3 Reasons Why I Finally Bought Nvidia Stock Last Week

Nvidia reported a 73% year-over-year revenue jump in its latest update and is trading around 20x forward earnings (15x on next-year estimates) while its trailing P/E is ~34x. Analysts expect revenue and EPS to rise about 71% and 74%, respectively, in the new fiscal year; the stock is ~21% below its all-time high but is ~12x higher over the past five years. The author discloses a new long position and argues the company remains attractively valued versus expected earnings growth, while flagging risks from Chinese trade restrictions and rising competition.

Analysis

Nvidia’s lead is increasingly a platform/stack advantage rather than just a product-cycle edge — its software ecosystems, interconnect standards, and data-center partnerships create demand stickiness that spills profits down the supply chain (TSMC/ASML capacity, PCIe-switch vendors, advanced cooling and power suppliers). That means a meaningful portion of incremental GPU TAM accrues to players who can scale manufacturing and systems integration quickly; conversely, discrete CPU incumbents without the software layer face deflationary pricing pressure and slower margin recovery over the next 12–36 months.

The principal macro/industry risks are policy-driven and structural: tightening export controls or an accelerated indigenous Chinese stack could compress long-term pricing power, while hyperscalers choosing vertically integrated accelerators (homegrown TPUs/ASICs) could cap ASPs for mainstream models. These threats operate on different horizons — trade-policy shocks can move markets within days, while native Chinese capability and hyperscaler internalization will materially alter economics over 18–36 months.

Given those dynamics, asymmetric trade structures that capture continued adoption while capping downside look attractive. A delta-focused options approach lets us participate in convex upside from model/enterprise adoption cycles while limiting exposure to short-term macro shocks. Equally, a cross-sectional pair that shorts legacy CPU execution risk (Intel) against Nvidia’s platform growth expresses the secular winner-takes-most thesis without requiring perfect timing on absolute equity direction.

The consensus is underweighting two second-order outcomes: (1) the used/gray market for datacenter GPUs meaningfully mutes near-term sell-through and clouds cloud-capacity demand signals, and (2) the server-infrastructure winners (switches, PSUs, cooling) will see structurally higher margins for 2–4 quarters — both create tradeable dispersion away from headline chip multiples.

Key Decisions for Investors

• Long NVDA call spread (12–18 months): buy a ~0.60–0.70 delta LEAP call and sell a farther OTM call to fund ~60–70% of the premium. Position size: 2–4% of risk capital. Reward: captures convex upside from continued AI adoption with capped premium loss; Risk: full premium if adoption disappoints or policy shock occurs.
• Pair trade — Long NVDA / Short INTC (equal notional): horizon 9–18 months. Position size: net market‑neutral notional with small long-delta tilt. Rationale: expresses platform vs legacy node execution divergence; Reward: asymmetric if NVDA retains share while INTC fails to monetize AI demand; Risk: macro selloff or simultaneous sector rerating compresses both names.
• Tail hedge: buy 6–12 month NVDA puts (small allocation ~0.5–1% of capital) or buy puts on a semiconductor ETF to protect against sudden export restrictions or an earnings shock. This caps single-event downside while leaving upside largely intact.
• Opportunistic long NFLX call spread (6–12 months): small tactical position (0.5–1% capital) to capture AI-driven content-cost deflation and better monetization experiments. Use a debit spread to limit cost; downside limited to premium if ad/ARPU initiatives stall.