Nvidia's Valuation Is Becoming Absurdly Cheap

Nvidia is rated Strong Buy, with Q2 FY27 revenue up 106% year over year and EPS up 120%; data center sales account for 92.5% of total revenue. The article expects Vera Rubin to reach 20% of data center revenue in Q3 and cites a $1.3T hyperscaler CapEx forecast for 2027. It also highlights a 14.5x FY28 forward P/E as undemanding.

Analysis

The key exposure is not simply chip demand: NVDA’s earnings path increasingly depends on hyperscalers converting large infrastructure budgets into deployed, revenue-producing systems. That creates a second-order bottleneck trade in power, cooling, networking and advanced memory, while also giving customers an incentive to shift suitable workloads toward custom silicon. The latter is a substitution risk to incremental GPU demand, not evidence that GPUs are broadly replaceable.

Near term (days), bullish positioning may persist, but the cited FY28 multiple is only as reliable as the earnings denominator. Verify the forecast’s margin, supply and share assumptions before treating it as valuation support. Over 1–3 months, watch Vera Rubin shipments and recognized revenue—not product announcements or management targets—and hyperscaler capex guidance. Over 6–18 months, the central test is whether customer returns on AI infrastructure sustain spending; a pause could pressure NVDA both through estimates and multiple compression.

Contrarian point: strong growth and a seemingly modest forward multiple can coexist with elevated execution risk if the market is capitalizing a peak buildout or assuming a smooth product transition. The bullish thesis is falsified by a material capex-guide-down from major buyers, delayed revenue conversion, or evidence that new-generation demand comes with weaker economics. No basis here to assume the cited ramp or capex forecast is independently confirmed.

Key Decisions for Investors

• Maintain a tactical long bias in NVDA, but add only on weakness or after shipment/revenue evidence confirms the Vera Rubin transition; avoid sizing off the forward P/E alone. Reassess if reported data-center growth or forward guidance materially decelerates.
• For defined-risk upside exposure, consider a call spread rather than outright calls ahead of the next earnings catalyst; cap premium at risk and avoid paying for an implied-volatility surge. The setup weakens if hyperscaler capex guidance is cut or product deliveries slip.
• Track hyperscaler capex guidance alongside NVDA revenue conversion, product mix and gross-margin commentary. Treat announced budgets as a watch item until deployment and customer returns are visible.
• Look for secondary beneficiaries in power, cooling, networking and memory only where order growth and capacity constraints are independently confirmed; the article provides no company-specific evidence to support a named supplier trade.