Warren Buffett Watched Alphabet's Stock Price Climb 9,000% Before He Decided to Invest. Here's Why He and Greg Abel Are Making It One of Berkshire Hathaway's Biggest Investments Now.

Berkshire Hathaway has built Alphabet into its third-largest publicly traded equity holding after initiating the position in Q3 2025 at roughly $200 per share. Alphabet plans about $200 billion of capex this year for AI data centers and servers; Google Cloud CEO Thomas Kurian said average server investments pay back in under two years, implying roughly a 20% annualized return over a six-year accelerator life. Custom TPUs reportedly have about half the payback period of average servers, while five-year customer contracts provide demand visibility; Alphabet trades below 17x forward earnings.

Analysis

The investable implication is less Berkshire’s ownership than Alphabet’s potential transition from an AI-capex discount to a demonstrated infrastructure-ROI premium. If TPU utilization and contracted Cloud backlog validate management’s economics over the next 2-3 quarters, GOOG can sustain elevated capex without the free-cash-flow multiple compression that has constrained peers; vertical integration would also reduce its dependence on NVDA’s pricing power. This creates a second-order risk for merchant GPU vendors and AI-server assemblers if large hyperscalers increasingly substitute internal accelerators for inference workloads.

The key skepticism is that server-level payback is not equivalent to fully loaded project return: power, networking, land, depreciation, idle-capacity risk, and customer concentration can materially extend cash payback. The relevant catalyst is not another capex announcement but disclosure of Cloud operating-margin progression, backlog conversion, TPU attach/utilization, and consolidated FCF after lease obligations. A weaker enterprise-spending environment or rapid model-efficiency gains could leave capacity underutilized within 6-18 months, converting an apparent moat into depreciation drag.

GOOG’s valuation support is strongest if AI monetization expands Search pricing and Cloud margins simultaneously; that combination would justify multiple expansion rather than merely offsetting higher depreciation. Consensus may underappreciate that proprietary silicon can improve gross margin even if Cloud revenue growth moderates, but it may also overestimate the signaling value of Berkshire activity, which is not a recurring fundamental catalyst. Near term, the stock is vulnerable to any evidence that AI answers cannibalize high-value search queries faster than monetization catches up.

Key Decisions for Investors

• Accumulate GOOG on post-earnings weakness rather than chase sentiment; target a 6-12 month long if Cloud operating margin expands while capex-to-revenue stabilizes. Thesis fails if consolidated FCF falls despite revenue growth or management cuts Cloud backlog/ROI commentary.
• Run a 3-6 month pair trade: long GOOG / short NVDA in equal beta-adjusted dollars as a hedge against proprietary-accelerator substitution. Cover the short if hyperscaler capex guidance accelerates materially or NVDA data-center gross margin remains resilient despite evidence of TPU deployment.
• Use a GOOG call spread 6-9 months out, financed only after confirming the next earnings release shows Cloud margin and durable advertising pricing; this caps exposure to a broad AI multiple derating. Avoid a standalone bullish options position ahead of results because the critical utilization data are not independently disclosed.
• Treat BRK.A as an indirect, lower-beta expression only; its incremental sensitivity to GOOG is diluted by the operating-company portfolio and insurance float. Prefer GOOG directly if the objective is AI-infrastructure ROI exposure.