Greg Abel Bought $39.4 Billion of Stocks in 6 Months, Up From the $7.1 Billion Berkshire Purchased a Year Earlier Under Warren Buffett

Berkshire Hathaway shifted to a net buyer in Q2 2026, buying $23.5B of stocks and selling $3.7B, reversing a multi-year net-selling trend. In the first six months of 2026, it cumulatively purchased $39.4B of equities and sold $27.8B, versus $7.1B bought and $11.6B sold in the same period last year, with Alphabet becoming its third-largest holding. The company still carried about $365.5B in cash and Treasuries as of June 30, suggesting continued selectivity, even as it seeks opportunities in a “frothy” market.

Analysis

The important signal is not the size of the purchase, but the type of asset chosen: a cash-rich platform with durable earnings power and AI optionality that does not require betting on frontier-model capex. That favors GOOGL relative to capital-intensive AI infrastructure names because incremental AI revenue can flow through at far higher margin and with less balance-sheet risk. Near term, expect a modest quality-rerating effect as value-oriented allocators borrow the signal, but the move only compounds if the next few quarters show AI monetization without a step-up in reinvestment intensity.

For BRK.B, the cash pile is both a weapon and an anchor. If risk-free rates stay elevated, the opportunity cost of sitting on Treasuries remains manageable, but the bigger issue is that the investable opportunity set still looks thin at current market multiples; that caps enthusiasm for a "Berkshire as aggressive redeployer" narrative. If the next 1-3 quarters show no meaningful buyback acceleration or new large equity additions, the stock is likely to trade more on insurance underwriting quality and less on capital-allocation mystique.

Contrarian view: the market may be overreading this as a broad AI endorsement when it is really a comment on valuation discipline and durable free cash flow. That makes the immediate reaction in GOOGL probably smaller than headline readers expect, but it also means any post-earnings evidence of accelerating cloud/search monetization could produce a second leg higher. Falsify the thesis if GOOGL's AI-related capex rises faster than revenue contribution, or if BRK.B's cash balance keeps growing without a visible deployment path.

Key Decisions for Investors

• Long GOOGL / short QQQ or XLK for a 1-3 month relative-value trade: use the Berkshire signal to express a quality rerating in a single name rather than broad tech beta. Risk/reward is attractive if GOOGL narrows its valuation gap versus mega-cap peers; exit if the next earnings cycle shows AI monetization lagging capex.
• Buy 3-6 month GOOGL call spreads on weakness rather than chasing strength. This keeps premium outlay modest while capturing any follow-through from buy-side mimicry and multiple expansion; thesis is invalidated if the stock fails to hold post-news support or if management guides to faster capex growth without corresponding revenue lift.
• Do not short BRK.B on the cash balance alone; treat it as a wait-for-confirmation name. Add only if future quarters confirm sustained equity deployment or repurchase acceleration, because otherwise the stock remains a low-beta compounder with a built-in cash drag that limits upside.
• Underweight BAC/COF relative to GOOGL in quality-tilted portfolios. The trims are not an outright sector call, but they reinforce a preference for businesses with cleaner earnings durability and less sensitivity to credit-cycle compression.