The Packard Foundation's Grantmaking Portfolio: One Foundation, Two Portfolios

Foundation Portfolio Analysis · 2022–2024

What three years of filings reveal about a flagship West Coast science and conservation funder: one foundation running two portfolios, the institutions it built and the grants everyone else competes for.

---

Download the full report (PDF)

---

A Note to the Reader

Throughout this report, claims are tagged according to how directly they are supported by the underlying data. [DATA] marks figures and patterns drawn straight from IRS 990 filings. [INFERRED] marks interpretations that go beyond the raw counts, where analytical judgment is required and alternative explanations remain possible. [REPORTED] marks a small number of claims about institutional affiliations and corporate history that are drawn from the named organizations' public materials and standard nonprofit registries, cited in the sources.

The portfolio described here is what Packard did fund, not what it set out to fund or what it considered. Selection criteria, declined applications, and the broader universe of eligible recipients are not visible in this data and should not be inferred from it. Three structural caveats shape the findings. First, 11 percent of grant dollars flow to international recipients that do not file U.S. returns, and the matched U.S. portion separates into 501(c)(3) nonprofits, universities, and for-profit/other entities (a full reconciliation appears in §02). Second, a three-year window cuts every multi-decade relationship at both ends. Third, Packard's grant-purpose text consists almost entirely of terse program labels rather than descriptive sentences, which makes program attribution unusually clean but makes keyword-based theme detection unusually conservative.

This report is part of a comparative foundation-analysis series; occasional contrasts with the Gates Foundation appear where they sharpen the structural reading of Packard's portfolio.

---

§01 · Executive Summary

The Portfolio at a Glance

This analysis examines every grant made by the David and Lucile Packard Foundation across tax years 2022 to 2024, drawn from IRS Form 990-PF Schedule I and resolved against the full U.S. 501(c)(3) population. The dataset captures 3,284 grants totaling $1.15 billion to approximately 1,324 distinct recipient organizations. Extracted grant totals reconcile to 100 percent of the official Part I grants-paid figure in all three years, to the dollar. The window is the three most recent complete annual returns available in the public IRS e-file extracts.

Five Findings

1. Packard runs two portfolios in one. [DATA] [REPORTED] A quarter of all giving ($277.0 million, 24.1 percent) sustains the institutions the family built. The Monterey Bay Aquarium Research Institute (MBARI), the oceanographic institute David Packard founded in 1987, is the single largest recipient at $199.3 million across just five grants in three annual funding cycles, 17.3 percent of all dollars and the three largest individual grants in the portfolio. The Lucile Packard Foundation for Children's Health, the fundraising arm of the Stanford children's hospital that bears Lucile Packard's name, received $60.0 million in two grants, and the Monterey Bay Aquarium Foundation another $17.7 million. The remaining $872 million is the external grantmaking program: the right starting denominator for any organization reading this portfolio as a prospect, though it still includes the intermediary layer and international giving; §02 reconciles the components.

2. Climate and environment is the dominant external theme, and a large share of it moves through intermediaries and platforms. [DATA] [REPORTED] Environment-and-science program lines carried $319.0 million, 27.8 percent of all dollars, the largest theme by nearly a factor of five. Among matched U.S. recipients, Packard over-indexes on environment (NTEE, the National Taxonomy of Exempt Entities, code C) at 4.8 times the sector base rate and on science and technology (NTEE U) at 8.0 times. At least 23.6 percent of matched dollars flow through intermediaries and pooled funds rather than operating organizations, led by Rockefeller Philanthropy Advisors ($42.5M), ClimateWorks ($33.9M), RF Catalytic Capital ($25.0M), Resources Legacy Fund ($18.2M), and the legacy Energy Foundation entity, now Energy Foundation China ($13.4M).

3. Half the dollars file to California addresses, and annual giving is shrinking. [DATA] California-addressed recipients received 49.5 percent of all giving, an unusually deep home-state concentration for a foundation of this scale (Gates's home-state share is 6 percent). Half the California total, though, is the family institutions, and another fifth is nationally focused platforms headquartered in the state; the place-based share is closer to a sixth of giving (see U.S. Geography). At the same time, total grants paid (the 990-PF cash-basis figure, which can differ from the foundation's published award totals) fell 19 percent across the window, from $432.1 million in 2022 to $350.0 million in 2024, with the grant count down 11 percent.

