Endowments & Foundations as LPs: GP Guide 2026

By Dawid, Founder of Altss. Writing about allocator intelligence and fundraising strategy.

US university endowments and private foundations collectively manage over $1.2 trillion in assets, with the largest endowments allocating 50–60% to alternatives. The "One Big Beautiful Bill Act," signed July 4, 2025, imposes tiered endowment taxes of 1.4%, 4%, or 8% on net investment income — Harvard could owe $368 million per year — accelerating PE secondary sales and reshaping how these LPs deploy capital. This article maps the market size, allocation models, tax implications, governance structure, decision chain, and targeting approach GPs need to understand before pursuing allocations that take 6 to 18 months to close.

Why do endowments and foundations require a different approach?

GPs approaching their first endowment meeting need to understand an allocator that is structurally different from the family offices most fundraising teams know well. Family offices can move in days. Endowments move in quarters. The governance layers, the LP decision cycle, the diligence expectations, and the relationship dynamics are fundamentally unlike what works in family office fundraising — and misreading any of them wastes months of pipeline time.

The Altss family office and institutional investors database covers 9,000+ family offices and an expanding universe of institutional LPs — endowments, foundations, pensions, insurers, sovereign wealth funds, OCIOs, and fund-of-funds — built on OSINT sourcing with human verification on every profile. The First-Time Fund Manager Playbook addresses endowment targeting in the context of Fund I fundraising. This article provides the complete endowment-specific map — from market size and tax changes through governance structure and practical approach strategy.

How large is the endowment and foundation market?

University endowments

US university endowments held approximately $837.7 billion across all institutions as of FY2024, per NACUBO. That number is heavily concentrated at the top: the 20 largest endowments control the majority of total assets.

The top five as of FY2025: Harvard University at $56.9 billion (11.9% return), the University of Texas System at approximately $47.5 billion, Stanford University at $47.7 billion (14.3% return), Yale University at $44.1 billion (11.1% return), and Princeton University at $36.4 billion (11.0% return).

Average FY2025 returns for $1B+ endowments came in at approximately 11.5% — the second consecutive year of double-digit gains following 11.2% in FY2024. The top three individual performers were public universities: University of Wisconsin-Madison at 16.2%, University of Michigan at 15.5%, and MIT at 14.8%.

Private foundations

US private foundations — a separate but related allocator category — hold roughly $1.1 trillion in assets across approximately 90,000 entities. The approximately 3,000 foundations with over $100 million in assets constitute the relevant allocator universe for GPs.

Their IRS 990-PF filings are publicly available and disclose total assets, investment income, grants, and investment management fees paid — making foundations among the most transparent institutional LPs. This is one of the most underutilized OSINT sources in fundraising, as detailed in the OSINT methodology article. Altss monitors 990-PF filings as part of its regulatory intelligence layer, flagging changes in foundation assets, new manager relationships, and fee structures that indicate deployment signals.

Combined, endowments and large foundations represent over $1.2 trillion in allocable capital with deep, structural commitments to alternatives.

What this market data cannot tell you

These aggregate figures mask wide variation. A $500 million endowment at a regional university and a $50 billion endowment at an Ivy League institution are categorized together in NACUBO data but behave like entirely different allocator types. Foundation asset figures include entities with zero alternatives exposure. The 3,000-foundation threshold for "relevant" allocators is an approximation — some smaller foundations have meaningful alternatives programs, and some larger ones are entirely indexed. GPs should use market-level data for orientation, not targeting.

How do endowments allocate to alternatives?

The "endowment model" — sometimes called the "Yale Model" after David Swensen's pioneering work at Yale from 1985 until his death in 2021 — rests on a straightforward insight: long time horizons and low liquidity needs allow endowments to earn an illiquidity premium by overweighting alternatives relative to a traditional 60/40 portfolio.

In practice, the largest endowments now allocate 50–60% to alternatives. Typical ranges: private equity at 15–20% of portfolio, venture capital at 10–15% at top-tier endowments, hedge funds and marketable alternatives at 15–20%, real assets (real estate, natural resources, infrastructure) at 8–12%, public equities at 20–30%, and fixed income at 5–15%.

