Is This Where AMD Catches Nvidia? Goldman Sachs Says It's Gaining Ground

Goldman Sachs upgraded AMD to a Buy and raised its price target to $640 from $450, citing accelerating AI infrastructure momentum. AMD shares jumped ~10% in morning trading, reaching 182% YTD and 306% over 12 months. The bullish setup is reinforced by reports of AMD GPU adoption by a Japanese autonomous driving startup (~10% of training on AMD GPUs) and a delay to Nvidia’s next-gen Kyber NVL144 system until 2028, plus upcoming catalysts around AMD’s Advancing AI 2026 conference (July 22–23).

Analysis

The important mechanism here is not “AMD beats Nvidia,” but that AI procurement is moving from a single-vendor architecture to a sourcing strategy. That favors AMD because every incremental design win comes off a much smaller base, so even modest share gains can drive disproportionate revenue growth and multiple expansion; for Nvidia, the first-order hit is less about lost absolute demand and more about a slower path to sustaining a premium multiple. The second-order beneficiaries are hyperscalers and enterprise buyers that gain negotiating leverage, plus alternative accelerator ecosystems that become more relevant once buyers insist on a credible fallback.

Near term, the 10% gap move looks flow- and narrative-driven, so the risk/reward is worse for chasing outright longs than for expressing relative value. The real catalyst is the July conference and the next earnings print: investors need concrete evidence on backlog, software traction, and deployment cadence, not just a broader “AI momentum” story. If AMD can show that Instinct is moving from pilot to repeatable production, the market will start underwriting a higher share of the AI capex pool over 1-3 months; if not, this is vulnerable to a post-event giveback.

Contrarian view: the market may be over-rotating a one-note narrative into a durable share shift. Nvidia’s installed base and software moat still matter more than roadmap timing, so a delayed next-gen product alone is not enough to justify a lasting change in leadership. The thesis is most fragile if Nvidia reasserts launch cadence or if AMD fails to raise AI revenue guidance; over 6-18 months, this is a competition story, but over the next few weeks it is mainly a sentiment trade.

Key Decisions for Investors

• Long AMD / short NVDA for 1-3 months as a relative-value expression into AMD’s July 22-23 conference; target further gap narrowing if AMD can validate AI design wins. Theses breaks if AMD offers no concrete monetization update or NVDA signals roadmap acceleration.
• Buy a limited-risk AMD call spread into the conference rather than chasing stock after the gap-up; this captures event-driven upside while defining downside if the market is disappointed by lack of guidance detail.
• If already long semis via SMH/SOXX, rebalance toward relative exposure and avoid adding AMD at current levels unless management raises AI revenue and margin assumptions. The move is likely to mean-revert if the event is “promising” but not numerically specific.
• Watch NVDA into the next earnings cycle for signs of real pricing pressure or share loss; if gross margin and demand remain intact, cover any NVDA underweight quickly because the market may be extrapolating one quarter of competitive noise into a structural trend.