Nvidia vs. Cerebras: Better AI Stock to Buy Now

Cerebras Systems debuted on May 14, rising 68% in the year’s biggest IPO, after raising more than $5.5 billion and reaching a nearly $67 billion market cap. The article argues Nvidia remains the stronger AI stock given its $215 billion recent revenue base, 70%+ gross margins, and roughly 28x forward earnings, while Cerebras offers higher risk but faster growth from $24 million in 2022 to more than $510 million last year. Overall tone is constructive on AI hardware, but it favors waiting for a dip in Cerebras versus buying Nvidia on valuation and scale.

Analysis

The market is still pricing AI as a one-dimensional GPU story, but the more important second-order effect is ecosystem lock-in. If model training and inference keep scaling, the scarce resource is no longer just flops per dollar; it is software compatibility, deployment reliability, and procurement friction. That favors the incumbent with the deepest developer moat and the widest attach rate into networking, software, and systems, while newer challengers may win isolated workloads but struggle to dislodge the default purchasing standard.

Cerebras is interesting less as a direct Nvidia replacement and more as a barometer for how much enterprise buyers are willing to diversify performance-critical workloads away from CUDA-centric infrastructure. The risk is that the IPO enthusiasm compresses the timeline investors are implicitly assigning to share capture: revenue can re-rate quickly on headline wins, but durable margin structure and repeatable gross profit tend to lag by quarters to years. If deployment economics are real, the first beneficiaries may be cloud intermediaries and systems integrators, not necessarily the chip vendor itself.

The contrarian take is that Nvidia’s valuation reset may be more important than the presence of a new competitor. When a category leader trades closer to a growth multiple rather than a scarcity multiple, the market is signaling skepticism about peak margins, but it also creates a better setup for compounding if demand stays broad-based. The real downside risk to Nvidia is not one startup; it is a spend digestion phase in which hyperscalers pause capex to absorb prior orders, which could pressure sentiment for 1-2 quarters even if the secular thesis remains intact.

For Cerebras, the post-IPO move likely front-ran fundamentals, so the trade is now about timing rather than conviction. The stock can work if it converts design wins into recurring usage and proves that its speed advantage survives real-world workload heterogeneity; otherwise, it risks becoming a high-beta trading vehicle tied to AI enthusiasm. In this tape, the highest-probability outcome is continued leadership by the incumbent, with the challenger creating volatility rather than a clean displacement narrative.

Key Decisions for Investors

• Maintain core long NVDA exposure on 3-12 month horizon; use any AI-capex headline selloff to add, targeting a better entry if forward multiple compresses another 10-15%.
• Avoid chasing CBRS after IPO pop; wait for a 20-30% retracement or evidence of repeatable enterprise deployments before initiating a position, as the risk/reward is skewed toward post-deal digestion.
• Relative-value trade: long NVDA / short a basket of high-multiple AI challengers or unprofitable compute names over 6-9 months, betting that software moat and margin durability matter more than point-in-time speed claims.
• Sell out-of-the-money covered calls on NVDA into strength for near-term premium capture; this monetizes elevated AI sentiment while preserving upside if the secular demand cycle continues.
• Set a catalyst watch on hyperscaler capex guidance over the next 1-2 quarters; if spending pauses, trim CBRS exposure first and treat any NVDA weakness as temporary rather than thesis-breaking.