Cerebras CEO says margin forecast was 'misunderstood' as stock plummets after earnings

Cerebras Systems fell 17% after its first post-IPO results, as investors focused on management's warning that core gross margin should be 38% to 41% for the full year versus 47% in Q1. CEO Andrew Feldman argued the guidance was misunderstood and said the company is still beating the plan, while analysts at Mizuho and Wedbush raised estimates after the call. The stock also faces upcoming insider lock-up expirations on about 28 million Class A shares and execution risk around renting back equipment and expanding data centers.

Analysis

The market is reacting less to the headline margin level than to the path dependency embedded in the business model. A company that is still digesting customer-related economics, asset redeployment, and infrastructure scaling is effectively telling investors that reported gross margin is an output of financing structure and utilization, not just chip economics. That creates a valuation problem: if the market had been underwriting a clean hardware gross margin ramp, the de-rating can persist until investors see a few quarters of stable utilization and no further “one-off” normalization.

Second-order, the pressure on adjacent AI infrastructure names is mixed. On one hand, the disclosure reinforces that AI compute remains capital intensive and operationally messy, which can weigh on speculative infrastructure multiples broadly. On the other hand, the specific issue here looks idiosyncratic rather than reflective of end-demand weakness, so the cleaner beneficiaries are incumbents with balance-sheet scale and supply chain leverage, not necessarily the niche challenger. For NVDA and TSM, the read-through is not demand destruction but that customers are still willing to absorb very large AI capex even when economics are opaque.

The overhang that matters most over the next 2-8 weeks is lock-up supply, not the margin print. A staggered unlock into a newly public, high-volatility name with uncertain fundamental normalization is a classic recipe for pressure even if the operating story is intact. If insider selling is muted, the stock can stabilize quickly; if there is meaningful distribution, the multiple can compress another leg before any true fundamental bottom is established.

The contrarian view is that the selloff may be overstating near-term downside because the market is conflating adjusted margin mechanics with terminal profitability. If management can demonstrate that the core gross margin range was always a planning assumption and that utilization improves as rented equipment is redeployed, the narrative can flip from “margin miss” to “guided conservatism.” That would likely matter more than the absolute first-quarter figure, because this is a credibility and cadence trade, not a secular demand trade.

Key Decisions for Investors

• Short the bounce in the post-IPO tape: initiate a tactical short in the name on strength over the next 1-3 trading sessions, targeting a 5-10% retracement if lock-up-related supply hits. Risk/reward is favorable because the next catalyst is supply, not a near-term upside revision.
• Use put spreads instead of outright short exposure: buy 1-2 month downside put spreads to capture another leg lower from insider distribution while capping carry cost. Best suited if implied volatility remains elevated but not fully pricing the unlock overhang.
• Pair trade: long NVDA / short Cerebras on a 1-2 month horizon. This isolates the difference between category leadership with supply-backed demand visibility and a smaller platform still proving margin durability; the trade benefits if AI capex stays strong but capital rotates away from lower-quality exposure.
• For TSM, avoid adding on this headline; the read-through is neutral-to-slightly negative for foundry sentiment if investors start questioning whether AI supply growth can outrun customer monetization. Wait for a better entry after the unlock window or on any broader semis selloff.
• If you want contrarian exposure, size a small long only after the first post-unlock trading window closes and only if management shows no material insider selling. The upside case is a rapid reset in expectations, but it needs evidence of stable core margins before the market re-rates the stock.