Juniper Ventures · internal

Public LP intelligence — likeliest Juniper LPs

Thirty-one named institutions, ranked by probability of writing into a small, scaling AI-assurance fund. Built by Griff's method: the LP bases that public pensions and endowments are legally obliged to disclose are the target list. Every row carries a real disclosure source; every estimate is marked.

Read this before the tables. The method works, but the arithmetic does not point where it first appears to. A US public pension writing into venture typically has a $25–100M minimum check and an internal cap of roughly 10–20% of any single fund. A $30M Fund II therefore cannot legally or practically absorb a CalPERS or a Texas TRS ticket — their smallest check would be the whole fund. So the list splits into two jobs: Tier A + Tier B are the Fund II close (foundations and fund-of-funds nascent sleeves, $500k–$10M checks, decisions in weeks-to-months); Tier C is the $1B path — cultivate now, on a 24–48 month clock, so that Fund III/IV at $150M+ has warm, pre-diligenced institutional demand. Mining the disclosures today is how you know who to start the 24-month relationship with. [inference — check-size and concentration-cap conventions are widely observed practice, not a single citable rule]

Scoring

Each candidate was scored on four axes, weighted for a fund at Juniper's stage:

Reachability (×3) — can they write a $0.5–5M check into a $30M vehicle at all?
Mission fit (×3) — documented appetite for AI safety, assurance, alignment or responsible-tech themes.
Emerging-manager appetite (×2) — a named programme, a nascent sleeve, or a track record of Fund I/II checks.
Disclosure quality (×1) — how cheaply you can verify and monitor them programmatically.

Ranks run continuously 1–31 across all three tables.

Tier A — mission-aligned foundations (the Fund II close)

These are the only large institutions with both a stated AI-safety/assurance mandate and the legal freedom to write a small check. Several are already LPs or direct investors in exactly Juniper's thesis.

#	Name · type	AUM / VC allocation	Evidence they LP venture (named)	Est. check	Disclosure source	Why likely for Juniper
1	Omidyar Network · philanthropic investment firm	~$1.5B+ deployed to date; invests off balance sheet, no 5% payout constraint [inference]	Runs Project Alexandria, explicitly assembling LPs to anchor companies in "AI infrastructure, assurance and alignment"; coordinated the $7.5M Anthropic investment with Ford and Nathan Cummings; co-chairs the $500M Humanity AI initiative	$2–10M	ImpactAlpha, Feb 2026 · Omidyar	The single best-matched institution on earth. Their published thesis uses Juniper's own word, "assurance". They are actively recruiting co-LPs. Approach as anchor, not as a line item.
2	Ford Foundation · private foundation	~$16B endowment; MRI programme carved out of corpus	LP in Ex/ante's $33M inaugural fund (spun out of Schmidt Futures, 2023); direct co-investor in Anthropic alongside Omidyar; Humanity AI founding member	$1–5M	ImpactAlpha · Fast Company, Oct 2025	Proven to write LP checks into sub-$50M AI-governance funds. Ex/ante is the closest structural precedent to Juniper that exists.
3	MacArthur Foundation · private foundation	~$8B; long-running PRI/impact-investing programme	Co-chair of the five-year, $500M Humanity AI initiative with Omidyar; established PRI book	$1–5M	Humanity AI launch	Institutional machinery for below-market and market-rate mission investments already exists; AI is now a named priority. Follows Omidyar's lead by design.
4	Schmidt Sciences / Eric Schmidt (family office) · private capital	Multi-billion; discretionary	Seeded Ex/ante alongside Union Square Ventures and Ford; personally invested in Goodfire's $150M Series B	$1–10M	Goodfire Series B, Feb 2026	Already sits on Juniper's cap table by proxy — same company, same round. The warmest possible intro path and no committee.
5	Siegel Family Endowment · family foundation	~$1B+ [inference]	Humanity AI founding member; tech-infrastructure thesis; David Siegel (Two Sigma) principal	$500k–3M	Humanity AI members	Quant-founder DNA means interpretability and evals read as legible engineering, not as ethics. Fast decisions, small committee.
6	Nathan Cummings Foundation · private foundation	~$500M; 100% mission-aligned portfolio	Co-investor with Omidyar and Ford in the $7.5M Anthropic position	$250k–1M	CB Insights / Project Alexandria coverage	One of very few foundations that committed its entire endowment to mission alignment — small but a genuine repeat AI-safety allocator.
7	Doris Duke Foundation · private foundation	~$2B	Humanity AI founding member	$500k–2M	Humanity AI	Newly committed to AI as a portfolio theme; venture exposure would be additive rather than duplicative for them. [inference on venture LP appetite]
8	David & Lucile Packard Foundation · private foundation	~$8B	Humanity AI founding member; established MRI/PRI practice	$1–3M	Humanity AI	Scale plus existing impact-investment plumbing. Slower, committee-driven. [inference on venture LP appetite]
9	Mozilla Foundation · foundation + Mozilla Ventures	Mozilla Ventures ~$35M vehicle	Humanity AI founding member; operates its own trustworthy-AI venture arm	$250k–1M	Humanity AI	More useful as a co-investor and credibility signal than as a large LP, but the thesis overlap is near-total.
10	Mellon · Lumina · Kapor (group) · foundations	~$8B / ~$1.5B / ~$150M	All three Humanity AI founding members; Kapor has a long emerging-manager and diverse-GP record	$250k–2M each	Mellon announcement	Treat as one outreach cluster on the Humanity AI relationship. Kapor is the most likely of the three to LP a small venture fund.
11	Patrick J. McGovern Foundation · private foundation	~$1.5B, AI/data-focused by charter	AI-and-data is the entire mandate; grant-heavy to date	$500k–2M	990-PF via ProPublica Nonprofit Explorer	Purest thematic overlap outside Omidyar. [inference — no confirmed venture LP position found; verify against the 990-PF Part II line 13 partnership schedule before spending time]
12	William & Flora Hewlett Foundation · private foundation	~$13B	Ran the Cyber Initiative; now funds AI governance research	$1–3M	Inside Philanthropy, Jan 2026	Deep AI-policy grantmaking but historically grant-side, not LP-side. [inference — treat as a cultivation target, verify MRI mandate first]
13	Alfred P. Sloan Foundation · private foundation	~$2B	Technology and scientific-integrity grantmaking	$250k–1M	990-PF via ProPublica	Lowest-confidence row in Tier A. Included because the charter fits; no venture LP evidence located. [inference]

