Jensen Huang Sounds Eerily Like Warren Buffett as He Tells Investors to Buy the Dip in AI Stocks

NVIDIA reported $81.61 billion in quarterly revenue, up 85% year over year, with non-GAAP EPS of $1.87 versus $1.77 consensus and Data Center revenue up 92% to $75.25 billion. Management guided to $91.0 billion in Q2 revenue, disclosed $119 billion in supply commitments, and authorized an additional $80 billion in buybacks while raising the dividend from $0.01 to $0.25 per quarter. The article argues the recent 6% pullback in NVDA is a buying opportunity, reinforced by strong AI demand signals and supportive institutional flows.

Analysis

The key market signal is not that AI demand is strong; it is that capacity is still the binding constraint, which keeps pricing power and vendor concentration unusually high. That favors the dominant platform vendor near term, but it also means the second-order winners are the picks-and-shovels names tied to deployment velocity: networking, power, thermal management, and high-bandwidth memory suppliers should see the next leg of orders before end-demand fully re-accelerates. In contrast, any AI beneficiary without a credible infrastructure spend path risks being punished as allocators rotate toward cash-generative execution rather than narrative exposure.

The bigger contrarian point is that this is becoming a capital-allocation trade, not just a growth trade. When hyperscalers and model labs are publicly signaling multi-gigawatt commitments, the market will increasingly ask which vendors can monetize demand fastest and which adjacencies get commoditized as scale rises. That creates a likely split: the platform leader and the most essential infrastructure enablers can keep compounding, while software-only AI names with weaker ROI evidence may underperform despite being in the same theme.

Risk is mostly a duration mismatch. Over days to weeks, momentum and buyback support should keep the tape firm; over months, the key reversal would be any evidence of order digestion, delayed deployments, or margin pressure from aggressive customer bargaining. The market is currently pricing “no pause” in AI capex; if one or two large buyers slow spending into the next print cycle, the multiple compression would hit the most crowded longs first, even if long-run AI adoption remains intact.

Key Decisions for Investors

• Add to NVDA on weakness over the next 1-2 weeks, but express via staggered entries; target a 2-3 month hold with a trailing stop under the recent pullback low because the setup is supported by both fundamentals and large-scale capital return.
• Long NVDA / short a basket of high-multiple AI application names with weak free-cash-flow visibility for a 1-3 month relative-value trade; this captures the likely market preference for monetization over narrative.
• Initiate a basket long in AI infrastructure adjacencies (networking, optics, power/thermal) on any further sector dip; these names should respond first if hyperscaler capex remains elevated over the next two quarters.
• Use GOOGL as a secondary long only on confirmation that AI infrastructure spend continues to translate into cloud backlog conversion; otherwise treat it as a lower-beta AI holder versus NVDA and prefer NVDA for risk-adjusted exposure.
• Avoid chasing generic AI beta after a strong bounce; wait for either a fresh pullback or the next major earnings/capex update, since the near-term risk is crowded positioning unwinding rather than fundamental deterioration.