Princeton's FY2025 allocation provides a useful reference: approximately 40% ($14.7 billion) of managed investments in private equity, with the balance across independent return assets, real assets, developed and emerging market equities, and fixed income. Stanford's $47.7 billion merged pool distributed $1.9 billion to operations — nearly double its 2015 payout — covering 22% of total university expenses. Harvard allocated $2.5 billion from its endowment in FY2025, a record, funding tuition and research as federal grants shrank.

What drove FY2025 outperformance?

FY2025's standout returns were not purely about broad alternatives exposure. The endowments posting 14–16% — significantly above the 11–12% cluster — had concentrated positions in AI infrastructure and crypto-adjacent companies. The University of Michigan is listed as an LP in Alt Capital II (a ~$275M AI enterprise software vehicle) and reportedly invested directly in OpenAI or an OpenAI-associated structure. That is not a diversified alternatives allocation — it is a targeted theme bet.

For GPs raising AI or digital infrastructure vehicles, the signal is clear: the most sophisticated endowments are comfortable making concentrated bets, sometimes going directly and bypassing the VC wrapper entirely. The January 2026 Deal Flow tracked comparable AI conviction bets across family offices and institutional allocators.

What the allocation data does not capture

Published allocation ranges describe targets, not actual exposure. Many endowments are above their stated PE targets — a condition that slows new commitments regardless of what the published allocation policy says. Allocation categories also vary between institutions: what one endowment classifies as "venture capital" another counts as "private equity." These classification inconsistencies make cross-institution comparison imprecise. GPs should verify actual exposure through direct conversations or 990-PF fee analysis rather than relying on published targets alone.

What changed with the endowment tax?

On July 4, 2025, the "One Big Beautiful Bill Act" was signed into law, transforming the endowment tax regime. The original 2017 Tax Cuts and Jobs Act imposed a flat 1.4% excise tax on net investment income for private institutions with endowments exceeding $500,000 per student and at least 500 students. The new law replaces that with a three-tier system.

The first tier applies a 1.4% rate to institutions with endowments between $500,000 and $750,000 per student — affecting schools like Emory, Duke, WashU, Penn, and Brown. The second tier imposes a 4% rate on endowments between $750,000 and $2 million per student — covering Stanford, Harvard, Notre Dame, Dartmouth, Rice, and Vanderbilt. The third tier sets an 8% rate on endowments exceeding $2 million per student — hitting Princeton, Yale, and MIT hardest.

The threshold was raised from 500 to 3,000 tuition-paying students — exempting many smaller institutions. Princeton, with a per-student endowment of $3.8 million, faces the maximum 8% rate. The tax now covers royalty payments and student loan interest income, broadening the base. The Treasury secretary is authorized to issue additional regulations to prevent tax avoidance strategies.

Effective date: taxable years beginning after December 31, 2025. For institutions with a June 30 fiscal year end, the new rates apply starting July 1, 2026.

Estimated annual costs: Harvard faces up to $368 million, Yale approximately $276–280 million, Princeton $217 million, and Stanford $202 million. Over five years, Harvard, Yale, Princeton, Stanford, and MIT could each pay over $1 billion in endowment taxes.

How does the endowment tax affect GP fundraising?

The tax creates three immediate effects on allocator behavior.

Accelerated PE secondary sales before the tax takes effect. Endowments are motivated to realize gains before July 1, 2026. Harvard announced plans to sell approximately $1 billion in private equity holdings on the secondary market, and Yale reportedly considered selling $3 billion worth of PE stakes. Yale's CIO Matt Mendelsohn publicly acknowledged the sales consideration. Princeton, with $14.7 billion (approximately 40%) in PE, faces similar pressure. The average LP buyout secondary sold at 94% of NAV in 2024. If elite endowments sell at discounts to avoid future taxation, it creates a temporary buyer's market for secondaries investors.

Portfolio restructuring toward tax-efficient structures. The tax applies to net investment income — interest, dividends, and realized gains — not unrealized appreciation. This creates an incentive to hold longer, avoid portfolio turnover, and favor structures that defer realization. GPs offering continuation vehicles, long-duration funds, or structures with deferred carried interest may find new receptivity from endowment allocators.

