Berkshire Hathaway's Favorite Artificial Intelligence (AI) Stock Looks Like a Genius Buy Now

Berkshire Hathaway bought more than 48 million Alphabet shares across both classes in Q2, lifting the combined holding to 12.6% of its equity portfolio and making Alphabet its third-largest position. The article argues Alphabet has diversified AI exposure through Google Search AI Overviews, Gemini subscriptions, Google Cloud compute demand, and its Anthropic investment, whose leaders reportedly target a $2 trillion valuation in a future IPO. Alphabet shares are stated to be below the more than $340 level reached after its late-April Q1 results, framing the current valuation as an attractive entry point.

Analysis

The disclosed ownership signal is unlikely to create durable incremental demand by itself: 13F information is backward-looking and any near-term price response should fade within days. The investable question is whether AI raises Alphabet's revenue per query faster than it raises inference cost and reduces monetizable outbound clicks. Search margin dilution is the key underappreciated risk; higher engagement is not automatically earnings accretive if answer-generation displaces high-value commercial queries.

Over the next 1-3 months, the clean catalyst is evidence that Google Cloud AI backlog converts into revenue without another step-up in depreciation or sales incentives. Cloud growth can support multiple expansion only if segment operating margin expands alongside it; otherwise, the market will view AI capacity additions as a capital-intensity problem similar to the hyperscaler trade. Anthropic exposure should be valued as optionality rather than core NAV until a financing or IPO produces a verifiable mark.

The 6-18 month competitive issue is inference economics. Alphabet's custom TPU ecosystem could lower its internal cost per AI query and eventually pressure the marginal demand pool for NVDA accelerators, but only if third-party Cloud adoption broadens beyond captive workloads. Consensus may be underweight the possibility that Alphabet wins AI distribution while monetization lags, producing strong usage metrics but muted EPS revision breadth.

A bullish thesis is falsified by two consecutive quarters of decelerating Search revenue growth, Cloud-margin contraction, or capex/depreciation growth materially exceeding Cloud revenue growth. Conversely, sustained Cloud backlog conversion plus stable paid-click or revenue-per-query metrics would justify adding exposure even if near-term AI investment depresses free cash flow.

Key Decisions for Investors

• Maintain a modest long GOOG position, sized at 1-2% NAV initially, rather than chasing the ownership headline. Add only after the next earnings release confirms both Cloud margin expansion and stable Search monetization; target a 12-18 month holding period with a 2:1 upside/downside framework.
• Express the AI-infrastructure differentiation through long GOOG / short QQQ on equal beta for 3-6 months, rather than an outright mega-cap technology long. The trade isolates Cloud and AI-monetization execution; exit if GOOG underperforms QQQ by 10% following earnings or if Cloud operating margin reverses.
• Do not underwrite a valuation uplift from Anthropic until a third-party financing, IPO filing, or independently disclosed carrying value establishes realizable economics. Set an event-driven alert for an Anthropic transaction rather than treating private-market marks as earnings-equivalent value.
• Monitor NVDA relative to GOOG as a 6-18 month watch trade: widening TPU adoption and improving Google Cloud AI margins would support a small long GOOG / short NVDA pair, but do not initiate without data on accelerator mix, external TPU availability, and Alphabet's committed GPU purchases.