Founders Fund
Peter Thiel, Brian Singerman, Napoleon Ta (partners); registered investment adviser
Contrarian frontier-tech: the marquee blend of space (SpaceX), defense (Anduril), and AI (Anthropic, OpenAI). The only top firm where space/defense and AI are co-equal flagship bets — closest match to the space+AI thesis in one fund.
- Not open — Not an open-ended vehicle — analysis only, not an available allocation.
- No public record — No published return series — performance cannot be verified in either direction.
- Private marks — Material value sits in private marks that have not been exit-tested.
Key positions
Single largest, most valuable position; filed for a record ~$1.75T IPO in 2026
~$1.25B invested from the prior fund; secondary mark has more than doubled past ~$800B
~$1B invested; led its June 2025 round at $30.5B, now raising at ~2x — the defense flagship
Joined the $8.3B Aug 2025 round
Foundational Thiel win, now public — the template for the space/defense thesis
Recent moves
Closed a record $6B growth fund (May 2026) after burning through its prior $4.6B fund in under a year on Anthropic, Anduril, and OpenAI — the most aggressive late-stage deployment pace of the marquee firms.
Our take
The single best fund-level match for an investor who wants BOTH iconic space/defense (SpaceX, Anduril) AND frontier AI (Anthropic, OpenAI) in one vehicle. Thiel's contrarian, concentrated style means fewer, larger, higher-variance bets.
The question is not whether Founders Fund has been good — SpaceX settles that for the 2007-2010 vintages.
It is whether the mechanism still operates. The historical edge was underwriting assets nobody else would, at prices nobody else would pay. There is nothing contrarian about co-leading Anthropic's $380B Series G in February 2026 alongside GIC, Coatue and MGX, the most crowded trade in private markets. And the sharpest datapoint cuts against the access story too: having co-led in February, Founders Fund is not named among the leads, co-leads or significant investors in May's $65B Series H at $965B post-money. Price discipline or lost allocation — either reading matters, because the edge has migrated from insight to access, and access is contested, rentable, and has a far lower ceiling than a seed stake compounding 18 years.
Thesis
The reputation rests on a $20M Series C check into SpaceX in 2008, a stake press-reported at roughly 1.5-3% of a company that IPO'd at ~$1.77T in June 2026 — tens of billions, against a firm whose entire AUM is a fraction of that. The vehicle raising money today is not that firm. The $6B growth fund closed May 2026 writes $500M+ checks into about a dozen already-famous companies at marks from $380B to $1.8T: a growth-equity return profile wearing venture branding, judged on its own arithmetic, not on a 2008 seed whose economics belong to different funds and different LPs.
Concentrated contrarian frontier tech (space, defense, AI) in few very large positions. Historically early and genuinely against consensus: SpaceX 2008, Palantir, Anduril at inception. The current growth vehicles invert that — roughly a dozen companies, checks above $500M, into late-stage rounds for Anthropic, OpenAI and Anduril. A small senior group led by Thiel, Napoleon Ta and Trae Stephens runs it; Thiel and partners put in $1.5B of the $6B themselves.
Assessment
- Real early conviction where it mattered: SpaceX 2008 and Anduril at inception were underwritable by almost no one else at the time.
- $1.5B of the $6B is Thiel and partner capital — unusually strong alignment for a growth vehicle.
- Concentration is deliberate and openly stated, not diversification quietly sold as selection skill.
- Space, defense and AI in one book is a genuinely rare combination among marquee firms.
- The record leans heavily on one 2008 check. Strip SpaceX and Palantir and the case for repeatable skill is thinner than the brand implies.
- Contrarianism does not survive $500M checks into consensus AI leaders at $380B+ entry prices.
- Co-led Anthropic at $380B in February, then absent from the named investor list of May's $965B Series H — a discipline or allocation question either way.
- A small senior group deploying $6B in $500M tickets concentrates judgment very tightly.
- Singerman went partner emeritus in Dec 2024 while capital scaled — senior bench thinning as ticket sizes grow.
Record
No fund-level net IRR or DPI is disclosed, and that absence is itself the finding. The SpaceX IPO in June 2026 (priced $135, ~$1.77T) is the first realization event of scale; lockups release in ~7% tranches at 70/90/105/120/135 days, with larger post-earnings unlocks and roughly 40% of shares free by the 2026-12-08 180-day mark. That distributions therefore land into 2027 is our inference, not a disclosed schedule. Anduril at $61B (May 2026, led by Thrive and a16z, not Founders Fund) is a healthier arm's-length markup than a round the holder helped price. Until cash lands, nearly everything above the 2007-2010 vintages is TVPI, not DPI. Attribution matters: the SpaceX gain accrues to 2008-era funds and their LPs, not to buyers of the 2026 vehicle.
- 2026-07Update, 31 Jul 2026: SPCX closed $108.37 (~$1.43T), 19.7% below the $135 IPO price and 52% off the 16 Jun high of $225.64 — the first realisation of scale is being marked well under the IPO print.
- 2026-07Update, Jul 2026: the release schedule reads differently — ~911.5M shares two days after first earnings, ~7% tranches late Aug to late Oct, ~28% after Q3, winding down early Dec; one 455.8M tranche needs $175.50.
Risks & fit
- AI entry prices from $380B to $965B leave little room for multiple expansion; returns depend on revenue compounding for years.
- Defense and gov-adjacent holdings are correlated to a policy regime that can change with an administration.
- Post-IPO SpaceX marks are now daily and public, so reported NAV becomes volatile in both directions.
- Key-person concentration in Thiel, whose attention spans politics and multiple ventures.
- Mega-check growth investing competes with Thrive, a16z growth, sovereigns and crossovers on price, not insight.
Two things would settle it. First, realized DPI from the 2021+ growth vehicles once SpaceX lockups fully release and Anthropic reaches liquidity: net multiples below roughly 2x would show that mega-check late-stage entry cannot reproduce seed-stage economics regardless of brand. Second, ex-SpaceX attribution — if the historical record excluding SpaceX and Palantir fails to beat a public tech index over the same period, the reputation is one extraordinary 2008 decision rather than a repeatable process. Conversely, a clean 3x+ net on a fund whose winners are not SpaceX would show the process travels.
This is analysis of a strategy, not a recommendation, and access is not the reader's decision to make: Founders Fund is a closed private partnership whose LPs are institutions and sovereign wealth funds. The useful reading is analytical — a case study in how a venture franchise's edge changes character as AUM scales, how power-law records get attributed to firms rather than to single decisions, and why the gap between TVPI and DPI is the most informative number in private markets.
Fee terms are not publicly disclosed. Arithmetic worth noting: on a $6B vehicle every 1% of annual management fee is $60M a year regardless of outcome, and carry on $380B+ entries is charged against a growth-equity return profile.