4. Packard is a renewing, mid-size-grantee funder. [DATA] The 302 recipients funded in all three years are 22.8 percent of recipients but received 59.8 percent of the dollars. The median grant is a stable $150,000, and the modal grant lives in the $100K-to-$500K band (55 percent of all grants). The recipient base skews small and mid-size: 72 percent of matched recipients report under $25 million in annual revenue. This is the layer of the nonprofit economy that Gates-scale grant sizes skip past.

5. The 2023 program restructure is visible in the filings. [DATA] [INFERRED] Packard's grant-purpose labels change cleanly between tax years 2022 and 2023: eight named 2022 program lines give way to four umbrella programs (Environment and Science, Just Societies, Families and Communities, Special Initiatives). Six of the eight consolidate legibly (Conservation and Science; Reproductive Health; Children, Families, and Communities; Local Grantmaking; Democracy, Rights & Governance; Justice and Equity); two (Special Opportunities; the President's Fund) disappear with no single labeled successor. The Organizational Effectiveness line, a long-running capacity-building program, collapses from $15.4 million to $1.2 million in 2024, a wind-down in plain sight.

---

§02 · Findings by Dimension

Scale and Giving

[DATA] The Packard Foundation held $8.55 billion in assets at the end of tax year 2024 and recorded qualifying distributions of $442.6 million, 5.2 percent of assets. Grants paid totaled $432.1 million (2022), $367.0 million (2023), and $350.0 million (2024), $1.15 billion across the window. Extracted grant totals reconcile to 100 percent of the official Part I grants-paid figure in all three years.

Annual grant statistics · 2022–2024

[DATA] The pooled median across all 3,284 grants is $150,000, essentially unchanged year to year. The mean runs 2.2 to 2.5 times the median, the signature of a long right tail.

[DATA] [INFERRED] Total grants paid fell 19 percent across the window while the median grant held flat, meaning the contraction came through fewer grants and a thinner large-grant tail rather than smaller checks. The single largest grant each year (always MBARI) actually grew.

Recipient-type reconciliation · $1.149B total [DATA]

Findings keyed to U.S. tax classifications (sector, maturity, revenue) describe the 501(c)(3) subset only; findings keyed to the foundation's own grant-purpose text (program lines, gift sizing, big bets) cover the full portfolio.

The Institutions the Family Built

[DATA] [REPORTED] Three recipients with direct Packard-family origins form an institutional portfolio that sits alongside, and should be sized apart from, the external grantmaking program. The filings abbreviate them as MBARI, LPFCH, and MBAQ; those abbreviations are kept where tables quote purposes as filed.

[DATA] MBARI's grants are few and enormous: $76.3 million (2022, two grants), $68.1 million (2023, two grants), and $54.9 million (2024, one grant). Every grant is labeled simply "MBARI." The three largest grants in the portfolio, all above $50 million, are these annual MBARI commitments, and they alone carry 13.9 percent of total giving.

[INFERRED] This is structurally different from ordinary grantmaking. MBARI is an operating research institute whose budget the foundation effectively underwrites year after year; the commitment functions as an internal institutional obligation rather than a competitive grant. Any read of Packard's "available" grant capital should start by setting aside roughly a quarter of the headline figure.

Program Architecture and the 2023 Restructure

[DATA] Packard's grant-purpose text consists almost entirely of program labels (21 distinct labels cover 100 percent of dollars), which permits a clean program-level decomposition that most foundations' filings do not support. The labels change abruptly between tax years 2022 and 2023.

Tax year 2022 · the old architecture ($432.1M)

Rows are rounded to $0.1M; the unrounded rows sum exactly to the $432.1M year total.

Tax years 2023–2024 · the new architecture ($717.0M)

[INFERRED] The mapping is legible: Conservation and Science became Environment and Science; Reproductive Health, Democracy, Rights & Governance, and Justice and Equity consolidated under Just Societies; Children, Families, and Communities absorbed Local Grantmaking to become Families and Communities. Special Opportunities and the President's Fund disappear as labels with no single labeled successor. Three structural signals stand out. The Just Societies umbrella grew 15 percent in its second year while Special Initiatives shrank 70 percent, suggesting the latter was a transition vehicle. The Organizational Effectiveness wind-down ($15.4M to $1.2M) ends a dedicated capacity-building line; recipients lose a distinct door into the foundation. And the Fellowships line, Packard's signature early-career science program, declined steadily ($18.5M to $15.7M to $11.7M).