Tier B — fund-of-funds and emerging-manager specialists

For a sub-$100M manager, fund-of-funds are, in the words of one 2026 market review, "the last institutional door still writing checks." These are the fastest yes/no cycles on the list and each one that says yes de-risks Tier C two funds later.

#	Name · type	AUM / VC allocation	Evidence they back emerging managers	Est. check	Disclosure source	Why likely for Juniper
14	Cendana Capital · FoF, seed-specialist	$470M raised in 2023 for seed GPs; ~$2B+ platform	Explicitly carved a "nascent manager" sleeve for ~$1M bets into pre-Fund I / pre-institutional GPs, alongside a $10–20M core cheque	$1M (nascent) → $10–20M (core)	FoF market review, Jun 2026 · cendanacapital.com	The nascent sleeve is precisely sized for Fund II and is the on-ramp to a core cheque at Fund III. Highest-probability institutional yes on the whole list.
15	Sapphire Partners · FoF (LP arm of Sapphire Ventures)	Multi-billion platform	Named among the FoFs that back first-time managers; publishes the OpenLP body of work on emerging-manager diligence	$5–15M	FoF market review	Deep-tech literate, writes into Fund I/II, and their public diligence framework tells you exactly what to prepare. Check may exceed 20% of a $30M fund — likely a Fund III target.
16	Screendoor · FoF, first-time-manager only	~$100M+ across vintages	Entire mandate is backing new managers; backed by a consortium of established GPs	$1–3M	screendoor.co	Purpose-built for exactly Juniper's stage, and the GP consortium behind it is a distribution network in itself.
17	Recast Capital · FoF + emerging-manager platform	~$200M+ [inference]	Named among active first-time-manager backers; runs an accelerator for emerging GPs	$1–5M	FoF market review	Programmatic emerging-manager pipeline; low friction to enter their funnel even before a formal LP conversation.
18	Top Tier Capital Partners · FoF	~$8B	Named among FoFs backing first-time managers	$5–20M	FoF market review	Institutional scale with genuine emerging-manager willingness. Fund III-sized ticket.
19	Industry Ventures · FoF + secondaries	~$8B	Historically active in the emerging-manager space	$5–15M	VC Lab, Jul 2026	Secondaries book gives them a reason to build relationships early with managers whose positions they may later buy.
20	Al Waha Fund of Funds · sovereign-backed FoF (Bahrain)	~$200M+	Named among FoFs backing first-time managers	$2–10M	FoF market review	Sovereign programmes chasing frontier-tech access are unusually open to thesis-differentiated Fund IIs. Longer diligence, geopolitical considerations.
21	Horsley Bridge Partners · FoF	~$4B+ venture	The archetypal venture FoF; historically anchored many now-tier-one franchises at Fund I	$10–25M	PitchBook LP coverage	Aspirational Fund III/IV target. Getting on their radar at Fund II is the play; a check now is unlikely. [inference on current emerging-manager posture]
22	TrueBridge Capital Partners · FoF	~$6B	Dedicated emerging-manager and seed strategies; co-publishes the Forbes Midas List	$5–15M	PitchBook LP coverage	Runs separate accounts for public pensions — a TrueBridge yes is often the mechanism by which pension money actually reaches a small fund.
23	StepStone Group (incl. legacy Greenspring) · FoF / advisor	~$700B AUM+AUA	Largest venture FoF platform; advises many of the Tier C pensions on their VC sleeves	$10–30M	stepstonegroup.com	Two-for-one: an LP in its own right and the gatekeeper who tells pensions whether you are diligenceable. Brief them early even if they cannot invest yet.
24	Hamilton Lane · FoF / separate accounts	~$950B AUM+AUA	Runs the emerging/diverse-manager mandate for MassPRIM's FUTURE Initiative	$5–25M (on behalf of clients)	Hamilton Lane / MassPRIM	The literal conduit through which a $1B emerging-manager pension programme is deployed. Reaching MassPRIM means reaching Hamilton Lane.
25	GCM Grosvenor · FoF / emerging-manager specialist	~$80B	Long-running emerging- and diverse-manager mandates for multiple US public plans	$5–25M (on behalf of clients)	gcmgrosvenor.com	Same logic as Hamilton Lane: an intermediary that converts pension emerging-manager budgets into small-fund checks.