Potential reduction in new commitments. If Harvard is paying $368 million in taxes and distributing $2.5 billion to operations while managing $56.9 billion, the math constrains new allocations. Yale announced cost-cutting measures including budget reductions, a retirement incentive for staff, and a hiring freeze. Princeton's President Eisgruber announced on February 4, 2026 that the university must "look for areas where we can consolidate or cut" — budgeting for 8% expected endowment returns, down from 10.2%, which could mean an endowment worth $11.3 billion less in ten years with annual payouts $500 million lower. For GPs, this means mega-endowments may pace commitments more slowly in 2026–2027.

The net signal: endowments are not leaving alternatives. They are restructuring around a tax reality that penalizes realized gains and portfolio turnover. Patient, long-duration managers who offer structural flexibility will find this environment favorable.

What this analysis cannot predict

Tax policy is subject to revision. The tiered rates may be adjusted in future legislation. Endowment responses to the tax are still forming — the secondary sales reported by Harvard and Yale are announced intentions, not completed transactions. Actual behavioral shifts will depend on market conditions, secondary pricing, and each institution's specific liquidity position. GPs should monitor developments rather than assume the current trajectory is fixed.

How does the endowment decision chain work?

Endowment governance follows a layered decision chain. Understanding each layer — and the LP decision cycle timelines — is essential for any GP planning an endowment approach.

Board of Trustees or Regents

Sets the institution's overall investment policy statement, including long-term return targets and asset allocation ranges. Approves the spending policy — typically 4–5% of a trailing average endowment value. Meets quarterly. Does not approve individual manager selections at most institutions.

Investment Committee

A subset of the board — typically 5–12 members — with direct oversight of investment strategy. This committee approves asset allocation shifts, co-investment policies, and large single-manager commitments. Meets monthly or quarterly. Understanding the committee cadence, documentation requirements, and escalation thresholds determines whether a fundraising timeline is realistic.

Chief Investment Officer and investment staff

The CIO and their team are the primary relationship interface for fund managers. They source managers, conduct initial screening, manage consultant relationships, and present recommendations to the investment committee. This is who GPs are pitching. At large endowments, the investment office may have 15–40 professionals organized by asset class. At mid-sized endowments ($500M–$2B), the team may be 3–8 people. At small endowments (under $500M), there may be a single CIO with one or two analysts — or the investment function may be fully outsourced to an OCIO.

Investment consultants

The most important gatekeeper that GPs underestimate. Major consulting firms — Cambridge Associates, NEPC, Wilshire, Meketa, Aon, Mercer — advise the vast majority of US endowments and foundations. For many mid-sized institutions, the consultant's recommendation is effectively the allocation decision. Getting on a consultant's approved list is often a prerequisite to receiving an allocation from their client base. The LP Due Diligence Checklist covers how to prepare for consultant-grade diligence.

What does the practical timeline look like?

The full cycle from first meeting to signed commitment typically runs 6 to 18 months. Initial screening by investment staff takes 2–4 weeks. The CIO and staff conduct a manager deep-dive over 1–3 months. If a consultant is involved, their review adds another 2–4 months. The investment committee presentation requires one meeting cycle. Legal and operational due diligence takes 1–3 months. Board approval, if required, takes one additional meeting cycle.

This is not a family office process. There is no "quick close" with an endowment. GPs who misread the LP decision cycle — expecting a commitment within one or two meetings — burn the relationship. The Family Office Investment Criteria Framework documents how family office processes differ from institutional ones.

How should GPs approach different endowment sizes?

Mega-endowments ($10B+)

Harvard, Yale, Stanford, Princeton, MIT, Penn, Columbia, Duke, Northwestern, University of Texas, and University of Michigan. These institutions maintain in-house investment teams of 15–40+ professionals with asset-class-specific allocation heads. They build direct GP relationships and typically work with consultants for diligence cross-referencing but do not depend on consultant approvals. They are co-investment capable and may invest directly in companies. Minimum ticket size thresholds typically range from $10M to $50M+.

