Bloomberg Daybreak: Nvidia Blowout (Podcast)

Nvidia delivered $96.2B in Q2 revenue and $2.22/share (ex-items), with data center revenue of $89B, and CFO Colette Kress guided fiscal 2028 revenue growth of ~70%—despite margin pressure from a memory cost surge. Central bankers at Jackson Hole will focus on a global inflation rethink and the risk of higher borrowing costs, where new price pressures from the Iran war, AI capex boom, and extreme weather have already driven some rate hikes. Separately, Fed Governor Lisa Cook denied Trump-linked mortgage fraud allegations, though the immediate market implication is likely smaller than the Nvidia outlook and the inflation/rate backdrop.

Analysis

The more important read-through is that NVDA is no longer just selling compute; it is extracting pricing power from a constrained AI supply chain. If memory inflation is strong enough to pressure near-term gross margin, that usually signals the ecosystem is still supply-bound, which is bullish for upstream component vendors but also raises the odds of customer budget discipline later this year. The first-order winner is NVDA's own earnings power; the second-order winner is memory and packaging capacity, while the second-order loser is anything downstream that needs cheap inference to justify rapid rollout.

On timing, the next 1-4 weeks are mostly about multiple risk rather than fundamentals: any hawkish Jackson Hole tone can compress high-duration AI valuations even if revenue momentum remains intact. Over 1-3 months, the key catalyst is whether NVDA can keep passing through input inflation without forcing hyperscalers to slow orders or shift mix toward custom silicon. If pricing power is real, NVDA can keep compounding; if it is just inventory scarcity, the market will start discounting peak growth before fiscal 2028 arrives.

The contrarian view is that consensus may be underestimating how inflationary the AI buildout is becoming. Rising memory, power, and financing costs can turn a great unit-growth story into a more mediocre free-cash-flow multiple story, especially if rates stay sticky. What would falsify the bullish setup is not a weak print, but a sequence of margin compression with no offsetting price increase, or evidence that hyperscaler capex is being delayed rather than accelerated.

Key Decisions for Investors

• Maintain a tactical long NVDA, but favor entry on 3-5% pullbacks or via 3-6 month call spreads; thesis works if margin guidance stabilizes and hyperscaler capex stays firm. Fails if gross margin trends stay below the implied recovery path for two consecutive quarters.
• Use a long NVDA / short QQQ pair into Jackson Hole over the next 2-6 weeks to isolate company-specific pricing power from macro duration risk. Risk/reward is attractive if rates volatility rises and AI multiples de-rate broadly.
• Set an alert on NVDA gross margin and average selling price commentary at the next update; if pricing power does not offset memory inflation, trim exposure and rotate to less valuation-sensitive semis. The key falsifier is a revised-down margin recovery path.
• Watch MU as a secondary beneficiary of memory tightness; if memory pricing keeps tightening while NVDA passes through costs, MU should outperform the broader semi basket over 1-3 months.