The Issue 05 Footnote We Missed — Berkshire Tripled Their Google Position to $15.6B, and the Whole Reading Changes

The Lead

Issue 05 shipped on Sunday with the headline thesis that the AI consensus was split on Google — three funds initiated, Gerstner exited, the smart money was divided. That reading was based on the eight Tiger-Cub-style funds we’d been tracking since the system began.

After shipping, we ran an honest reconsideration of which funds we should be tracking, applying an independence test: are these funds genuinely independent decision-makers, or are they all in the same idea-sharing network? The eight funds we’d been tracking are almost entirely Tiger Cubs: Robertson-lineage growth funds that hire from each other, share analysts, and often coordinate informally. Convergence across that universe is partly just “the Tiger Cub network agreed.”

We added a single new fund — Berkshire Hathaway — as the only large-AUM filer outside the Tiger Cub idea-sharing network that passed the independence test. The first quarter we tracked Berkshire produced a signal that re-rates the entire Issue 05 Google reading: Berkshire tripled their Alphabet position from 17.85 million shares to 54.25 million shares — a +204% increase that adds approximately $10 billion of cost basis and takes the position from $5.2B to $15.6B.

The smart-money consensus on Google isn’t split. It’s bullish — and Gerstner is the outlier, not Berkshire.

Falsifiable form: Berkshire’s holding period on this size of add is multi-year, but the directional read is testable inside 18 months. If GOOG’s Q3 2026 combined ad-revenue + Google Cloud growth reaccelerates to ≥15% YoY, the Buffett team thesis was right. If the DOJ Search remedy + TPU capex compression combine to drag growth below 10%, Gerstner’s exit was the right call.

Threads

Berkshire tripled GOOG to $15.6B (+204% Q-over-Q)

Per Berkshire Hathaway’s 15 May 2026 13F-HR (accession 0001193125-26-226661):

• Q4 2025: 17,847,400 GOOG shares ($5.16B)
• Q1 2026: 54,249,798 GOOG shares ($15.60B)
• Delta: +36,402,398 shares, +$10.44B implied cost basis at quarter-average prices

GOOG is now Berkshire’s seventh-largest position behind AAPL ($57.8B), AmEx ($45.9B), Coca-Cola ($30.4B), Bank of America ($25.0B), Chevron ($17.5B trimmed -35%), and Occidental ($17.2B held flat).

This is almost certainly a deputy-managed position — Todd Combs and Ted Weschler each run ~$20-30B portions of the Berkshire equity book, and Buffett has historically been tech-averse beyond Apple. Adding $10B to a tech mega-cap is a deputy decision with Buffett sign-off. Either way: the call is “Google is cheap relative to its franchise quality even with the DOJ overhang and the TPU capex compression.”

The three plausible underwriting reads:

1. DOJ remedy is overdiscounted. Final judgment December 2025 imposes restrictions on Search distribution + requires sharing search data with competitors. The market priced this as material; Berkshire is calling it priced.
2. TPU capex narrative cuts both ways. GOOG’s Q1 2026 capex +108% YoY to $35.7B compresses near-term margins, but it also means GOOG owns the largest internal silicon program of any hyperscaler. Berkshire is betting on the long-run franchise, not the near-term margin.
3. YouTube + Cloud are durable cash machines. YouTube ad revenue + Cloud subscription revenue grow regardless of the Search overhang. The deputies likely modelled franchise value of those two alone and concluded it justifies the multiple.

Lab-calibrated: this is a strategic_shift × neutral setup (Berkshire’s apparent style: long-term franchise quality at value entry points), which carries the strongest historical edge in our 4,669-row sample (+3.03% mean direction-adjusted alpha at 20d, 58.6% hit rate, n=111). Without a multi-quarter lab calibration for Berkshire specifically, the strength of the signal isn’t reduceable — but the setup category is historically the best in our data.