How to reach them: warm introductions through existing GP relationships or co-investors. Conference presence — CIOs speak at institutional conferences but rarely take cold meetings. The Harvard and Yale endowments have multi-decade manager relationships; displacement is rare. For emerging managers, the most realistic path is demonstrating differentiated strategy in a category where the endowment has an identified gap.

Large endowments ($1B–$10B)

Vanderbilt, Rice, Dartmouth, Emory, WashU, Notre Dame, and most large public university systems. Smaller teams of 5–15 professionals with meaningful consultant involvement. These institutions typically maintain 15–30 GP relationships per asset class and have more capacity for emerging managers than mega-endowments.

How to reach them: consultant relationships are the primary channel. Conference attendance at NACUBO, ILPA, and regional CIO gatherings provides initial exposure. OSINT signals from 990-PF filings reveal which managers they have historically allocated to and what fees they are paying — useful for identifying where they have gaps in coverage.

Mid-sized endowments ($250M–$1B)

Hundreds of universities, teaching hospitals, and cultural institutions. Often 2–5 investment professionals or a single CIO. Heavy consultant dependency. Many use an OCIO (Outsourced Chief Investment Officer) model — and if the endowment uses an OCIO, the GP's relationship is with the OCIO platform, not the institution itself. Direct outreach is less effective because the institution's staff does not have bandwidth to evaluate new managers independently.

Small endowments (under $250M)

Almost always use an OCIO or simplified portfolio. Alternatives allocation is low — typically 5–15% — and concentrated in fund-of-funds or co-mingled vehicles. Direct GP outreach is rarely productive at this tier. The realistic access points are fund-of-funds, OCIO platforms, or aggregator vehicles.

What this segmentation does not capture

Endowment size is a rough proxy for sophistication, not a rule. Some $500 million endowments have experienced CIOs running concentrated alternatives portfolios. Some $5 billion endowments are heavily consultant-dependent. Public university endowments face state-level governance constraints — open-meeting laws, political oversight, procurement requirements — that private endowments do not. GPs should verify the actual decision structure rather than assuming size dictates process.

How do foundations differ from endowments?

Private foundations share similarities with endowments but differ in three critical ways.

The 5% distribution requirement. The IRS requires private foundations to distribute at least 5% of net investment assets annually as charitable grants. This creates a structural liquidity need that endowments do not face — and creates tension with illiquid alternatives exposure. Understanding the spending policy constraints is essential for sizing appropriate ticket sizes. A foundation with $500 million in assets and a 5% distribution requirement needs $25 million in annual liquidity; a GP asking for a $20 million commitment to a 10-year lockup vehicle should understand where that liquidity comes from.

Family involvement. Many of the largest foundations — Ford Foundation ($16.5B), Lilly Endowment ($48.4B), Robert Wood Johnson Foundation ($13.5B) — have evolved beyond family control into professionally managed institutions. But thousands of smaller private foundations are effectively family offices with a charitable wrapper. The founding family sits on the board, influences investment decisions, and may have sector or geographic preferences that mirror the family's business interests. For these allocators, the targeting approach maps more closely to the Family Office Targeting Strategy than the institutional process described above.

Mission-related investing. Foundations increasingly integrate investment mandates with programmatic goals. Ford Foundation committed $1 billion to mission-related investments. In January 2026, Julia Thiele-Schürhoff of Stella Vermögensverwaltungs GmbH committed €200 million to responsAbility's emerging markets strategy — explicitly aligned with UN SDGs. For GPs raising impact, climate, or emerging markets vehicles, foundations represent a distinct LP category with genuine mission alignment — not the greenwashing that characterizes some institutional ESG mandates.

What 990-PF filings reveal about foundation investment behavior

Every foundation's annual 990-PF filing discloses investment assets, income, manager fees, and grant activity. A foundation paying $3 million in investment management fees on $200 million in assets is allocating approximately 1.5% in fees — suggesting meaningful alternatives exposure. A foundation whose assets grew 40% in a single year likely received a major gift and is deploying fresh capital. These are the kinds of signals Altss monitors through its OSINT regulatory intelligence layer — cross-referencing filing data against news, professional signals, and contact verification to build a current picture of each allocator's activity.