Recipient-Address Geography

[DATA] Packard's giving is the inverse of a globally distributed portfolio: 88.9 percent of dollars file to U.S. recipient addresses, and 49.5 percent to California ones. Address geography describes where recipients are headquartered, not the ultimate use of funds; for the intermediary layer, ultimate geography is not observable in the filings.

Domestic vs international distribution

International recipients · by country

[DATA] [INFERRED] Indonesia is the largest international destination by a wide margin, with 189 grants, and the second tier features coastal and fisheries nations (Mexico, Japan, Chile). This is the footprint of an oceans and marine-conservation program delivered in-country through many grants averaging about $176,000, not a multilateral co-financing pattern. The Netherlands figure is dominated by a single recipient, Stichting European Climate Foundation ($7.4M), a European climate regranting platform. The tail below the top seven brings the international portfolio to 36 countries in total. The main non-oceans line is reproductive-health delivery in East Africa, led by Ethiopia.

U.S. Geography

[DATA] Within the United States, the portfolio is built around three poles: home-state California, the Washington, DC intermediary cluster, and New York philanthropy infrastructure.

Top U.S. jurisdictions by dollars

[DATA] The top three jurisdictions carry 72.9 percent of all giving; 1,144 grants went to California recipients, 35 percent of all grants.

[DATA] [INFERRED] The California total should not be read as place-based focus at face value. It decomposes into $277.0 million to the family institutions (48.7 percent of the state total), $106.0 million to nationally and globally focused platforms headquartered in the state (ClimateWorks, Resources Legacy Fund, the two Energy Foundation entities, Multiplier, Liberty Hill), and roughly $186 million, about 16 percent of all giving, to other California-based recipients, a figure that itself includes universities and the globally oriented Climate Breakthrough awards program. The genuinely place-based California portfolio is therefore on the order of a sixth of giving, not half, though that is still several times Gates's entire home-state share.

Who Drives Each State Total

[DATA] [INFERRED] Top recipients in each leading state, three-year totals:

• California · $568.8M: MBARI ($199.3M), Lucile Packard Foundation for Children's Health ($60.0M), ClimateWorks ($33.9M), Climate Breakthrough ($23.9M), Resources Legacy Fund ($18.2M).
• Washington, DC · $148.6M: Hopewell Fund ($8.1M), New Venture Fund ($8.1M), Windward Fund ($6.2M), the American Association for the Advancement of Science ($4.0M), DKT International ($3.9M).
• New York · $119.9M: Rockefeller Philanthropy Advisors ($42.5M), RF Catalytic Capital ($25.0M), Planned Parenthood Federation of America ($6.1M), NEO Philanthropy ($2.7M), Guttmacher Institute ($2.7M).
• Mississippi · $12.4M: Hope Enterprise Corporation ($1.5M), Mississippi Coalition on Black Civic Participation ($1.2M), Mississippi Votes ($1.1M), Women's Foundation of Mississippi ($1.1M), Converge ($1.1M).
• Louisiana · $12.2M: Tulane ($2.4M), Institute of Women & Ethnic Studies ($1.4M), Power Coalition for Equity and Justice ($1.0M), Deep South Center for Environmental Justice ($0.8M).

[INFERRED] The DC and New York totals are not regional giving at all; they are intermediary infrastructure. Every leading DC recipient is a pooled fund or fiscal sponsor, and the New York list is led by two philanthropy-services platforms. By contrast, Mississippi and Louisiana are genuine place-based portfolios: dozens of grants averaging around $250,000 to civic-participation, women's, and community organizations across the Gulf South, a regional concentration that has no parallel elsewhere in the portfolio outside California.

Issue-Area Concentration

[DATA] Because Packard's purpose text is program labels rather than descriptive sentences, keyword clustering detects only two themes, and each maps exactly onto program lines. Clusters overlap by design elsewhere in this series; here they do not, because the labels are disjoint.

[DATA] The climate-and-environment figure is exactly the sum, at full precision, of the two environment program lines (Conservation and Science 2022, $109.8M; Environment and Science 2023–24, $209.3M). The reproductive-health figure is the 2022 program line alone; after the restructure, that work continues inside Just Societies ($147.7M across 2023–24) where the label no longer isolates it.

[INFERRED] A keyword read therefore understates Packard's thematic concentration. On a program-line basis, environment and science (including MBARI and the science fellowships) carried roughly $564 million, just under half of all giving, and the Just Societies and reproductive-health lines together carried roughly $213 million.

