Viking Global Investors dumped Amazon in Q3. Then bought it back in Q4. The $55 billion Tiger Cub — led by Ole Andreas Halvorsen — completely liquidated its 3.9-million-share Amazon position last September, only to re-enter with a brand-new $722 million stake three months later. In the same filing, Halvorsen exited Meta Platforms entirely while opening a fresh $835 million Alphabet position. If this sounds like a hedge fund with an identity crisis, that's because it might be one.
The Q4 2025 13F, filed February 17, 2026, reveals a portfolio in violent transition: 21 new positions opened, 23 positions closed, and 55 continued holdings reshuffled — that's a 56% position turnover rate in a single quarter. Among the most striking moves: Halvorsen slashed his bank holdings by 60%+ across JPMorgan, Bank of America, and Capital One, while piling into Chewy (+147%), Ross Stores (+56%), and Visa (+38%). The result is a $37.7 billion portfolio that looks almost nothing like it did two quarters ago.
My read: This is a fund searching for its footing. Viking returned just 5.8% through November 2025, per Business Insider — lagging both the S&P 500 and its Tiger Cub peers. This marks the first full year without former investment chief Ning Jin, who departed in 2024 after a decade of co-managing the portfolio. The frenetic trading — average holding period under 19 months, with entire mega-cap positions opened and closed within a single quarter — suggests a team stress-testing thesis after thesis, unable to find the same conviction that once generated 89% returns in Viking's inaugural year.
The Amazon Boomerang: Sell Everything, Buy It Back
Let's talk about the elephant in the room. In Q3 2025, Halvorsen sold all 3,897,092 shares of Amazon — a position that had been the fund's eighth-largest holding. The Motley Fool noted at the time that profit-taking was one explanation, but also flagged concerns about "lofty expectations that go unmet" in the AI trade and Amazon's pricey P/E ratio of 34.
Three months later? Viking opened a brand-new 3,126,816-share position in Amazon worth $722 million. That's 80% of the shares they just dumped, bought back at Q4-end prices. The Q4 Amazon allocation (1.9% of portfolio) is smaller than the Q2 2025 position that preceded the exit (~2.8%), but the message is clear: the thesis changed, then changed back.
I'm skeptical. Either the Q3 exit was a mistake — which the Q4 re-entry tacitly acknowledges — or there was a genuine fundamental catalyst in the intervening 90 days that justified the round trip. Given that Amazon's stock price barely moved between September 30 and December 31 (from ~$187 to ~$231), the more likely explanation is execution friction: trading costs and opportunity cost from being out of the name during a modest rally. This is the kind of whipsaw that erodes alpha over time.
The Meta-to-Alphabet Swap: Choosing Your AI Champion
If the Amazon boomerang is puzzling, the Meta-to-Alphabet swap is at least intellectually coherent. Viking exited its entire 929,003-share Meta position ($682 million in Q3) and simultaneously opened a new 2,666,272-share Alphabet Class A stake worth $835 million.
The logic isn't hard to reconstruct. Both are AI beneficiaries with massive advertising businesses, but their risk profiles diverge sharply. Meta is spending $65 billion on AI infrastructure in 2025 with uncertain returns, while Alphabet has a more visible AI monetization path through Search and Cloud. At the time of the swap, Alphabet was trading at roughly 20x forward earnings vs Meta's 22x — a slight discount for arguably more diversified AI exposure.
This move is notable because it puts Viking on the opposite side of the trade from Pershing Square's Bill Ackman, who took a $1.76 billion new position in Meta during the same quarter, calling it "deeply discounted." When two of the world's sharpest hedge fund managers disagree this violently on the same stock in the same quarter, it tells you something about the genuine uncertainty in AI valuations right now.
The Great Bank Unwind
Viking Global built its 2024-2025 reputation on a massive financials bet. At one point, the fund's top four holdings were all banks: PNC, JPMorgan, Schwab, and Capital One. In Q3 2025, these four alone represented over 16% of the portfolio.
Q4 tells a different story. Halvorsen cut aggressively across the banking sector:
- JPMorgan Chase: -62.3% of shares (5.06M → 1.91M), dropping from #2 to #20
- Bank of America: -60.9% of shares (11.5M → 4.49M), collapsing from 2.1% to 0.66%
- Capital One: -60.0% of shares (7.39M → 2.95M), shrinking from 4.1% to 1.9%
- Charles Schwab: -16.3% of shares (16.6M → 13.9M), dipping from 4.1% to 3.7%
Combined, these four banks went from ~$6.3B (16.4% of Q3) to ~$3.96B (10.5% of Q4). That's $2.3 billion in bank exposure unwound in a single quarter. The exits of BlackRock ($722M) and Deutsche Bank ($215M) compound the financial sector retreat.
Was this prescient caution or premature selling? Banks had a strong 2024-2025, driven by rising net interest margins and deregulation hopes under the new administration. But the trade was getting crowded. Berkshire Hathaway also trimmed Bank of America repeatedly through 2024-2025, suggesting the smart money consensus was shifting away from financials. I think Halvorsen read the room correctly — bank stocks had largely priced in the rate environment, and the risk/reward had tilted unfavorably.
What Viking's Buying: The Conviction Cluster
While the exits grab headlines, the buys reveal where Halvorsen actually wants to deploy capital. The pattern is clear: quality mega-caps and consumer cyclicals.