What 990-PF data cannot tell you. Filings are annual and typically 12–18 months stale by the time they are publicly available. Manager fee disclosures reveal aggregate fee load but not individual manager allocations. A foundation reporting $5 million in management fees could have two managers or twenty. Grant activity reveals programmatic priorities but not necessarily investment mandate alignment. GPs should treat 990-PF data as a starting indicator, not a complete portrait.

What structural shifts are reshaping endowment and foundation allocation in 2025–2026?

Beyond the endowment tax, three structural shifts are reshaping how these allocators deploy capital.

CalPERS Total Portfolio Approach. In November 2025, CalPERS became the first US pension fund to adopt a Total Portfolio Approach, replacing its strategic asset allocation model. Funds using TPA outperformed the traditional model by 1.3% per year over a decade. While CalPERS is a pension fund — not an endowment — its governance influence is enormous. If TPA demonstrates superior results, endowments will follow, which means managers may increasingly be evaluated against total-portfolio benchmarks rather than asset-class-specific track record metrics.

Overallocation to PE. Over half of global pension funds exceeded their PE allocation targets in Q1 2025. CalSTRS had a $7.49 billion overallocation — the largest globally. Endowments face the same dynamic: after years of increasing PE targets, many are now above target, which slows new commitments and creates openings for private credit, infrastructure, and co-investment vehicles that offer alternatives exposure without adding to the PE bucket.

Concentrated AI and crypto exposure as a differentiator. The FY2025 data reveals that endowments posting 14–16% returns had concentrated AI and crypto positions. Michigan invested directly in AI platforms. Duke showed pre-IPO Coinbase exposure. This marks a shift from the diversified alternatives allocation that characterized the endowment model for two decades — and it creates opportunities for GPs raising thematic vehicles. The January 2026 Deal Flow documents the same pattern across family offices — with Bezos Expeditions and Thiel Capital deploying directly into AI infrastructure.

What these trends cannot predict

TPA adoption by endowments is plausible but not confirmed — no major endowment has publicly announced a switch. PE overallocation may correct through distributions rather than secondary sales, leaving commitment pace unchanged. The concentrated AI bets that drove FY2025 outperformance carry concentration risk that could reverse. GPs should read these trends as directional signals, not certainties.

How should GPs practically approach endowment fundraising?

Determine the decision chain. Is the endowment managed in-house, by an OCIO, or through a consultant-driven model? If OCIO, the relationship target is the OCIO platform. If consultant-driven, start with the consultant. Getting this wrong at the outset means building a relationship with someone who cannot make or influence the allocation decision.

Research public filings. 990-PF filings (foundations), annual reports (public university endowments), and investment office disclosures reveal current allocations, manager relationships, and fee structures. The OSINT methodology article explains how Altss extracts fundraising intelligence from each filing type — and why OSINT-sourced intelligence captures allocator behavior that survey-based providers systematically miss.

Map the consultant network. If the endowment works with Cambridge Associates, NEPC, Wilshire, or Meketa, the path to an allocation likely runs through the consultant's recommendation. Building consultant relationships is a 12–24 month process — but once on an approved list, GPs gain access to the consultant's entire client base. For GPs who cannot access consultant channels, see How to Raise a Fund Without a Placement Agent for alternative approaches.

Align timing to governance cycles. Investment committees meet quarterly. Asset allocation reviews happen annually. Approach endowments when their allocation review cycle is open — not mid-cycle when no new commitments will be considered regardless of strategy quality. Altss tracks LP decision cycle signals — including investment committee cadence, personnel changes, and mandate shifts — through its OSINT regulatory intelligence layer.

Prepare for institutional-grade diligence. Endowments expect ILPA-aligned reporting, audited financials, a complete data room, reference checks with existing LPs, and operational due diligence covering fund administration, valuation policy, compliance, and business continuity. The LP Due Diligence Checklist maps every item GPs will need.

What are the limitations of this analysis?

This article synthesizes publicly available data from NACUBO, IRS filings, university annual reports, and news coverage. Several caveats apply.

Data staleness. NACUBO FY2024 data reflects June 30, 2024 values. FY2025 returns are preliminary. Foundation 990-PF filings are typically 12–18 months old. Between publication and the time a GP acts on this information, allocator positions may have changed.

