Sands Capital
Frank Sands Jr. (CEO, CIO) and team
Long-only / long-horizon growth manager (technically an asset manager, included as a growth-tech crossover) concentrating in secular AI, semiconductor, and global growth compounders.
- Not open — Not an open-ended vehicle — analysis only, not an available allocation.
Top holdings
Top position; added ~$2.5B in Q1 2026
AI foundry
Gemini / cloud AI
AWS hyperscaler
Commerce growth compounder
EUV litho monopoly
Recent moves
Added ~$2.5B to NVIDIA in Q1 2026, lifting it to ~12.4% and the clear top position, while holding a broad AI/semi spine (TSM, GOOGL, AMZN, ASML) — a conviction add into the AI compute leader at scale.
Loading disclosed positions…
Our take
Sands is the institutional, long-horizon end of this spectrum — a large $2.5B NVDA add signals that even a patient growth shop sees the AI compute leader as a core multi-year compounder rather than a trade, anchoring a diversified secular-growth book.
The tension is between Sands's discovery origin story and a $50B book.
Its largest disclosed position is Nvidia — 12.38% of the $25.49B Q1-2026 13F, 14.8% of the flagship fund at 31 Mar 2026, trimmed to 8.9% by 30 Jun 2026. That is the most widely-held stock on earth and a top weight of the very index the strategy is measured against. But the convergence case is weaker than that headline: at 31 Mar 2026 the fund held 59.2% giant-cap vs the benchmark's 67.3%, 45.0% info tech vs 51.1%, and its top ten included Spotify and Shopify — not index top-decile names. Active share was 59. This is not a closet indexer; it is a different book that happens to share the benchmark's single largest bet. The harder question is whether the difference was additive. Over ten years it was not — the non-index exposure cost roughly 4.7 points a year. Two readings compete: 2022 was one bad draw inside a sound long-horizon process, or $50B mechanically pushed a 30-stock strategy up-cap while the residual active bets stopped paying. The filings cannot separate them.
- 2026-06Top ten at 30 Jun 2026: Nvidia 8.9%, Seagate 7.2%, Micron 5.9%, cash 5.8%, Alphabet 5.7%, SanDisk 5.3%, Quanta 4.3%, Carpenter Tech 4.1%, Meta 3.9%, Block 3.6%. TSMC, Amazon, Shopify, ASML and Spotify are out.
- 2026-06At 30 Jun 2026: giant-cap 45.2% vs 59.0%, info tech 53.5% vs 54.5%, so the technology underweight is about 1pp, not 6. Active share 63, 32 holdings; the ten-year drag is 3.96pp a year, 14.62% vs 18.58%.
Thesis
Sands is not a hedge fund — it is a long-only, unlevered, no-short growth manager (D.C. area, founded 1992, staff-owned, $50B firm AUM as of 30 Jun 2026 per its own disclosure). No 2/20, no shorting, and its 13F is unusually representative rather than a fragment. So the critique that matters is not whether leverage is decaying an edge, but whether a concentrated growth shop has been paid for its concentration. On the flagship vehicle's numbers, over ten years it has not: 12.17% annualized vs 16.83% for the Russell 1000 Growth it chose (INST shares, to 31 Mar 2026).
Concentrated (~30 names), long-horizon 'growth-business' investing: buy companies judged to sit at the front of secular structural change, hold through multi-year compounding, accept high tracking error and deep drawdowns as the cost. Active share ~59-63; portfolio turnover 55% annualized — long-horizon in intent, not literally buy-and-hold. Public equity (Select Growth, Global Growth, EM) plus a private/venture arm. The US flagship benchmarks to the Russell 1000 Growth.
Assessment
- Long-only and unlevered: no financing risk, no short squeeze, no gate. Losses show up as drawdowns, not blowups — a structurally different failure mode from a levered fund.
- Genuinely differentiated book, not a closet index — 59 active share, 30 holdings, and underweight both giant-cap and technology versus its own benchmark.
- Because it is long-only and US-heavy, the 13F is a far more faithful read on the real portfolio than it would be for a levered or derivatives-heavy filer.
- Discloses the unflattering numbers plainly — the 1-star rating and the -50.42% 2022 sit on the sponsor's own fact sheet, next to the benchmark.
