margin compression from renting back compute is the key tell — it suggests growth is being bought with capex intensity, which usually means the stock trades like a long-duration asset until the market sees operating leverage.

Competitive dynamics are more nuanced than a simple Nvidia-disruption trade. NVDA remains the default beneficiary of broad AI spend because it owns the software ecosystem and can serve both training and inference; CBRS has to prove it is a carve-out winner, not a niche alternative. AMZN is a second-order winner if AWS distribution lowers customer acquisition costs for CBRS, but the economic uplift to Amazon is likely modest unless it can monetize the traffic with higher-margin cloud attach.

The contrarian risk is that backlog is being confused with visibility. In infrastructure-heavy AI, signed demand can be revised, delayed, or repriced if power, deployment, or model economics shift; that makes 1-3 month sentiment fragile even if the 12-18 month thesis survives. Falsifiers are straightforward: sustained gross-margin stabilization, evidence that revenue converts without outsized capex, and management reaffirming growth without pushing profitability further out.

Key Decisions for Investors

• Stay on the sidelines for now; CBRS is a hold-not-buy until the next earnings print confirms backlog conversion and stops the gross-margin bleed. If margins do not inflect over the next 1-2 quarters, the stock should de-rate despite top-line growth.
• Fade strength in CBRS with a 3-6 month put spread or starter short on any rally back above the recent post-IPO trading range; target 15-25% downside if the market refocuses on capex intensity and unprofitability. Stop out if management delivers a clean margin bridge or material guide raise.
• Relative-value: long NVDA / short CBRS in a small size as a quality-vs.-speculation pair for the next 1-3 months. The trade works if AI spend remains robust but investors rotate toward ecosystem winners with proven cash generation; it breaks if CBRS shows durable first-party margin expansion.
• Watch AMZN as a secondary beneficiary, but do not over-underwrite it yet. Only add exposure if AWS commentary shows CBRS-driven demand is incremental and not dilutive to cloud margins; otherwise the catalyst is too small to matter.