Tier C — public pensions & endowments (the $1B path; cultivate from now)

These are the institutions Griff's method surfaces directly: every name below publishes fund-level commitments, so you can read their existing venture book, find the overlap with Juniper's thesis, and start the relationship two funds early. Almost none can write into a $30M vehicle today.

#	Name · type	AUM / VC allocation	Evidence they LP venture (named funds)	Est. check	Disclosure source	Why likely for Juniper
26	CalPERS · US public pension	~$550B total; PE ~$111B (Jan 2026); VC raised from 2% → 6% of PE since 2022	Publishes a complete, quarterly, fund-by-fund PE list with commitments, vintages and IRRs. Named venture/growth GPs include Andreessen Horowitz ($400M into a16z's California Innovation Opportunities — a16z's first major CA pension money), General Catalyst, TA Associates, EQT	$100M–400M	CalPERS PEP Fund Performance Review · Fortune	The most transparent large LP in the world and structurally short of venture exposure after a deliberate strategy pivot. Also committed $1B in 2023 to seed new investment firms as an anchor LP — the one CalPERS door a small manager can actually knock on.
27	NYC Retirement Systems (Comptroller) · five public pensions	~$290B; $13.02B emerging-manager exposure in FY2025, up from $10.36B	Publishes an annual MWBE & Emerging Manager report naming managers and fund-of-funds conduits; updated EM definitions effective Jul 2024 covering direct, co-invest and FoF	$25–75M	NYC Comptroller EM/MWBE report	The largest emerging-manager budget in US public pensions, growing 26% year over year, with a documented FoF route that can reach a small fund indirectly.
28	NY State Common Retirement Fund · US public pension	~$280B; $1B+ emerging-manager allocation	Discloses monthly; recent emerging-manager allocations to Demopolis Equity Partners, Cross Rapids Capital, Palm Peak Capital (Mar 2026)	$25–75M	New Private Markets, Mar 2026	Monthly transaction disclosure makes it the cheapest large LP to monitor programmatically. Genuine, funded emerging-manager mandate.
29	Illinois SURS · US public pension	~$25B; deploys $500M+/yr to emerging managers	Statutory emerging-manager targets; board materials name every manager hired	$10–40M	LP allocation statistics, Mar 2026	Smaller plan size means smaller minimum checks — one of the few US public pensions whose emerging-manager check could fit a ~$100M Fund III.
30	MassPRIM · US public pension	~$110B; FUTURE Initiative: $1B target to emerging/diverse managers; $14B+ with diverse managers (~13% of PRIT, Dec 2024)	Board minutes name each FUTURE commitment; deployed via Hamilton Lane	$10–50M	MassPRIM diversity page · P&I	Legislatively mandated (Act Enabling Partnerships for Growth, 2021), so the budget cannot quietly disappear in a down cycle. Reach via row 24.
31	SFERS · US public pension	~$36B; PE at 28% of assets vs a 20% target; venture is the highest since-inception performer at +20.8%	CIO reports name every commitment: Trinity Ventures 2024 ($24.4M), NEA Secondary Opportunity Fund ($20M), Resurgens Technology III ($35M)	$20–50M	SFERS PE update, Sep 2025 · SFERS investments	Bay Area plan, venture-native, writes cheques as small as $20M, and publishes board packets that read like a target list. Caveat: currently over-allocated to PE, so near-term new relationships are constrained.

