Jane Street
Secretive, employee-owned quantitative market-making/prop-trading firm (no single public face). Dominant in ETF and options market-making globally.
A market maker, not a conviction investor — its 13F is flow and inventory, ETF-heavy, NOT a portfolio of bets. The top lines are SPY/QQQ puts and calls, TSLA/GLD calls, and ETF positions (GLD, SPY, TLT) — the residual inventory of a firm that quotes both sides of millions of trades. Single-name equity (AVGO, PLTR appear) is small and incidental. This filing is the clearest case in the whole set where 13F must NOT be read as a view on the AI trade.
- Not open — Not an open-ended vehicle — analysis only, not an available allocation.
Top holdings
Market-making inventory + option overlays on both sides — flow, not a directional view.
ETF inventory; GLD calls also held. Not an AI position. (largest reported equity line)
Small single-name AI-semis line — incidental market-making inventory, not conviction.
AI name appears in the book but as residual inventory, not a bet.
Large-cap inventory and option underlyings; flow, not conviction. (top reported lines)
Recent moves
Q1 2026 13F (period ending Mar 31, 2026): top lines SPY/QQQ puts+calls, TSLA/GLD calls, GLD/SPY/TLT ETFs; reported gross book ~$777B vs ~$662B prior. Source: 13f.info / Holdings Channel / Grufity.
Loading disclosed positions…
Our take
This is the page's textbook market-maker caveat: Jane Street's 13F is inventory and flow from quoting both sides of ETF/option markets — the SPY puts AND SPY calls sitting side by side prove it is not directional. Apparent AI exposure (AVGO, PLTR) is incidental residual inventory, not a thesis; reading conviction into it would be a category error. Honest limit: 13F captures only the long US-listed snapshot of a book that is delta-hedged across shorts, options and global instruments — the AI-trade signal here is essentially zero.
The critique isn't the 13F caveat — it's the drift in what Jane Street is.
Classical market making is capital-light and self-funding. Jane Street has issued past $5.4bn of senior SECURED notes since January 2024, plus a senior secured term loan B upsized ~$500m in 2025, against ~$45bn members' equity (FY2025 year-end, reported May 2026). Firms that only earn spread don't pledge assets like that. The borrowing funds inventory duration: multi-day positions, credit, illiquid private stakes — a hedge fund grafted onto a market maker, with a market maker's disclosure regime. Capacity is the second issue: spread capture scales with volume, not capital, so incremental equity gets pushed into lower-Sharpe, longer-horizon risk. Revenue per head above $11m and a $9.38bn comp pool (~24% of 2025 revenue) say the marginal dollar is still very profitable — and that the firm must keep finding new risk to deploy against, each arena (India index options, private credit, venture) further from the original edge.
Thesis
Jane Street is not a fund and its 13F is not a portfolio — it is ~45-day-lagged, longs-only, US-listed residue of a delta-hedged global market-making book (SPY puts and SPY calls side by side). The story the filing hides: a firm that booked $39.6bn net trading revenue in 2025 (vs $20.5bn in 2024) and $16.1bn in Q1 2026 alone, funded increasingly by secured term debt and moving from microsecond quoting into multi-day risk and private AI stakes. The question is whether that is durable edge or a levered claim on a volatility regime.
Employee-owned proprietary trading and market making across ETFs, options, equities, credit, FX and crypto — quoting two-sided in size, monetising bid-ask and creation/redemption arbitrage, then hedging residual delta. Increasingly supplemented by 'medium-frequency' positions held days to weeks and direct private AI investments (CoreWeave, Anthropic per Bloomberg). Closed to outside capital; no fee exists because there are no outside investors.
Assessment
- Structural edge is real: two-sided ETF/options quoting in size is a genuinely capacity-constrained business with few credible global competitors.
- Employee ownership removes redemption risk and external fee drag — capital is permanent and incentives sit directly on firm P&L.
- ~$45bn members' equity (FY2025 year-end, reported May 2026) lets it warehouse risk when others withdraw — exactly when spreads are widest.
- Diversification across ETFs, options, credit, FX and crypto, and across geographies, is real rather than marketing.
- Revenue nearly doubled 2024→2025 and again YoY in Q1 2026 — that tracks the volatility regime far more cleanly than any improvement in skill.
- $5.4bn+ of senior secured notes since Jan 2024 plus an upsized secured term loan B is a capital structure a pure spread-capture business should not require.
- Migration into multi-day positions and private AI stakes (CoreWeave, Anthropic) is directional risk-taking marked without an audited NAV.
- SEBI alleged ₹4,843.57 crore of manipulative index-options profit — a material profit pocket depended on conduct a regulator later challenged.
- No audited return series, Sharpe, or drawdown history — every performance claim about this firm is press-sourced, not verifiable.
Record
Reported (press, not audited): $20.5bn net trading revenue 2024 → $39.6bn 2025 → $16.1bn revenue and $10.3bn net income in Q1 2026 (Bloomberg, Apr–May 2026). Attribution matters more than level. Market-making revenue is volume × spread × volatility, and 2025–26 delivered all three. Bloomberg attributes the Q1 2026 figure substantially to medium-frequency positions held days to weeks — directional risk, not spread capture — plus private AI marks. Context tempers the 'surge' framing: Q4 2025 alone was ~$15.5bn, so Q1 2026 is an incremental record, not a step-change. Meanwhile Indian index options, a press-reported ~$2.3bn profit pocket before the July 2025 SEBI order, remain untraded. With no audited series, no volatility figure and no drawdown record, no risk-adjusted judgment is possible — only the observation that the slope steepened when the vol regime did.
Risks & fit
- Volatility mean-reversion: a calm regime compresses spreads and volumes while debt service, headcount and comp expectations stay fixed.
- Regulatory: SEBI's probe expanded in Sept 2025 and the SAT appeal continues into 2026; Indian index-options trading remains unresumed.
- Leverage plus illiquidity: secured debt funding private AI stakes and multi-day inventory is a duration mismatch in a stress event.
- Key-person and retention: after ~$2.68m average 2025 payouts, a down year tests whether talent stays or spawns competitors.
- AI-trade concentration via both private stakes and correlated hedging books — a drawdown hits marks and flow simultaneously.
Two tests. First, a full year of low realised volatility (VIX persistently sub-15) with revenue holding near $35–40bn would falsify the vol-beta reading; reversion toward $20bn confirms it. Second, disclosure: the senior secured notes carry reporting obligations, so if bondholder financials show revenue holding while the multi-day and private-investment lines shrink, the edge is the quoting franchise as claimed. If the growth line is instead medium-frequency plus venture marks, the firm has quietly become a leveraged hedge fund and should be judged as one.
Analytically relevant to anyone modelling ETF/options liquidity, market-structure risk, or the private AI funding chain — Jane Street sits upstream of all three. Not investable: employee-owned, closed to outside capital, never took external money; only its bonds trade. Its 13F is the worst filing in this universe to read as a conviction signal — treating its SPY/QQQ option lines as a market view is a category error. Watch it as infrastructure and a volatility-regime barometer.
No management or performance fee — closed, employee-owned, no external LPs. The analogue is the comp pool: $9.38bn for 2025, ~24% of net trading revenue, ~$2.68m per employee (Bloomberg, May 2026). Outside exposure exists only via its senior secured notes.