Behavioral uncertainty. The endowment tax was enacted in July 2025 and takes effect for most institutions in July 2026. Endowment responses — secondary sales, restructuring, commitment pacing — are announced intentions that may not execute as planned. Market conditions, pricing, and legal interpretations will shape actual behavior.

Generalization risk. The endowment model describes how the largest, most sophisticated institutions allocate. It does not describe the median endowment, which is smaller, less alternatives-heavy, and more consultant- or OCIO-dependent. GPs should verify individual allocator profiles rather than assuming aggregate patterns apply to specific targets.

Altss coverage scope. Altss institutional LP coverage — including endowments, foundations, pensions, insurers, sovereign wealth funds, OCIOs, and fund-of-funds — went live in February 2026. Coverage depth is expanding continuously. As with all Altss data, every profile originates from OSINT sources and passes through human verification before production. Where coverage gaps exist, the platform reflects them rather than filling them with unverified data.

FAQ

How large is the US endowment market? US university endowments held approximately $837.7 billion as of FY2024 (NACUBO data). Including private foundations with over $100 million in assets, the combined allocator universe exceeds $1.2 trillion.

What returns did endowments earn in FY2025? The average for $1B+ endowments was approximately 11.5%. Top performers: University of Wisconsin-Madison (16.2%), University of Michigan (15.5%), MIT (14.8%). Harvard earned 11.9%, Yale 11.1%, Stanford 14.3%, Princeton 11.0%. Performance dispersion was partly driven by concentrated AI and crypto positions at outperforming institutions.

What is the new endowment tax? The "One Big Beautiful Bill Act" (signed July 4, 2025) replaces the flat 1.4% excise tax with a three-tier system: 1.4% ($500K–$750K per student), 4% ($750K–$2M), 8% (over $2M). The student threshold was raised from 500 to 3,000. Applies to tax years beginning after December 31, 2025.

How much alternatives exposure do endowments have? The largest endowments allocate 50–60% to alternatives. Princeton allocates approximately 40% to PE alone. Mid-sized endowments typically allocate 25–40% to alternatives. However, many institutions are currently above their stated PE targets, which slows new commitments.

How long does it take to get an endowment allocation? 6–18 months from first meeting to signed subscription agreement. The LP decision cycle involves initial screening, CIO deep-dive, consultant review (if applicable), investment committee presentation, and legal/operational due diligence. There is no equivalent of a family office "quick close."

Do I need a consultant relationship to access endowments? For mid-sized endowments ($250M–$2B), consultant relationships are often essential. For mega-endowments ($10B+), direct GP relationships are the primary channel — but displacement of existing managers is rare. For small endowments, OCIO platforms and fund-of-funds are the access point.

How do I research a foundation's investment activity? IRS Form 990-PF filings are publicly available and disclose total assets, investment income, manager fees paid, and grants made. Fee analysis can indicate alternatives exposure levels. Altss monitors 990-PF filings as part of its OSINT regulatory intelligence layer, cross-referencing filing data with news and professional signals to build a current allocator picture. The key limitation is that filings are typically 12–18 months stale.

Are endowments reducing PE allocations? Not broadly. However, many are above target allocation, which slows new commitments. The new endowment tax is driving accelerated secondary sales — Harvard plans to sell approximately $1 billion and Yale considered selling approximately $3 billion in PE holdings before the higher rates take effect in July 2026. These are announced intentions; actual execution depends on secondary market pricing and liquidity conditions.

What can this article not tell me about a specific endowment? Aggregate data describes the market, not individual allocators. Each endowment has its own governance structure, allocation targets, consultant relationships, and personnel dynamics. GPs should use this article for market orientation and then research specific targets through OSINT-derived intelligence, 990-PF filings, annual reports, and direct conversations. The OSINT methodology article explains how Altss approaches this at the individual allocator level.

The Altss family office and institutional investors database tracks endowment and foundation allocator signals alongside 9,000+ family offices and institutional LPs globally — built on OSINT, verified by human analysts, enriched by allocator self-reporting. To see how the platform maps decision-makers, mandates, and timing signals across the institutional allocator universe, see the platform.