What these portfolios tell us

As of 17 September 2026, the research base contains 17,277 company–program records and a separate profile for every one of the 19 entries. This is a broad public-source research base, with material gaps documented in coverage.md, rather than an all-history investment census.

The most useful distinction is the kind of company each organization is equipped to help and the evidence it expects at entry. The post’s “top tier / AI / deep tech” labels are too coarse to decide where a startup should apply. Portfolio resemblance is evidence about fit, not a probability of acceptance.

Practical findings

AI is a starting point, not a sufficient match. Speedrun’s current public descriptions frequently identify AI and interactive/media products; Founders Inc. also has a substantial hardware/building pattern. AI Grant provides a direct route for AI products, while AI Fund requires evaluating a venture-studio relationship. An enterprise workflow tool, an AI game and an industrial robot should get different shortlists even when all use the same models.

Broad accelerators require a sharper company argument. YC and Techstars span many sectors. Being in a sector they have funded is weak evidence on its own. Customer urgency, team execution, differentiation and a credible path to a large business should dominate. Techstars requires a second step: choose a current individual program whose theme, geography and partner needs actually fit.

Deep technology has different proof requirements. HAX and IndieBio are useful analogues for real physical or biological innovation. Technical experiments, IP, lab/manufacturing milestones and capital needs matter more than calling the company “AI-powered.” SOSV’s historical software/crypto investments are not evidence of current appetite: its present guidance focuses on physical and biological deep technology. Current SOSV guidance.

Activate and Greentown answer a different funding/support question. Activate supports scientists through a non-dilutive fellowship with specific eligibility conditions; Greentown is a climate incubator and member community. They may be excellent choices for a startup without being evidence that an equity investor will fund it. Activate’s healthcare-only restrictions also make “biotech” an insufficient eligibility test. Activate criteria, Greentown members.

Founder circumstances can change the ranking substantially. SPC can be appropriate before a settled idea, HF0 currently emphasizes repeat founders, and PearX focuses on pre-seed teams. A product-only brief cannot fully assess these routes. Current program fit gets greater weight than a superficial product resemblance. SPC, HF0, PearX.

Support programs should be considered alongside the fundraising shortlist. Google’s regional accelerators, NVIDIA Inception and Microsoft for Startups can supply technical support and benefits without making an equity investment. Membership does not establish investment by Google, NVentures or M12. Their public examples are especially subject to marketing selection bias.

Measured portfolio signals

These are transparent keyword tags from product descriptions plus source sector labels, not a hand-verified classification of every company. They are nonexclusive: one company can have several themes. The denominator is records with a description or sector label in that program’s analysis scope; network-only Techstars rows are excluded. A missing keyword means “not detected,” not “not in this sector.” Comparisons are affected by different directory taxonomies, historical windows, pivots and missing data. Do not treat the percentages as allocation targets or acceptance probabilities.

Percentages are deliberately omitted here for sparse/selectively collected HF0, Arc, PearX, NVIDIA and Microsoft samples. AI Grant and AI Fund are AI-focused by mandate even when a short company description does not contain an AI keyword. SOSV’s parent row overlaps its two child rows.

Concrete portfolio contrasts

These are illustrative precedents from the collected records, not proof of what an investor will fund next. Product descriptions are current/source-reported rather than reconstructed at admission.

Each example is searchable in the explorer and linked to source evidence in the dataset. Funding amounts for these individual companies were not inferred from current program headline checks.

Current program profiles

The following combines checked terms with analyst interpretation of fit. Headline checks are not guaranteed offers; credits, staged commitments and equity instruments are different. Terms and intake should be rechecked when evaluating a real applicant.

Y Combinator

Route: Equity investor. Funding/support: $500K: $125K for 7% plus $375K on an uncapped MFN SAFE.

Fit interpretation: Broad early-stage technology; assess technical execution, customer demand, speed and venture-scale potential.

What to check: Check current batch timing, founder commitment and participation requirements.

