Greg Abel Just Broke Berkshire's 14-Quarter Streak of Net Stock Sales, Buying a Net $20 Billion Last Quarter. What That Reversal Means for Berkshire's Portfolio.

Berkshire Hathaway shifted from roughly three years of net equity sales to nearly $20 billion of net stock purchases in Q2 2026, buying $23.5 billion and selling $3.7 billion. About $17 billion of purchases were Alphabet shares, acquired amid the company's secondary offering to fund AI investment. Greg Abel also made a $5.8 billion Taylor Morrison Home acquisition, signaling a more hands-on operational approach as he establishes his post-Buffett strategy.

Analysis

The investable signal is not that Berkshire has broadly turned risk-on, but that capital allocation may be becoming more concentrated and transaction-driven under Abel. If the reported GOOG purchase is confirmed in Berkshire’s next 13F, it would account for the overwhelming majority of equity deployment; that is materially different from a diversified endorsement of the equity market. For BRK.A/BRK.B, the near-term valuation question is whether investors reward reduced cash drag or apply a governance discount for a more interventionist operating model.

GOOG faces a mixed second-order read: a large strategic buyer can absorb supply and support the stock over the next reporting cycle, but an equity financing event tied to AI investment would shift the debate from AI monetization to capital intensity, dilution, and return on invested capital. That debate is more consequential over 6-18 months for mega-cap software multiples than the incremental demand from one buyer. Confirmation of issuance terms, use of proceeds, and Berkshire’s ownership stake is essential before treating the purchase as a durable demand signal.

TMHC is the cleaner operating-risk expression of the leadership transition. Greater integration can improve purchasing leverage, land discipline, and overhead absorption, but it also raises Berkshire’s direct exposure to housing-cycle volatility precisely when mortgage rates, incentives, and land impairments can move earnings sharply. The contrarian view is that investors may be over-crediting a single quarter of activity as a new regime: a concentrated block purchase and a homebuilder integration do not yet establish a repeatable acquisition or public-equity playbook.

The next 1-3 month catalyst is Berkshire’s subsequent holdings disclosure and any commentary on repurchases, cash deployment, and the operating structure of its housing assets. A return to net equity sales, renewed large buybacks, or an AI-capex guidance increase without corresponding GOOG revenue/FCF acceleration would weaken the constructive interpretation.

Key Decisions for Investors

• Maintain BRK.B as a watchlist/neutral core holding rather than add on the allocation headline; add only if the next filing confirms diversified deployment or repurchases accelerate. Thesis is falsified if cash remains elevated while operating integration increases conglomerate complexity, supporting a valuation discount versus the S&P 500.
• Do not chase GOOG solely on reported Berkshire demand. Establish a 30-60 day event watch around the 13F and any financing documentation; consider a tactical long only if the purchase is confirmed and AI capex guidance is funded without material dilution. Exit the thesis on rising capex with downward FCF estimates or evidence that the issuance reflects financing stress rather than opportunistic balance-sheet management.
• For housing exposure, prefer a small TMHC short versus long ITB only after reviewing Berkshire’s integration plan and TMHC’s land/inventory marks. Risk/reward improves if mortgage rates remain restrictive and incentives rise; cover if rates fall decisively or TMHC demonstrates measurable gross-margin and SG&A benefits from centralized procurement.
• Avoid extrapolating the reported move into broad AI longs such as NVDA. The relevant 6-18 month risk is that equity-funded AI spending raises scrutiny of hyperscaler returns, which would favor selective software beneficiaries over incremental semiconductor beta if monetization lags capex.