Tier C watchlist — same method, lower current priority

Name	Why it's on the list	Disclosure source
UTIMCO (Univ. of Texas, ~$80B)	LP in ~293 funds with a dedicated venture book; fund-level venture returns have surfaced publicly, including outsized Thrive positions. Endowments face no pension-style minimum-check floor.	utimco.org · Newcomer, Mar 2026
University of Michigan (~$21.2B endowment)	The single best free data source on this page: Regents' communications name each new venture commitment and its size — e.g. $164.5M across five managers in two months, incl. Plural (€15M), Accel ($15M), Sequoia China ($75M), Matrix ($36M). ~30% of the endowment is in venture.	regents.umich.edu meeting files · Markets Group
New Mexico State Investment Council (~$71B SWF)	Commits across PE and VC monthly and discloses each tranche ($120M Jun 2026, $861M May 2026, $545M Mar 2026); explicit in-state innovation mandate.	NM SIC private equity
Washington State Investment Board (~$179–211B)	Publishes monthly/quarterly/annual reports plus a standing private-equity IRR file; approves $500M–1.5B of PE per meeting. Large minimums, low emerging-manager appetite.	sib.wa.gov reports
Teacher Retirement System of Texas (~$210B)	Dedicates $500M+ to emerging PE managers; publishes its full private-markets book.	LP allocation statistics
LACERA · LACERS · LAFPP · SCERS · Hawaii ERS · Alaska Permanent	Mid-size West Coast and small-state plans: smaller minimum cheques, active emerging-manager programmes, full board-packet disclosure. The most realistic pension targets for a Fund III in the $100–200M range. [inference — verify each plan's current VC pacing and minimum before outreach]	Individual board agendas; aggregated in Dakota Marketplace

How to pull this programmatically

Roughly 40–45% of US public pensions run a formal emerging-manager programme, and the disclosure obligations that make them mineable are statutory, not voluntary. A working pipeline in rough order of cost-per-signal:

1. Structured, free, machine-readable — start here

CalPERS Private Equity Program Fund Performance Review — every active partnership with commitment, vintage, cash in/out, net IRR and multiple, refreshed quarterly with a ~2-quarter reporting lag. There is a printer-friendly view that renders the entire table in one HTML document, which is trivially parseable. printfunds view
CalPERS AB 2833 / AIV fee & expense disclosure — California statute compels per-vehicle fee, carry and expense disclosure; filed as board-agenda attachments each November. Gives you fee terms peers actually accepted. Nov 2025 filing
NYC Comptroller MWBE & Emerging Manager report — annual PDF, names managers and FoF conduits, gives programme-level totals. FY2025 PDF
WSIB reports hub — standing monthly/quarterly/annual files plus a dedicated private-equity IRR report. sib.wa.gov/reports
SFERS board packets — CIO reports and PE updates as predictable dated PDFs on a CDN path (cdn.mysfers.org/uploads/YYYY/MM/…), each naming new commitments and amounts.
U-Michigan Regents — monthly meeting files at a fully predictable URL pattern (regents.umich.edu/files/meetings/MM-YY/YYYY-MM-IX-N.pdf). The investment communication is almost always item IX. This is the highest-density venture-commitment feed available for free anywhere.

2. IRS filings — for the foundations in Tier A

Form 990-PF, Part II line 13 ("other investments") is where a private foundation's partnership interests live; the attached detail schedule is what names actual funds. Note the real limitation: the instructions permit aggregation by category ("domestic partnerships", "foreign partnerships"), and the AICPA has pushed to loosen it further — so some foundations name every LP position and some disclose one aggregate line. Coverage is uneven by filer, not by rule. IRS 990-PF instructions
ProPublica Nonprofit Explorer exposes e-filed returns as parsed XML with a documented API, so you can grep across thousands of filers for fund-name strings rather than opening PDFs. Caveat: XML exists only for electronic filings, and the bulk summary set lags. Nonprofit Explorer · get_the_990 parser
University endowments (Form 990) are a weaker source than pensions: Schedule F reports offshore investment by region, not by fund name, which is why UTIMCO's and Michigan's board disclosures — not their 990s — are the usable channel.