Sector Comparison Against Base Rates

[DATA] The figures below describe the NTEE-coded portion of the portfolio: $712.0M, 62 percent of total giving. Of the $974.0M resolved to a registered U.S. nonprofit, $262.0M sits with recipients that carry no NTEE code in the registry, and MBARI alone is $199.3M of that, so the foundation's single largest recipient is absent from this table. Grants resolved only to a university record and not to a registered nonprofit ($12.4M) are likewise outside it, which modestly understates Education (B).

NTEE major code · Packard weighting vs sector base rate

"Packard %" is the dollar share among NTEE-coded matched dollars ($712.0M). "Sector %" is the share of $1M+ revenue 501(c)(3)s in each category by organization count. The comparison is a rough alignment indicator, not a strict like-for-like ratio.

[DATA] [INFERRED] Two readings stand out. The single largest matched category is T, philanthropy and grantmaking, at 18.6 percent of NTEE-coded dollars: nearly one classified dollar in five goes to other grantmaking entities, the intermediary pattern quantified below. And the U over-index (8.0×) is driven almost entirely by one recipient: ClimateWorks ($33.9M of the category's $41.5M), which is NTEE-coded as science and technology but functions as a climate pooled fund. Mental health registers zero matched dollars; housing, religion, and recreation barely register.

Gift Sizing and the Big-Bet Tail

[DATA] The median grant is $150,000 and the mean is about $350,000. The modal grant by count sits in the $100K-to-$500K band; the dollars are bimodal, with 31 percent in that working band and 48 percent in grants of $1 million or more.

[DATA] The 24 grants of $5 million or more (0.7 percent of grants) carry 33.7 percent of all dollars. The three grants above $50 million, all MBARI, carry 13.9 percent on their own.

[DATA] [INFERRED] Read against Gates, the shape is a compressed barbell. Gates's median is $500,000 with 43 percent of dollars in $5M+ grants spread across dozens of institutional platforms; Packard's median is $150,000 with a $5M+ tier that is mostly two family institutions plus a handful of climate intermediaries. Outside that tail, Packard is a $100K-to-$500K grantmaker.

The 24 Big Bets

[DATA] All 24 grants of $5 million or more in the three-year window:

[INFERRED] The tail decomposes cleanly. The family institutions (MBARI, LPFCH, the Aquarium Foundation) account for 10 of the 24 grants; climate intermediaries and platforms (ClimateWorks, the Energy Foundation, Climate Breakthrough, Windward, the European Climate Foundation) account for 9; and philanthropy-services vehicles (Rockefeller Philanthropy Advisors, RF Catalytic Capital) account for 4. Planned Parenthood's single $5.0 million grant is the only big bet to an operating nonprofit outside the family institutions and the intermediary layer.

Largest Partners

[DATA] The top recipients by total dollars received across 2022 to 2024, deduplicated by matched entity (ID-first; see Recipient matching in §05), which merges affiliated filers such as UC Berkeley and its foundation:

[DATA] The top 10 recipients absorbed 38.6 percent of all dollars. Excluding the three family institutions, the leading recipients are nearly all intermediaries: philanthropy-services platforms and climate pooled funds rather than operating organizations.

Top University Recipients

[DATA] University grants total $86.6 million, 7.5 percent of giving, spread across many institutions in $100K-to-$500K grants. Three-year totals:

[DATA] [INFERRED] No university approaches MBARI's scale; the largest university relationship (UC Berkeley, $9.6M, merging the campus and its affiliated foundation) is under 5 percent of the MBARI total. The many-small-grants pattern, with 37 grants to Berkeley averaging about $260K, is consistent with the Packard Fellowships and project-level science funding rather than institutional capital. The foundation's real research bet is its own institute.

Recipient Profile

[DATA] Maturity. The median formation year among matched recipients is 1996 (865 of the 867 matched recipients have a known formation year). This section keys directly to matched nonprofit records, so it was already at the entity level; the dedup-key correction described in §05 does not move these figures. The distribution by era:

[DATA] Scale. By revenue, the portfolio skews to small and mid-size organizations: 99 matched recipients report under $1 million in annual revenue, 269 report $1 million to $5 million, and 255 report $5 million to $25 million. Together that is 72 percent of matched recipients under $25 million. Only 35 recipients report over $500 million.

