The market has a tidy narrative about Alphabet right now. AI capex is about to nearly double, and free cash flow - $91 billion last year - could collapse by 90% in 2026. The stock already ran roughly 65% in 2025 on AI enthusiasm, so the story goes: you are late, the spending is enormous, and the cash returns are a question mark.
Greg Abel disagrees. In his first quarter as CEO of Berkshire Hathaway - January 1 through March 31, the first time Buffett's successor has controlled the equity book - Abel tripled Berkshire's Alphabet stake by 224%, adding roughly 40 million shares worth nearly $17 billion. He also exited 16 other positions, including Amazon. The Alphabet stake now sits among Berkshire's seven largest holdings.
Abel isn't known for momentum-chasing. He's an energy and infrastructure operator who measures capital in decades, not quarters. When he buys, he usually does it when the numbers are getting better but the crowd is still uncomfortable. That's the setup here.
Alphabet reported its Q1 2026 earnings on April 30. Revenue came in at $109.9 billion, up 22% year-over-year. Operating income grew 30%. Operating margin expanded by two percentage points to 36.1%, even while the company raised its full-year 2026 capex guidance to $180–190 billion - roughly double the $91.4 billion spent in 2025.
Cloud revenue growth accelerated to the low-to-mid 60% range, the fastest among the hyperscalers. Search - still the cash engine - grew 19% with no sign of AI cannibalization at scale. Gross profit was $68.6 billion. Operating profit hit $39.7 billion in a single quarter.
The headlines focus on what FCF might look like next year. The income statement says the revenue base is growing fast enough to absorb the spending if the trend holds. The market can't hold both narratives at once.
The operating numbers don't match the FCF anxiety
Here's the trap. Investors see the capex guidance jump from roughly $91 billion to $180–190 billion and immediately subtract it from operating cash flow to get a terrifying FCF projection. Pivotal Research modeled a 90% FCF drop to roughly $8 billion for full-year 2026. That number is technically plausible and emotionally powerful. It also assumes revenue stays flat while spending doubles.
Revenue isn't staying flat. Q1 already showed 22% top-line growth with margin expansion. If that pace - or something close to it - carries forward, the revenue base itself will be large enough to soften the FCF whiplash. The real question isn't whether FCF drops. It's whether it bottoms quickly and then recovers as Cloud spending translates to Cloud revenue, which then compounds through Search and ads.
The market is still pricing Alphabet as a company spending $180 billion on AI with no proven return. That was the fair thesis six months ago. After a quarter where margins expanded and Cloud accelerated, the setup is cleaner than the tape reflects.
Why the market is still anchored to the wrong number
This is the one thing that matters: Cloud ROI. Google Cloud is the growth engine that justifies the capex cycle. If Cloud revenue keeps growing in the 40–60% range while reaching profitability at scale, the capex builds infrastructure that generates incremental revenue for years. The FCF trough becomes a one-year hit, not a structural problem.
If Cloud growth slows to the low 20s and margins don't follow, the $180 billion bet starts looking like infrastructure with no tenants. Then the valuation - roughly 25 times forward earnings - stops looking patient and starts looking expensive.
The financial bridge
Abel's bet is on the first scenario. Berkshire bought when the capex headline is at its peak and the operating numbers are at their strongest. That's the asymmetry he looks for.
Scorecard
- Thesis: Cloud-driven revenue growth absorbs the capex spike, and FCF recovers faster than the market prices.
- Timeframe: 12–18 months. The FCF trough should be visible by Q4 2026, with recovery signals emerging in 2027 Cloud margins.
- Target: Not a price target. The stock is already near all-time highs at roughly $387. The case isn't that today's price is cheap - it's that the operating trajectory justifies holding through the FCF scare. If revenue growth holds and Cloud ROI materializes, a re-rating above current levels follows naturally from earnings catching up.
- Tripwire: Full-year Cloud growth falls below 25% with no margin improvement. That suggests the capex isn't converting to revenue, and the thesis breaks.
What could still break it
I can be wrong again. The capex number is genuinely large - $180–190 billion is not a rounding error. If the AI spending cycle doesn't produce Cloud revenue as fast as management expects, or if Search advertising takes a structural hit from AI search products, the revenue base shrinks right when the bill comes due. That's the scenario where the FCF collapse becomes structural rather than cyclical, and a 25x forward multiple stops looking patient.
Berkshire has