Wall Street's Favorite New AI Stock vs. the One They're Ignoring: The Numbers Aren't Even Close

Cerebras jumped 68% on its first day of trading and is now about 60% above its $185 IPO price, but the article argues the rally is ahead of fundamentals. The company reported just $500 million in 2025 sales and a nearly $76 million non-GAAP loss, versus Microsoft's nearly $83 billion in quarterly revenue, 21% adjusted EPS growth, and $32 billion in cash. The piece frames Cerebras as a speculative AI upstart while favoring Microsoft as the stronger long-term AI beneficiary.

Analysis

The market is rewarding narrative optionality over realized economics: Cerebras is trading like a venture-style call option on AI compute scarcity, while Microsoft is being priced like a mature platform despite still compounding high-teens revenue and mid-20s earnings growth. That gap creates a second-order read-through: capital is rotating toward “scarcity” hardware exposure and away from the picks-and-shovels incumbents that already monetize demand, which can widen valuation dispersion across the AI stack even if the end-demand picture stays intact.

The real issue for Cerebras is not product differentiation, but adoption friction. A wafer-scale architecture can win benchmarks and win pilot commitments, yet the path from technical superiority to repeatable gross margin expansion is constrained by software compatibility, customer concentration, and manufacturing yield risk. If the OpenAI relationship is more financing signal than durable volume signal, the stock can remain elevated for weeks, but any delay in converting headline partnerships into backlog and gross margin inflection would likely compress the multiple quickly.

Microsoft’s setup is more interesting than the tape suggests because the stock’s relative weakness creates a cleaner asymmetry than chasing the IPO. Azure and Copilot are already monetizing AI through a broad distribution layer, so even modest acceleration in enterprise adoption can drive operating leverage with far less execution risk than a single-product chip start-up. The contrarian miss is that investors may be underweighting how much AI infrastructure spend ultimately accrues to the hyperscalers that control demand routing, not to the newest hardware entrant.

Near term, the base case is continued volatility rather than a straight-line reversal: Cerebras can stay momentum-driven as long as the IPO lockup/air-cover dynamic persists, but the first real catalyst is disclosure quality around bookings, margins, and customer concentration over the next 1-2 quarters. Microsoft’s catalyst path is slower but more durable; if the market regains focus on cash generation and AI monetization, the name should rerate over months, not days.

Key Decisions for Investors

• Long MSFT vs short CBRS as a relative-value pair for the next 1-3 months: favor monetized AI exposure over pre-profit hardware optionality; target 10-15% spread capture if sentiment normalizes.
• Sell CBRS strength into IPO momentum: trim or short into any post-listing extension above the first-day range; downside risk rises sharply if the market starts pricing in fundamental disclosure instead of narrative.
• Add MSFT on weakness over the next 2-4 weeks with a 6-12 month horizon: asymmetric setup if AI revenue contribution becomes more visible; downside is limited by cash flow and balance sheet support.
• For higher-risk accounts, buy MSFT call spreads 6-9 months out rather than outright stock: preserves upside to a valuation re-rate while limiting drag if AI spend growth moderates.
• Avoid chasing NVDA here unless there is a separate valuation reset: the article implies the market is already comfortable paying for AI monopoly economics, so incremental upside is less compelling than in the lagging hyperscaler.