[DATA] [INFERRED] This is the sharpest single contrast with Gates, whose matched portfolio concentrates in nine-figure institutions. Packard's working capital flows to organizations in the $1M-to-$25M revenue band, the operating scale of many regional conservation groups, reproductive-health providers, and civic-participation organizations. The 2010-and-later cohort (244 organizations, 28 percent of matched recipients) shows the foundation continues to take positions in young organizations, typically at $150K-scale checks.

Repeat-Recipient Structure

[DATA] Funding concentration by years funded

[DATA] [INFERRED] The 302 recipients funded in all three years received 59.8 percent of the dollars. The concentration is meaningful but gentler than Gates (30 percent of recipients holding 83 percent of dollars): Packard renews a core, but roughly half its recipients in any window are single-year, and those single-year relationships still carry a real 16 percent of dollars. There is more room at the edge of this portfolio than at the edge of Gates's.

One caveat on the one-year cohort: the three-year window cuts off recipient histories at both edges. An organization funded in 2021 and again in 2025 would appear here as a single-year recipient. The 49.0 percent single-year share is an upper bound on genuinely one-time giving; the concentration finding is conservative.

Intermediary and Regrantor Channels

[DATA] At least 23.6 percent of matched dollars ($230.2M of $974.0M) went to intermediaries: organizations coded as grantmakers, flagged conduits (donor-advised funds, fiscal sponsors, flow-through vehicles), or curated pooled funds. The $974.0M denominator is every grant resolved to a registered U.S. nonprofit; it is broader than the reconciliation table's $898.5M 501(c)(3) line because it retains the $74.2M of university-bucket grants that also match a nonprofit record and $1.4M of matched grants with foreign addresses. The share is a lower bound; the curated pooled-fund list is not exhaustive, and 990-PF data cannot trace where these vehicles subgrant. The largest channels, with their filed purposes:

[REPORTED] The two Energy Foundation rows are legally separate organizations, not a name variant. The original San Francisco entity (EIN 94-3126848) retained the legacy name "The Energy Foundation" after a separation completed in 2020 and now operates as Energy Foundation China; the United States Energy Foundation (EIN 83-1740146) is the successor entity for the U.S. programs. Packard helped establish the China program in 1999. [4][5][6]

[INFERRED] On that history, the $13.4M to the legacy entity is best read as China-focused climate regranting.

[DATA] [INFERRED] Setting aside the employee-match vehicle, the channel list maps onto two strategies. The climate intermediaries (ClimateWorks, the two Energy Foundation entities, Windward, Multiplier, and, abroad, the European Climate Foundation) are field-level pooled funds through which Packard co-finances climate work alongside peer funders; this is how a $350M-a-year foundation participates in a problem whose financing runs to billions. The philanthropy-services platforms (Rockefeller Philanthropy Advisors, RF Catalytic, Hopewell, New Venture Fund) carry the politically sensitive Just Societies and reproductive-health portfolios, where pooled vehicles offer administrative reach and, plausibly, a degree of separation. For a prospective grantee in either field, the practical implication is the same: a meaningful share of Packard's money in your field is accessed through these named intermediaries, not through Packard directly.

[DATA] [INFERRED] The channels' own filings show where the layer deploys. Across the ten largest intermediaries (the channel table above, excluding the employee-match vehicle and the smallest entry, Liberty Hill), onward grants in the same window totaled $5.4 billion (all of their funders' money pooled, not Packard's alone): 24.4 percent went to foreign recipients, with Washington, DC (17.5 percent) and California (13.3 percent) the largest U.S. destinations. The legacy Energy Foundation entity's own grants are 95 percent foreign, consistent with the China reading above, and Resources Legacy Fund is the only California-weighted channel (40 percent in-state). Because pooled vehicles commingle many funders, this describes the layer's overall deployment, never the path of Packard's specific dollars.

[DATA] For series context, the equivalent lower-bound figure, computed from the same dataset and definitions, is 6.1 percent for Gates and 25.4 percent for Ford: Packard sits at the intermediary-heavy end of large-foundation practice, alongside Ford.

Mission Language Patterns

[DATA] Packard's purpose text is program labels, so term frequencies reconstruct the program architecture rather than describing grant content. Top substantive terms (word boundaries applied; generic terms excluded):

[DATA] Caveat: paired terms ("just" and "societies"; "organizational" and "effectiveness") return identical figures because they come from single recurring program labels; they are one funding line each, not two.