The biggest increases by share count:
- Chewy: +146.9% shares (5.48M → 13.54M) — the pet e-commerce play doubled, now $447M
- Ross Stores: +56.5% shares (2.02M → 3.17M) — off-price retail conviction, now $571M
- Mid-America Apartment: +46.7% shares — REIT exposure expansion, now $524M
- Visa: +37.5% shares (2.90M → 3.98M) — payments dominance, now $1.40B (#4 holding)
- Microsoft: +32.4% shares (2.43M → 3.22M) — doubling down on the Q3 entry, now $1.56B (#1 holding)
- Boeing: +31.3% shares (3.01M → 3.95M) — turnaround bet, now $858M
- Air Products: +30.4% shares (3.67M → 4.78M) — industrial gas compounder, now $1.18B (#7)
- DraftKings: +27.4% shares (15.0M → 19.1M) — sports betting growth, now $659M
- Taiwan Semiconductor: +24.6% shares (3.94M → 4.91M) — AI chip fabrication, now $1.49B (#3)
The Microsoft doubling-down is the clearest signal. Halvorsen opened this position in Q3 at ~$1.26B and immediately added 32% more shares, pushing it to the fund's #1 holding at $1.56B. Given that he sold both NVIDIA and Amazon in Q3 to fund the Microsoft entry, the continued accumulation confirms this isn't a placeholder — it's a thesis. Microsoft's 39% Azure cloud growth rate, combined with a forward P/E of 25 (16% discount to its 5-year average), makes it Halvorsen's chosen vehicle for AI exposure.
Taiwan Semiconductor at #3 ($1.49B) reinforces the AI hardware theme without the Nvidia premium. TSM manufactures chips for virtually every major AI player — NVIDIA, AMD, Apple, Alphabet — and trades at roughly 18x forward earnings, a significant discount to its customers.
21 New Bets: Shotgun or Strategy?
Opening 21 new positions in a single quarter is aggressive by any standard. For context, Berkshire Hathaway typically adds 1-3 new positions per quarter; Pershing Square added just 2 in Q4. Viking's 21 new buys span:
- Mega-cap tech re-entries: Alphabet ($835M), Amazon ($722M)
- Financial infrastructure: Intercontinental Exchange ($778M), Tradeweb ($298M), Progressive ($303M), Chubb ($316M), Arthur J. Gallagher ($269M)
- Consumer cyclicals: Dick's Sporting Goods ($509M), Carnival ($429M), Stellantis ($434M)
- Healthcare: UnitedHealth ($395M), Thermo Fisher ($460M), Medline ($292M)
- REITs: Digital Realty ($233M)
- Industrials: Lennox International ($302M), Lowe's ($220M)
The 23 exits are equally telling. Out go Meta, Netflix, Nike, Philip Morris, BlackRock, KKR, Danaher, Chipotle, Intuit, and Deckers. The pattern in the exits: concentrated tech/growth names and financial managers (BlackRock, KKR) being replaced by diversified industrials and financial infrastructure.
I'll be honest: 21 new positions from a fund running 76 total names concerns me. That's not conviction investing — it's hypothesis testing at scale. When over a quarter of your portfolio is brand new, you're not compounding value; you're paying market-making fees and spread costs on position entry. For a fund that returned just 5.8% through November, the high turnover isn't generating alpha — it's consuming it.
Performance in Context: The Ning Jin Question
The uncomfortable truth about Viking's 2025 is the personnel question. Ning Jin, who served as co-chief investment officer for roughly a decade, departed in early 2024. The 2025 performance — 5.8% through November, well behind the S&P 500's 25%+ return — marks the first full calendar year without his influence.
This matters because Viking's historical returns were extraordinary. Halvorsen's fund returned 89% in its first year (2000), compounded at roughly 20% annualized through 2023, and managed $46 billion at peak. The fund's current AUM of ~$55 billion (with $37.7B in disclosed 13F equity holdings) reflects past performance, not current momentum.
The Q4 filing provides a window into the adjustment. The portfolio's top-10 concentration sits at approximately 34% — diversified by Tiger Cub standards but lacking the concentrated conviction bets that historically drove Viking's alpha. Compare this to Pershing Square's 72% top-5 concentration or Berkshire's 67% Apple concentration. Viking's shotgun approach to position-taking in Q4 reads as a fund that hasn't yet found its new equilibrium.
What to Watch in 2026
Three things will tell us whether Viking is finding its footing or continuing to drift:
1. Microsoft and TSM conviction: These are now the #1 and #3 holdings, representing the fund's clearest thesis — AI beneficiaries at reasonable valuations. If Halvorsen adds to both in Q1 2026, it signals genuine conviction rather than tactical trading. If he trims either, the identity crisis deepens.
2. The bank unwind pace: PNC remains the #2 holding at $1.52B, but nearly every other bank position was slashed 60%+. Does PNC get the same treatment next quarter? Or does Halvorsen maintain it as his single financial anchor while rotating the rest into insurance and fintech (ICE, Progressive, Visa)?
3. Turnover rate: If Q1 2026 shows another 20+ new positions and 20+ exits, the pattern becomes structural, not tactical. High turnover at Viking's scale (~$38B in equities) generates enormous transaction costs. At some point, the churn becomes the performance problem rather than the solution.
Viking Global remains one of the most storied names in the hedge fund world, and Halvorsen's two-decade track record commands respect. But the Q4 2025 filing reads more like a fund experimenting than one executing with precision. The Amazon boomerang, the Meta-for-Alphabet swap, the 56% turnover rate — these are the moves of a team recalibrating in real time. The question is whether the recalibration finds a destination, or becomes the permanent state.
Data sourced from SEC EDGAR 13F filings (accession 0001103804-26-000002 for Q4, 0001103804-25-000010 for Q3). Portfolio values as of December 31, 2025. Performance data from Business Insider (December 2025 reporting).
Originally published on 13F Insight
Top comments (0)