Sources: Official reference 1, Official reference 2.

a16z Speedrun

Route: Equity investor. Funding/support: Up to $1M advertised; credits separate. Headline is not a guaranteed check or a full description of dilution.

Fit interpretation: AI applications, infrastructure, consumer products, games and interactive experiences; strong product-building evidence matters.

What to check: Verify cohort terms and location for the upcoming intake.

Sources: Official reference 1, Official reference 2.

Techstars

Route: Equity investor. Funding/support: $220K standard offer: $20K for 5% common equity through a CEA plus $200K uncapped MFN SAFE. Asia-Pacific offer differs ($120K total).

Fit interpretation: Broad sectors; select the individual location, corporate partner and thematic accelerator, not only the Techstars brand.

What to check: Program-specific geography, cohort availability and industry restrictions; recheck final terms.

Sources: Official reference 1, Official reference 2.

Founders, Inc.

Route: Equity investor. Funding/support: $100K–$250K first checks on portfolio page; individual deals may vary.

Fit interpretation: Technical builders across AI, developer products, consumer, robotics and hardware; a working demonstration is particularly informative.

What to check: Assess willingness to use the builder community and program facilities; not Founders Fund.

Sources: Official reference 1, Official reference 2.

Sequoia Arc

Route: Equity investor. Funding/support: Current Arc is an open call for early Sequoia partnership; investment terms vary. A fixed $1M deal is not verified.

Fit interpretation: Ambitious early-stage companies with differentiated insight, a large opportunity and convincing founder-market fit.

What to check: Separate historical eight-week cohorts from the current Arc open-call/Intensive format.

Sources: Official reference 1, Official reference 2.

South Park Commons

Route: Equity investor. Funding/support: $400K for 7% upfront plus $600K in the next external round; $1M is not all upfront.

Fit interpretation: Technical founders exploring ideas before product-market fit; frontier ambition across AI, chips, robotics, biology and software.

What to check: Eight-week in-person bootcamp at an SPC office. Residency without funding is a different path. Fall 2026 application deadline has passed.

Sources: Official reference 1, Official reference 2.

HF0

Route: Equity investor. Funding/support: Current standard investment terms could not be verified from the official pages; do not rely on the post’s $1M figure.

Fit interpretation: Current homepage emphasizes repeat founders and an intensive technical building environment.

What to check: A similar product alone does not establish fit; founder history and willingness to participate in residency are central.

Sources: Official reference 1, Official reference 2.

PearX

Route: Equity investor. Funding/support: $500K–$2M on current PearX page; older materials give different ranges.

Fit interpretation: Pre-seed teams needing close help with customers, hiring and fundraising; software, AI and selected scientific businesses.

What to check: 12-week San Francisco program. Most enter before fundraising; some have raised up to $2M. W27 page lists October 4 regular deadline.

Sources: Official reference 1, Official reference 2.

500 Global Flagship

Route: Equity investor. Funding/support: The post’s $150K offer was not reverified for the current cohort; confirm investment, equity and any program fee together.

Fit interpretation: Broad scalable businesses; customer acquisition, business model and growth evidence should drive comparisons.

What to check: Do not attribute all regional 500 Global investments to Flagship. Confirm the target cohort and participation requirements.

Sources: Official reference 1, Official reference 2.

AI Grant

Route: Equity investor. Funding/support: $250K on an uncapped, no-discount MFN SAFE; credits are additional benefits.

Fit interpretation: AI-native products with useful demonstrations; product insight and execution beyond generic model access.

What to check: Seed/pre-seed; Delaware incorporation needed for investment. Last published batch applications closed; recheck next opening.

Sources: Official reference 1.

AI Fund

Route: Venture studio / investor. Funding/support: Company-specific venture studio/cofounding arrangements; no universal $1M check verified.

Fit interpretation: Applied AI companies where technical capabilities, a concrete business problem and studio collaboration can create value.