[DATA] The word "climate" appears in zero grant purposes. The foundation's dominant theme, carrying $319.0 million, is filed entirely under "Environment and Science" and "Conservation and Science" program labels; the word itself never appears in the purpose text. Searches for "artificial intelligence" and "racial" also return zero grants.

[INFERRED] This is the clearest illustration of why purpose-text analysis must be read against filing style. A keyword screen for "climate" run naively across 990 data would conclude Packard does not fund climate work, while it is in fact a major climate funder. Packard's filings describe its bureaucracy, not its grantees' work.

---

§03 · Practical Implications

[INFERRED] This analysis identifies portfolio composition, not selection criteria. The considerations below describe alignment with the demonstrated portfolio.

Strongly represented profiles

• Ocean, marine-science, and conservation organizations, especially in California and the Pacific
• Climate organizations reachable through the intermediary layer (ClimateWorks, the United States Energy Foundation, Climate Breakthrough, Windward, Multiplier)
• Reproductive-health organizations, domestically and in East Africa, often via philanthropy-services platforms
• Civic-participation, women's, and community organizations in California and the Gulf South (Mississippi, Louisiana)
• Early-career science: the Packard Fellowships line, though it is shrinking
• Small and mid-size organizations ($1M to $25M revenue) able to put $100K-to-$500K annual grants to work
• Organizations positioned for multi-year renewal: the three-year core carries 60 percent of dollars

Underrepresented profiles

• Mental health (0.0× the sector base rate, zero matched dollars), housing (0.1×), religion-related (0.1×), recreation (0.1×)
• Arts and culture (0.4×) and direct human services (0.5×)
• Education organizations seeking institutional-scale support (0.6×; university grants average under $250K)
• Organizations outside California, DC/NY intermediary channels, the Gulf South, or the oceans-program countries
• Anyone seeking the $5M+ tier: it is effectively reserved for family institutions and named climate platforms

For organizations seeking entry [DATA] [INFERRED] [REPORTED] Part XV does not identify the foundation as preselected-only (it does not check the "contributes only to preselected organizations" box), though the filing itself says nothing about how often unsolicited proposals lead to grants. The foundation's own grantseeker guidance reports that in a typical year about 15 percent of grants go to first-time grantees and less than 1 percent originate as unsolicited proposals; as of this writing it lists no open requests for proposals. [7] The demonstrated paths into the portfolio are three: a program-fit grant in the $100K-to-$500K working band, where most of the foundation's relationships live; subgrantee or partner status within one of the named intermediaries that carry the climate and Just Societies portfolios; or the place-based portfolios in California and the Gulf South. The contraction in grants paid (down 19 percent across the window) and the wind-down of the Organizational Effectiveness line both argue for entering through current program priorities rather than capacity-building or legacy doors.

---

§04 · Limits of the Analysis

Selection criteria. Portfolio composition reflects both deliberate strategy and existing relationships. This analysis cannot distinguish between what was chosen and what was already in the funnel. The 2023 restructure means the 2022 program labels describe an architecture the foundation has explicitly moved away from.

Probability of selection. A nonprofit cannot infer odds of Packard funding from this analysis. The portfolio captures roughly 1,324 funded organizations against an eligible universe in the hundreds of thousands, and a quarter of the dollars are pre-committed to family institutions before any external applicant is considered.

Causation. Correlation between organizational characteristics and selection is not evidence of preference. The Gulf South concentration may reflect a deliberate regional strategy or the networks of specific program staff; the data is consistent with both.

The unmatched portion. Sector, maturity, and revenue findings describe only the matched 501(c)(3) subset ($712.0M of NTEE-coded dollars, 62 percent of giving). International recipients (11.1 percent), universities (7.5 percent), and for-profit/other entities (3.2 percent) are absent from analyses keyed to U.S. tax classifications.

Purpose-label terseness. Packard's purpose text is program labels. This makes program attribution exact but theme detection conservative: work on, say, climate or racial equity inside an umbrella label is invisible to keyword analysis. The program-line decomposition is the more reliable lens for this foundation; the keyword clusters are a floor, not an estimate.

The three-year window. Recipient histories are truncated at both edges of the 2022-to-2024 window. The repeat-recipient concentration finding is conservative, and multi-decade relationships (MBARI's history with the foundation spans nearly four decades) appear here only as their most recent three years.