What changes in the Issue 05 reading

Issue 05 published yesterday with these convergence counts:

Ticker	Issue 05 # funds	Updated # funds	Combined $
GOOG	4 funds	5 funds	$2.20B → $17.78B
MSFT	3 funds	4 funds	$1.38B → $2.46B
AMZN	5 funds	5 funds (Berkshire exited the small Q4 position)	$2.84B (unchanged)

The numerically biggest change: GOOG’s combined exposure jumped from $2.2B to $17.8B purely from the Berkshire add. That single move is larger than the combined GOOG holdings of the prior 8 funds put together.

The qualitative change is bigger. Issue 05 framed the GOOG signal as “split — three funds initiated, Gerstner exited.” With Berkshire visible, the framing should be: “Smart money is decisively bullish on GOOG. Gerstner’s exit was the outlier.” Worth holding two readings in tension though:

• Gerstner’s exit is tactical — driven by the TPU capex margin compression visible over the next 4 quarters
• Berkshire’s add is strategic — driven by the franchise value over the next 5-10 years

Both can be right. The market often resolves these by GOOG trading sideways for several quarters as the margin pressure plays out, then re-rating up as the capex cycle plateaus. The position implication: if you have a 1-year horizon, Gerstner’s exit may be correct. If you have a 5-year horizon, Berkshire’s add is.

The rest of Berkshire’s Q1 rotation — Visa + Mastercard exits, Delta re-entry

The GOOG move wasn’t the only Berkshire signal this quarter. Berkshire’s position count dropped 42 → 29 (-31%) and AUM declined $274B → $263B — the most aggressive Berkshire rotation since the 2020 pandemic re-positioning. Major moves:

Position	Action	$ change	Implied read
GOOG	+204%	+$10B	The big call
VISA	EXITED	-$2.91B	Full exit on payment network
MASTERCARD	EXITED	-$2.28B	Same — Berkshire is OUT of card networks entirely
UnitedHealth	EXITED	-$1.66B	Full exit, post-controversy
Chevron	-35%	-$9.40B	Major energy reduction
Domino’s Pizza	EXITED	-$1.40B	Consumer cyclical exit
AON, POOL, HEICO, CHARTER, LAMAR, ALLEGION	EXITED	-$2.85B combined	Cleanup of smaller positions
Delta Airlines	NEW	$2.65B	First airlines position since the 2020 mass exit
Apple, AmEx, Coke, BAC, OXY, CHUBB, MCO	Held flat	—	Core book unchanged

The Visa + Mastercard combined $5.2B exit is its own signal worth noting separately. Both card networks are quality compounder positions Berkshire has held for years. Exiting both simultaneously suggests:

• Either deputies have a structural view on payment-network disruption (stablecoin / CBDC / AI-mediated commerce changing the payment-rail value chain)
• Or they’re funding the GOOG + Delta adds and the card networks are the funding source (capital-rotation rather than thesis-change)

The Delta re-entry is unusual. Berkshire famously exited all four major US airlines in April 2020 during the COVID drawdown (“the world has changed for the airlines”). Reentering Delta in Q1 2026 at a $2.65B position is a meaningful directional reversal on the airlines thesis — they’re at least one of (a) cheap on cash flow, (b) recovering from a multi-year industry restructure, (c) AI-resistant business model. Worth its own dedicated read in a future issue.

Why this find justifies the fund-watchlist expansion methodology

The Issue 05 reading missed the GOOG signal entirely because our 8-fund universe was structurally biased toward the Tiger Cub network. Adding Berkshire — the only filer we deemed genuinely independent of that network — produced first-quarter signal value that justifies the inclusion criterion.