What to check: Evaluate willingness to co-build, founder role, company maturity and ownership arrangement; not a standard accelerator application.

Sources: Official reference 1, Official reference 2.

NVIDIA Inception

Route: Support, not equity investment. Funding/support: Free program; no equity taken. Benefits and credits are not a venture investment.

Fit interpretation: Technical startups that benefit from NVIDIA software, computing and ecosystem support.

What to check: Incorporated, under ten years old, website and at least one developer; excluded business categories apply. NVentures is separate.

Sources: Official reference 1.

Microsoft for Startups

Route: Support, not equity investment. Funding/support: Up to $150K in credits advertised; eligibility and benefit levels vary.

Fit interpretation: Software/AI companies benefiting from Azure, enterprise tooling and go-to-market support.

What to check: Assess program and benefit eligibility separately from fundraising. M12 investment is not established by membership.

Sources: Official reference 1.

Google for Startups Accelerator

Route: Support, not equity investment. Funding/support: Equity-free accelerator support; credits and benefits depend on track.

Fit interpretation: AI-enabled or technology businesses with a meaningful technical scaling problem and regional program fit.

What to check: Apply to a particular region/theme. Revenue, geography and seed-to-Series-A/B requirements vary; old alumni are not all AI-track alumni.

Sources: Official reference 1, Official reference 2.

SOSV

Route: Equity investor / parent. Funding/support: Current stated initial investment up to $550K, not the post’s $525K.

Fit interpretation: Current strategy emphasizes deep technology in physical and biological sciences. Historical software/crypto holdings do not prove current appetite.

What to check: Typically first VC, early technical teams; usually four-to-six-month US program. Current guidance excludes ordinary software, fintech, crypto and similar businesses.

Sources: Official reference 1, Official reference 2.

SOSV SF & NY (IndieBio)

Route: Equity investor / SOSV program. Funding/support: SOSV terms are the parent reference; do not count this as a second check on top of SOSV.

Fit interpretation: Biotechnology, bioindustrial processes, food, materials and life-science platforms with experimental substance.

What to check: Assess scientific feasibility, IP access, lab needs, milestones, regulation and US program participation.

Sources: Official reference 1, Official reference 2.

SOSV HAX

Route: Equity investor / SOSV program. Funding/support: SOSV terms are the parent reference; company-specific economics require confirmation.

Fit interpretation: Hardware, robotics, advanced manufacturing, energy, materials and other hard technology.

What to check: A real physical technology and clear technical milestones are stronger evidence than an AI label; verify facilities and participation fit.

Sources: Official reference 1, Official reference 2.

Greentown Labs

Route: Support / incubator membership. Funding/support: Membership, workspace, labs and commercialization support; portfolio listing does not establish a cash investment.

Fit interpretation: Climate and cleantech companies benefiting from prototyping, labs, industrial partners and community.

What to check: Assess climate impact, technical readiness, physical workspace needs and membership costs/locations.

Sources: Official reference 1, Official reference 2.

Activate

Route: Non-dilutive fellowship. Funding/support: US fellowship advertises more than $300K over two years across stipend, R&D and other support; no equity.

Fit interpretation: Research-based hard technology that needs substantial technical de-risking before commercial scale.

What to check: US track: full-time, US work eligibility, bachelor’s plus four years of relevant experience, first hard-tech commercialization, and no more than $2M private debt/equity at deadline. Healthcare-only/FDA pathways generally excluded; software exceptions narrow.

Sources: Official reference 1, Official reference 2.

How to use this with your unnamed startup

Send a description using startup-brief.md, or write it conversationally. I will assess all 19 routes, rank the most credible investment candidates, cite close portfolio examples, explain where the analogy breaks, identify likely reasons to pass, and give the milestones that would improve the case. Support-only opportunities will appear separately.

The assessment protocol is saved in evaluation-method.md. It uses qualitative likelihood and explicit confidence. Numerical acceptance probabilities would require rejected-applicant and admission-time data that this research does not contain.