Cathie Wood Just Bet $46 Million on a Newly Public AI Chip Stock. Should You Follow Her?

Cerebras Systems raised more than $6 billion in what was described as the biggest U.S. tech IPO of the year, with shares debuting at $350 and closing at $311.07 after pricing at $185. The company reported 2025 revenue growth of 76% to $510 million and a swing to $238 million in net profit, helped by a multi-year OpenAI compute deal worth more than $20 billion and an AWS term sheet. However, the stock trades at about 120-125 times sales and roughly 700 times earnings, leaving valuation risk high despite strong AI momentum.

Analysis

The key second-order issue is not whether the technology is real, but whether the business is transitioning from a scarce-chip vendor into a capital-intensive compute utility just as sentiment is peaking. That shift typically compresses returns on capital because revenue recognition becomes gated by datacenter buildout, energy procurement, and customer uptime obligations, all of which introduce longer working-capital cycles and more execution failure points than a pure hardware model.

This also changes the competitive set. If Cerebras gains traction as an AI cloud layer, the pressure lands less on Nvidia’s silicon moat and more on hyperscalers’ pricing power and integration advantage, especially AWS. Amazon benefits if external demand validates Bedrock as a distribution channel for differentiated inference capacity, while Intel remains the obvious relative loser because the market will read any win by non-Nvidia architectures as further evidence that legacy accelerators are not the only path to scale.

The near-term stock dynamic is likely driven more by positioning than fundamentals. A post-IPO float with high narrative intensity can stay disconnected for weeks, but once initial lockup or underwriter support fades, the market usually reverts to a cash-flow lens; at >100x sales, even minor slippage in customer concentration, export-policy friction, or capex intensity can drive a multiple reset of 30-50% without any change in the core technology story. The most important catalyst is not another customer announcement, but evidence that revenue is broadening beyond a small number of state-linked or frontier-AI buyers.

The consensus is missing that the OpenAI relationship is simultaneously a validation event and a margin-risk event. Co-design and hosted capacity may deepen stickiness, but they also give a single customer unusual influence over roadmap, pricing, and deployment cadence, which tends to lower long-term negotiating leverage for the supplier. In other words, the market is pricing a platform monopoly when the more likely intermediate outcome is a premium niche infrastructure provider with lumpy demand and heavy reinvestment needs.

Key Decisions for Investors

• Avoid initiating a long in CBRS for the next 4-8 weeks; if forced to express the view, use a bearish call spread or short-dated put spread to target a 25-35% valuation compression on any post-IPO de-risking.
• Pair trade: long AMZN / short CBRS over 1-3 months. AWS gets paid to distribute third-party AI infrastructure while bearing less single-name execution risk; CBRS has higher downside if growth must be financed via capex-heavy service revenues.
• Maintain or add modest NVDA exposure on weakness rather than chasing CBRS. A valid alternative-architecture narrative does not break Nvidia’s positioning; it more likely expands total AI spend while leaving NVDA as the primary liquidity beneficiary.
• Tactically underweight INTC versus AMZN/NVDA for the next quarter. Any market re-rating of non-Nvidia accelerators should still accrue to platform and ecosystem leaders first, not legacy x86 incumbents.
• Set a catalyst watch for the next earnings cycle and any disclosure on customer concentration. If top-customer dependency remains above ~50% of revenue, treat any rally as fadeable and consider shorts on strength.