Berkshire Hathaway Owns AI Exposure in 3 Different Ways. Here's Why That's Not the Same as Owning Nvidia.

Berkshire Hathaway has AI exposure through Apple, Alphabet and Moody's, which represent 20.2%, 7.4% and 3.4% of its portfolio, respectively, but it holds no Nvidia shares. Over five years, Nvidia returned 912%, far exceeding Apple’s 113%, Alphabet’s 138% and Moody’s 35%; analysts project Nvidia revenue and EPS CAGR of 59% from fiscal 2026 through fiscal 2029. The article argues Berkshire’s conservative preference for less cyclical AI beneficiaries may limit upside versus Nvidia while retaining indirect exposure to AI growth.

Analysis

The relevant market question is not whether Berkshire owns NVDA, but whether AI capital spending broadens from accelerators into power, cooling, networking and enterprise monetization. GPU suppliers capture the first wave of spending, while the next 6-18 months should favor VRT, ETN, GEV and CEG if data-center construction remains capacity-constrained. This is a more durable earnings pathway than extrapolating training-chip demand, because utility interconnection queues and electrical-equipment lead times create supply-side scarcity.

GOOG has the clearest rerating path among Berkshire-linked holdings: successful AI search monetization would remove the market's principal concern—query economics and share loss—while Cloud margin expansion offers a separate upside lever. AAPL's AI payoff is less directly monetizable and likely appears first through lower churn, mix and upgrade-cycle support rather than a discrete revenue line; its premium multiple therefore leaves less room for execution misses. MCO's opportunity is operational leverage and higher product stickiness, but its credit-cycle sensitivity remains materially more important than AI features over the next year.

The article's comparison is not investable without verifying current portfolio weights and consensus estimates; public-equity holdings are not the sole driver of BRK.A/BRK.B intrinsic value. Consensus may also be underpricing the possibility that AI infrastructure spending shifts from general-purpose GPUs toward custom silicon and inference optimization, which would compress NVDA's incremental pricing power even if total AI spending remains strong. Near-term sentiment is likely insufficient to move these securities absent hyperscaler capex, AI-product monetization, or semiconductor guidance.

Key Decisions for Investors

• No standalone BRK.A trade on this item. Treat any claimed portfolio exposure as a holdings-verification alert; reassess only after Berkshire's next 13F/quarterly filing and capital-allocation update.
• Prefer a 6-12 month long GOOG / short AAPL pair, sized beta-neutral, if GOOG maintains Search revenue growth without a material deterioration in operating margin. The thesis is AI monetization and Cloud upside versus AAPL's more indirect AI return; exit if GOOG's Search share or TAC-adjusted margin weakens for two consecutive quarters.
• For AI-spending exposure, favor a diversified basket of VRT, ETN and GEV over adding to NVDA after sharp hardware-led rallies. Enter around hyperscaler capex results; key risk is a data-center buildout pause, evidenced by falling order backlog, longer project timelines, or capex guidance cuts.
• Keep NVDA versus AVGO/AMD positioning tactical rather than structural over the next 1-3 months. A long NVDA position requires confirmation that gross margin and supply commitments remain resilient; rotate toward AVGO if hyperscaler custom-accelerator commentary indicates material 2027 deployment acceleration.