Will Alphabet Break Warren Buffett's Cardinal Rule of Investing?

Berkshire Hathaway boosted its Alphabet stake from about $5.6B at end-2025 to nearly $37.8B by the end of Q2 this year, but Alphabet’s capital efficiency is deteriorating as AI capex rises. Post-tax ROIC fell from over 58% (2024) to ~42% (2025) with average estimates declining further to 38.3% this year, supported by guided ~ $200B capex for 2026 and expected increases in 2027. Analysts project ROIC dropping below 31% by 2029, creating a cautious setup despite Alphabet still generating >30% ROIC on a longer horizon.

Analysis

The market is likely misreading this as a pure quality-bearish story; the bigger mechanism is a rotation of value capture from platform owners to infrastructure suppliers. If Alphabet keeps front-loading AI spend, the near-term winners are the semiconductor and networking stack that monetize capex immediately, while GOOG/GOOGL equity holders absorb the depreciation and amortization lag before revenue payback is visible.

For Alphabet, the risk is not solvency or balance-sheet strain; it is multiple compression as the market discounts a lower near-term ROIC trajectory and asks for proof that incremental AI spend can outrun WACC. That makes this more of a 1-3 month sentiment and guidance trade than a 3-5 year fundamental thesis: if management signals capex re-acceleration without a corresponding monetization bridge, the stock can de-rate even if absolute profitability stays high.

Contrarian view: consensus may be overfocusing on headline ROIC and underweighting optionality value. A business with dominant cash generation can deliberately accept lower near-term ROIC to defend search and cloud economics, and Berkshire’s accumulation suggests long-duration capital still sees that tradeoff as acceptable; the falsifier is not a lower ROIC print alone, but evidence that AI spend fails to improve engagement, monetization, or cloud backlog within the next 2-4 quarters.

Key Decisions for Investors

• Short GOOGL on rallies into the next earnings/guidance window; use a 1-3 month horizon and cover if management explicitly narrows the gap between AI capex and monetization or if free cash flow re-accelerates. Risk/reward is decent because the stock can de-rate on narrative alone, but upside is limited by balance-sheet strength and buybacks.
• Pair trade: long NVDA vs short GOOGL for 1-3 months. Thesis: Alphabet's capex intensity is a transfer of value to the AI supply chain before it is a value creator for shareholders. Falsify if Alphabet shows clear AI monetization KPIs or if hyperscaler capex guidance rolls over.
• If you want defined risk, buy GOOGL put spreads into earnings rather than outright short stock. Structure for a post-print multiple compression move; exit if the company signals capex flattening in 2027-2028 or if cloud/search growth proves the payback period is shortening.
• Watch BRK.B as a sentiment buffer, not a standalone trade: Berkshire ownership can slow downside by validating the long-duration thesis, but it does not change near-term ROIC math. If BRK.B continues to add while GOOGL weakens, that is a sign to lean into the pair rather than fight the tape.