Tiger Global Management
Chase Coleman (founder); Scott Shleifer (former private head)
Crossover public/private growth investor; a concentrated public book in mega-cap AI/internet plus one of the largest private AI venture franchises (PIP funds).
- Not open — Not an open-ended vehicle — analysis only, not an available allocation.
- Partial 13F — The disclosed book is longs-only and ~45 days lagged — a slice, not the strategy.
- Concentrated — Top disclosed positions are 50%+ of the disclosed book.
Top holdings
Top position; Gemini / AI search + cloud
Core AI compute
AWS hyperscaler
AI foundry
AI capex / Llama
Custom AI silicon
Recent moves
Top-10 concentration ~70% in mega-cap AI/semis; made just ~9 new private investments in 2025 with big marks on AI-linked names, and is shrinking its next VC fund to ~$2.2B in a deliberate 'small and beautiful' reset. Listed among SpaceX (xAI)'s investor roster.
Loading disclosed positions…
Our take
Tiger's tilt toward the 'plumbing of the AI economy' over apps, plus a disciplined smaller VC vintage, signals a higher-conviction, more patient posture after the 2022 drawdown — they want AI infra exposure on both sides of the public/private line.
The most instructive number in Tiger's history is its own: the first ten PIP vintages each held under $3B, averaged fewer than 50 investments, and returned ~34% gross / ~23% net IRR.
Then the vehicles became $6.7B and $12.7B, deal count hit the hundreds, and PIP 15 landed in the bottom 10% of 2021 funds with paper losses over 15% as of 6/30/24. That is a textbook capacity curve. The correction is real but its authorship is contestable: PIP 16 closed at $2.2B against a ~$6B target — a 63% shortfall, which is LPs declining as much as Tiger choosing — and PIP 17 targets roughly the same, explicitly modelled on PIP 16. So the reset is now two vintages old, and the 'disciplined return to small' framing is partly a narrative fitted to a fundraising outcome. Meanwhile the 2025 rebound (PIP 16 ~+33%, PIP 15 ~+16%) is a marks story sitting on OpenAI and Waymo, priced by primary rounds rather than exits, and the public book has migrated to positions so large and so consensus that the honest question is what fraction is stock-picking versus AI beta expressed with concentration.
Thesis
Tiger is two firms wearing one name. The public sleeve ($22.85B 13F, 3/31/26) is a concentrated mega-cap AI/semis basket — top 5 ≈48% — that a cap-weighted tech index largely replicates. The private sleeve holds the differentiation and the damage: PIP 15 ($12.7B, 2021 vintage) sat in the bottom decile of its vintage, and PIP 16 then closed at $2.2B against a ~$6B target. Our read: the 2025 recovery rests on private marks (OpenAI, Waymo) untested by exit, and the downsizing was at least partly LPs voting, not the firm choosing.
Crossover growth investing: a long-biased public book in mega-cap internet/semis run alongside a private venture franchise (Private Investment Partners, PIP I–XVII). The thesis is that seeing the same companies pre- and post-IPO compounds information advantage. Currently tilted to AI infrastructure — Alphabet 13.4%, Nvidia 9.2%, Amazon 9.1%, TSMC 8.2%, Meta 7.7% — rather than application-layer AI, on both sides of the public/private line.
Assessment
- Deployment discipline is visible in behavior, not just letters: ~9 new investments in 2025 after reviewing hundreds, versus the hundreds-per-vintage pace of 2021.
- Owning the AI infrastructure layer (TSMC, Nvidia) rather than the application layer is a defensible read on where AI margin currently accrues.
- Genuine crossover reach: PIP 16's OpenAI and Waymo stakes are access most public-equity managers simply cannot obtain.
- Long-cycle evidence that the underlying process worked at small scale — ~23% net IRR across the first ten sub-$3B vintages.
- Top-5 positions ≈48% of the 13F book in the largest, most-covered stocks on earth. Hard to argue proprietary insight into Alphabet or Nvidia at that weight.
- The downsizing reads as much like an LP verdict as a strategy choice: PIP 16 missed its ~$6B target by 63%, and PIP 17 targets the same reduced size.
- 2025's rebound is unrealized marks. PIP 15 sold 85+ companies for just over $1B against $12.7B committed — realized cash return remains very thin.
- OpenAI and Waymo marks are set by primary financing rounds, not exits. A funding-market repricing would move them fast and in one direction.
- 13F book fell $29.71B → $22.85B in a single quarter; redemptions, de-grossing, or hedging are indistinguishable from the filing alone.
Record
Attribution matters more than the headline. 2022 was −56% in the hedge fund, an arithmetic hole requiring roughly a 127% gain to recover — so investors who entered near the 2021 peak may still sit below high-water mark even after strong 2023–25 years. Both the collapse and the recovery are consistent with high-beta growth-factor exposure rather than demonstrated selection skill; the same factor tide explains both directions. On the private side, PIP 16 ~+33% and PIP 15 ~+16% in 2025 are appreciation in carrying value, not distributions. Distinguish the three possible sources cleanly: beta (owning the AI complex), leverage/concentration (sizing it hard), and skill (owning it before consensus). Tiger's public book today looks like the first two. The pre-2020 PIP record looks like the third — at a fund size the firm no longer runs, and, per PIP 16's shortfall, may no longer be able to raise.
Risks & fit
- AI capex cycle turning: the public book is a near-pure expression of one theme with little apparent diversifying ballast.
- Private marks converging downward to realizable value if the late-stage AI funding market cools.
- Liquidity mismatch — venture-duration assets in vehicles whose LPs have finite patience for a J-curve now four-plus years old.
- Key-person dependence on Chase Coleman across both books after Scott Shleifer moved to senior advisor.
- Concentration in names widely held across the crossover peer group, raising correlated-unwind risk in a drawdown.
Realized cash settles this. If PIP 15/16 convert marks into distributions — OpenAI or Waymo exiting at or above carrying value, DPI climbing past 1.0x — the 'marks are untested' critique fails and the crossover model is vindicated. Second falsifier: if the public book compounds meaningfully ahead of a cap-weighted mega-cap tech basket over a full peak-to-trough-to-peak cycle net of fees, that is skill, not beta, and our concentration critique is wrong. Third: a PIP 17 that closes at or above target would show the small-fund reset was chosen, not imposed. All observable within 24–36 months.
Aimed at readers tracking crossover funds as a market signal — how a large AI-exposed book is positioned, and what its private marks imply about late-stage AI valuations. Also a case study in fund capacity: ~23% net IRR at sub-$3B degraded at $12.7B, and a 63% fundraising shortfall shows who enforces that limit. Tiger's vehicles are private, institutional, and not generally accessible; nothing here describes an available investment or suggests one.
Fee terms are not publicly disclosed. Crossover managers of this type typically carry a management fee plus carry on both hedge and private vehicles; after a −56% year, high-water-mark mechanics also affect collection. Treat any specific figure as unverified.