- A decade of paid-for concentration that did not pay: 12.17% versus 16.83% annualized for the Russell 1000 Growth (INST shares, 10 yr to 31 Mar 2026). The 5-year figure is worse — 0.16% versus 12.76%.
- Drawdown severity, not just direction. 2022 was -50.42% against a benchmark -29.14%; 2021-2023 compounded to roughly -21% while the index rose. That is concentration working in reverse.
- 1-star overall Morningstar rating among 998 large-growth funds (1Q-2026 fact sheet; 978 funds as of 30 Jun 2026) — a risk-adjusted verdict across three, five and ten years, not one bad cycle.
- 55% annualized portfolio turnover sits awkwardly beside the long-horizon framing. That implies a name replaced roughly every two years, not multi-decade ownership.
- Frank Sands Jr. holds CEO and CIO simultaneously at a founder-family firm, concentrating business and investment authority in one seat with no public succession plan.
Record
Attribute carefully: the record is vehicle-level, not firm-level — Sands runs many mandates and no single 'firm return' exists. The cleanest public proxy is the Touchstone Sands Capital Select Growth Fund ($1.9B, sub-advised), whose 1Q-2026 fact sheet reports INST-share annualized returns of 10.70% (1 yr), 17.16% (3 yr), 0.16% (5 yr) and 12.17% (10 yr) against Russell 1000 Growth figures of 18.81%, 21.18%, 12.76% and 16.83%. It trails at every horizon shown. The shape is explained largely by one hole: -50.42% in 2022 versus -29.14% for the index, after +69.56% in 2020. The 3-year number is measured from the bottom of that hole and flatters the process; the 5- and 10-year numbers contain it and are the honest read. Recent recovery has leaned on mega-cap AI names the benchmark also owns heavily, so relatively little of it reads as differentiated selection.
- 2026-07Update 30 Jun 2026: 18.4% sat in memory and storage — Seagate 7.2%, Micron 5.9%, SanDisk 5.3% — driving Q2's +23.34% against +16.74%, then roughly -7.8% in July. Differentiated selection, both ways.
- 2026-06Amended to the 2Q-2026 sheet (30 Jun, $2.2B): INST 6.80% / 20.30% / 2.32% / 14.62% vs 17.71% / 22.58% / 13.71% / 18.58%. Behind at all four, but year to date it led, +6.87% vs +5.33%.
Risks & fit
- Rate and duration: long-horizon growth is a long-duration asset. A higher-for-longer or re-accelerating rate path compresses exactly this book, as 2022 demonstrated.
- AI capex concentration — semiconductors and semi equipment were 26.4% of the fund at 31 Mar 2026. NVDA, TSMC and Broadcom are one correlated bet, not three.
- Persistent benchmark lag plus a 1-star rating invites redemptions, which in a 30-stock book force selling into the same names everyone else is exiting.
- Private and venture holdings carry unrealized marks that cannot be tested from outside until realizations occur.
- Key-person concentration in a combined CEO/CIO founder seat, with succession undisclosed publicly.
Our read — that the active portion of this book stopped paying for its risk — is wrong if, over the next three to five years, Select Growth's excess return versus the Russell 1000 Growth comes mainly from its non-index positions (the Spotify/Shopify tier, and initiations outside the benchmark's top decile) rather than from the size of the Nvidia weight. Tests: a rolling ten-year number that closes the ~4.7pt annualized gap, a five-year number that stops being near-zero, and 13F disclosures showing smaller initiations that later scale. If the gap closes on non-index names, 2022 was one draw.
Understanding what a large concentrated long-only growth mandate looks like from the inside, and as a case study in whether high-conviction growth investing survives a single catastrophic year. This is analysis of a strategy's logic and record, not a recommendation. Sands's vehicles are institutional separate accounts plus sub-advised mutual funds; nothing here is a view on whether any reader should hold, buy or avoid any of them, the underlying stocks, or the manager.
Asset-manager economics, not hedge-fund — a flat fee on assets, no carry on the public strategies. Touchstone net expense ratios per the 1Q-2026 fact sheet: 0.68% R6, 0.79% INST, 1.15% A. The institutional separate-account schedule is not public.