3. Manual / legal, when a plan is stubborn

State public-records requests. Most plans disclose commitments at open board meetings and then, as Fortune once put it, "the veil of secrecy begins to descend" — a targeted FOIA/PRA request for the private-markets commitment schedule usually recovers the fund-level list.
Trade-press commitment feeds — Pensions & Investments "searches-hires-mandates", Markets Group, Alternatives Watch, New Private Markets. Highest signal per minute of any source; effectively a human-curated diff of board agendas.
Do not bother with SEC Form D for LP identity — it discloses the fund raise, not who wrote the cheques. And note the ceiling on this whole exercise: a16z, like most tier-one firms, does not name its backers. You only ever see the disclosing side of the relationship, which is exactly why the method works from the LP end and not the GP end.

Suggested build: a scheduled scraper over the ~12 predictable sources above → normalise to (lp, gp, fund, vintage, amount, source_url, retrieved_at) → join against a hand-maintained list of AI-adjacent GPs → score each LP on venture share, cheque-size floor, emerging-manager programme presence, and mission-fit keywords. Two days of work, and unlike a subscription it produces a ranked list rather than a searchable one.

Validation: does a service already do this? (per Griff)

Short answer: three services own the raw data, none owns the ranking. Buying one is a reasonable accelerant; building one is still differentiated.

Service	What it actually covers	Price	Verdict for Juniper
Preqin	Deepest institutional-allocator dataset — pensions, endowments, foundations, SWFs, insurers. Tracks commitments, allocation targets and history, built from FOIA requests plus voluntary submissions from 2,200+ managers.	~$25k/yr single-asset-class; $40–60k multi-asset; institutions reportedly pay $50–150k	The closest thing to "already built". Prohibitive for a $30M fund's budget, and it still won't tell you who cares about AI safety.
PitchBook	Best-in-class deal and company data; LP allocation data present but thinner than Preqin at commitment level.	~$28k per seat/yr	Buy for deal work, not for LP work.
Dakota Marketplace	Specialised in public plans: 1,497 pension funds and 5,575 verified contacts, filterable by decision chain — board, staff, consultant — not just commitment history. Publishes free monthly and quarterly commitment summaries.	Mid four figures to low five figures/yr [inference — not publicly listed]	The one worth paying for. Decision-chain routing is the part that is genuinely tedious to scrape, and the free monthly summaries are a usable substitute in the meantime.
Altss · LPbacked · PipelineRoad	Newer, emerging-manager-oriented tools; family-office depth and faster refresh cycles. Altss is reported as the most common choice among Fund I–III managers.	Low four figures/yr	Cheap enough to trial. Treat vendor-published comparisons sceptically — several of these sites rank themselves first.

The gap, stated plainly. Every one of these answers "who commits to venture funds?" None answers "who commits to venture funds and has an AI-safety, assurance or responsible-tech mandate and can write a cheque small enough for a $30M fund?" That third clause is the entire product. It cannot be bought, it took this page's research to assemble by hand, and it is maintainable by one scraper plus a scoring function. Recommendation: subscribe to Dakota for contacts and coverage, build the mission-fit scoring layer in-house, and skip Preqin until Fund III.

Sequencing

Now (Fund II): rows 1–4 and 14–17. Omidyar's Project Alexandria is the anchor conversation; Eric Schmidt is the warmest intro (shared Goodfire cap table); Cendana's nascent sleeve is the highest-probability institutional yes.
Next 6 months: rows 5–13 as a Humanity AI cluster — one relationship, eleven institutions, and they visibly follow each other.
24–48 months (Fund III/IV, the $1B path): Tier C. Start now by reading their published books, finding the AI-adjacent GPs already in their portfolios, and getting introduced through those GPs and through rows 22–25, who are the actual gatekeepers of pension money into small funds.

One number to keep in view while doing this: experienced firms captured 90.9% of US VC capital raised in Q1 2026, up from 73.7% across 2025, and on emerging-manager platforms nearly 90% of H1 2026 commitments went to funds under $15M. Capital is concentrating at both extremes. A $30M Fund II with a 50x mark sits in the squeezed middle — which is precisely why the mission-aligned foundations, who are not benchmarking Juniper against Sequoia, are the right first door.