Five candidates we rejected as NOT worth adding (with reasons preserved for future reference):

Fund	Why we rejected	Verdict holds
Tiger Global	Apex of the Tiger Cub network we’re trying to diversify away from	Yes
Light Street	Glen Kacher is a Tiger Global alumnus — same idea-sharing network	Yes
Soroban	83% watchlist overlap; positions are infrequent due to hyper-concentration	Yes
Pershing Square / Third Point	Public theses published via shareholder letters before 13F lands	Yes
VC firms (Craft, a16z, Founders Fund, Sequoia)	13F = IPO distribution record, not active stock picking	Yes

The lesson preserved for the next watchlist expansion: independence of decision-making is the constraint, not AUM size or AI relevance. Berkshire’s deputies (Combs, Weschler, Abel) are demonstrably outside the Tiger Cub network, demonstrably opinionated, and demonstrably willing to make $10B+ contrarian bets. That’s the profile that produces signal.

The other addition we noted as potentially worth adding: Appaloosa (David Tepper) — only because he has explicit power-utility positions (VST $304M + NRG $253M = 9.4% of book) that none of our other funds hold, expressing the Aschenbrenner thesis via direct utility ownership rather than secondary-construction names. Whether to add Tepper is a separate decision for Issue 07+.

Watchlist Updates

Ticker	Direction	Driver	Lab-calibrated reading
GOOG	bullish (upgraded from neutral in Issue 05)	Berkshire +$10B add; consensus is positive, Gerstner is the outlier	Strong — strategic_shift × neutral has +3.03% mean (n=111). Berkshire signal trumps Gerstner exit.
V (Visa)	bearish (NEW signal)	Berkshire exit $2.91B	Tactical — bearish strategic_shift has -4.25% mean alpha. Treat as hypothesis.
MA (Mastercard)	bearish (NEW signal)	Berkshire exit $2.28B	Same — paired signal with Visa exit
DAL (Delta)	bullish (NEW signal)	Berkshire NEW $2.65B (first airlines re-entry since 2020)	Out-of-universe but worth tracking
UNH	bearish	Berkshire exit $1.66B post-controversy	Tactical

The Week Ahead

The next material catalyst is NVDA Q1 FY26 earnings on Wednesday 28 May (US after-close) — fiscal year ending late January, reporting calendar Feb-Apr quarter. Consensus revenue ~$48-52B (vs $44B prior quarter), Data Center segment expected to dominate. The single highest-information event of the next two weeks.

Specifically what NVDA’s print will test:

• Whether the AMD warrant disclosures (Issue 04) are visible yet in NVDA’s customer-concentration commentary
• Whether the AWS-OpenAI joint collaboration (Issue 03) is visible in NVDA’s Q1 hyperscaler revenue mix
• Whether the broader hyperscaler capex flow (Issue 03 thesis: $400B+ annualised) is converting to NVDA revenue at expected silicon mix
• Whether NVDA’s guidance for Q2 (calendar May-Jul) reflects any deceleration in any of the four major hyperscaler buy programs

Also in this window:

• Tuesday 19 May: No major scheduled watchlist events
• Friday 22 May: Any deferred 13F filings — Atreides + Aschenbrenner both non-filed deadline, possibly Form NT extensions land
• Monday 26 May: US Memorial Day, US markets closed
• Wednesday 28 May: NVDA earnings (the big one)

Methodology and disclosures

Filings Intel Digest is a bona fide financial publication. Nothing here is personalised investment advice. The editor may hold positions in companies discussed; current positions and policy are at /about. All claims are sourced to publicly filed documents linked inline. Backtested or historical figures are direction-adjusted and calibrated against the signal lab; calibration is partial — treat any forward statement as a hypothesis, not a forecast.

Frozen as of close-of-trading 15 May 2026 (Friday). This is an early “correction + addendum” issue — Issue 05 shipped Sunday morning AEST with the 8-fund analysis; Berkshire was added Sunday afternoon AEST after applying the independence-test methodology. Future fund-watchlist expansions will be documented in the relevant issue as they happen.

Watchlist now tracking 11 funds (8 Tiger-Cub-style growth + Aschenbrenner + Druckenmiller + Berkshire). Atreides and Aschenbrenner Q1 2026 13F non-filings remain outstanding; will integrate when filed.