Intermediary opacity. At least 23.6 percent of matched dollars route through regrantors whose onward grants are invisible in Packard's 990-PF; the channels' own Schedule I filings (summarized in §02) show where the layer as a whole deploys, but which dollars are Packard's cannot be traced. Findings about recipient maturity, revenue, and geography describe the intermediaries themselves, not the organizations they ultimately fund.

---

§05 · Methodology

Source data. Grants extracted from the David and Lucile Packard Foundation's IRS Form 990-PF Schedule I filings for tax years 2022 to 2024, accessed via the SciRise Funding Intelligence dataset. Extracted annual totals reconcile to 100 percent of the official Part I grants-paid figure in all three years (3,284 grants totaling $1.149 billion).

Recipient matching. Each grant recipient was resolved against the full U.S. 501(c)(3) population by EIN where possible. Recipients were deduplicated ID-first: by matched university record, then matched nonprofit record, with a case-normalized name-plus-jurisdiction key used only for entities outside those registries (government agencies, international recipients, unmatched filers). Recipient-type splits assign each grant to exactly one bucket (international by country code; then university; then 501(c)(3); then other), so the reconciliation table sums to the portfolio total without double-counting.

Entity resolution. An earlier name-keyed dedup split 52 matched recipients across filed name variants (such as a dropped "Inc"), covering $62.2M (5.4 percent of dollars). All aggregates in this report use the ID-first entity key described above, which resolves those splits; recipient counts, tenure buckets, and ranked tables are at the entity level throughout. The name-based key remains in use only for recipients that do not exist in our registries, where distinct filed names can still represent one organization.

Program-line decomposition. Packard's grant-purpose text consists of 21 distinct program labels covering 100 percent of dollars. Program tables aggregate the labels as filed, normalized for case only. The 2022-to-2023 label transition is reported as filed; the mapping between old and new programs is an inference from dollar and grant continuity.

NTEE classification and base rate. Sector shares use NTEE major categories for matched U.S. recipients. The base-rate denominator is the population of U.S. 501(c)(3) organizations reporting $1M+ revenue in their most recent filing. Foundation share is measured by dollars; sector share by organization count.

Intermediary share. Intermediaries are recipients coded NTEE-T (philanthropy/grantmaking), flagged conduits (donor-advised funds, fiscal sponsors, flow-through vehicles), or curated pooled funds (program-area-coded funder collaboratives such as ClimateWorks and New Venture Fund). The share is a lower bound on routed-through-intermediary giving.

Intermediary onward deployment. For the largest intermediary recipients, we separately analyzed their own public grant-recipient filings (IRS Form 990 Schedule I) over the same tax-year window. These figures describe the intermediaries' total onward grantmaking across all funders, not the path of Packard-specific dollars.

Country and state attribution. Country and state attributions follow the IRS filing's address fields. A headquarters address reflects institutional location, not necessarily where the work happens; this matters most for California, where several nationally and globally focused platforms are headquartered.

Rounding. Dollar figures are displayed to $0.1M; component rows can therefore differ from a stated total by up to $0.1M. All sums reconcile exactly at full precision.

Evidence tags. [DATA]: claim directly supported by counts or figures derived from the underlying filings. [INFERRED]: pattern interpreted from data requiring analytical judgment; alternative explanations exist. [REPORTED]: institutional-affiliation and corporate-history facts drawn from the named organizations' public materials and standard nonprofit registries.

Sources

1. MBARI. "History." https://www.mbari.org/about/history/
2. The David and Lucile Packard Foundation. "MBARI: Innovation Beneath the Waves, Inspiration on Shore." https://www.packard.org/insights/grantee-story/mbari-innovation-beneath-the-waves-inspiration-on-shore/
3. Lucile Packard Foundation for Children's Health. "About." https://lpfch.org/about/
4. GuideStar. "Energy Foundation China" (EIN 94-3126848). https://www.guidestar.org/profile/94-3126848
5. GuideStar. "United States Energy Foundation" (EIN 83-1740146). https://www.guidestar.org/profile/83-1740146
6. InfluenceWatch. "Energy Foundation China (EFC)." https://www.influencewatch.org/non-profit/energy-foundation-china/
7. The David and Lucile Packard Foundation. "Funding Opportunities." https://www.packard.org/grantees/funding-opportunties/ (accessed June 2026).

---

§06 · Reference

About the Foundation

---

This is an independent analysis. Findings reflect public 990 filings and the cited institutional materials only, and do not represent any communication from the David and Lucile Packard Foundation. Errors and interpretations are the author's.

End of report · 2026