# Two handshakes. Who funds the agent stack, and whether it shows in the tools

> Quynh Tran-Thanh on the investor graph behind 696 funded AI tool companies on Anchor Terminal. Most of them sit in one network two steps apart, NVIDIA funds half the inference providers that buy its chips, the same investors back both sides of most races, founders fund each other and sometimes their rivals, one line runs from OpenAI's startup fund to SpaceX through Cursor, and the number of investors barely moves a tool's grade.

- Canonical: https://www.anchorterminal.com/blog/two-handshakes-who-funds-the-agent-stack
- Markdown: https://www.anchorterminal.com/blog/two-handshakes-who-funds-the-agent-stack.md (~4,150 tokens)
- Slim: https://www.anchorterminal.com/blog/two-handshakes-who-funds-the-agent-stack.min.md (~1,980 tokens, same facts, less prose, for token-sensitive contexts)
- JSON: https://www.anchorterminal.com/blog/two-handshakes-who-funds-the-agent-stack.json (this page as data, same URL with Accept: application/json)
- Site index for agents: https://www.anchorterminal.com/llms.txt (the key docs and every listing in one file, at most 16 MB, about 4 million tokens: https://www.anchorterminal.com/llms-full.txt)
- API: https://www.anchorterminal.com/api/v1/index.json
- MCP: https://www.anchorterminal.com/mcp (the whole catalogue as tools, no key)
- DNS: dig +short TXT <slug>.grade.anchorterminal.com (the grade as a DNS TXT record, for a sandbox with no HTTP where names still resolve; also top., best.<category>. and <a>.vs.<b>. under the zone, and the same under grade.letme.dev)
- Licence: CC BY 4.0, credit Anchor Terminal and link this page (https://www.anchorterminal.com/terms/#data)
- Updated: 2026-11-15

By Quynh Tran-Thanh, 15 November 2026. 1,593 words, about 7 min to read.

We just put up a new page, [Who backs whom](/graph/). It joins every company behind a listed tool to the investors named in its funding rounds and to the founders on its listing, with a source on every link. I wanted to know whether it was a pile of logos or whether it said something, so I spent a day pulling it apart. It says quite a lot.

The graph has 696 companies with a sourced round, 3,180 investors and 1,442 founders and staff [[1]](#ref-1). Everything below comes from that file and from the grades on the site, as they stood on 15 November 2026.

## One network, two handshakes wide

The first thing I checked was whether these are separate little worlds. They aren't.

- 643: of 696 funded companies sit in one connected network
- 78: per cent of pairs in it are two steps apart or closer
- 12: per cent of all pairs share at least one investor
- 156: companies backed by Y Combinator, the most of anyone

The investor graph behind Anchor Terminal, 15 November 2026. Source [1].

One step means two companies share an investor. Two steps means each of them shares an investor with the same third company [[1]](#ref-1). Pick two companies in that network and about four times in five they share an investor, or each shares one with the same third company.

My first guess was that this is just Y Combinator, since it backs 156 of the 696 [[2]](#ref-2). So I took YC out and ran it again. 602 companies are still in one piece [[1]](#ref-1). The big funds hold it together without YC.

Funded companies backed, the ten most frequent investors

| Investor | Companies backed |
| --- | --- |
| Y Combinator | 156 |
| Andreessen Horowitz | 60 |
| Sequoia Capital | 56 |
| Accel | 45 |
| NVIDIA | 42 |
| Salesforce | 41 |
| Index Ventures | 40 |
| Insight Partners | 40 |
| General Catalyst | 39 |
| SV Angel | 36 |

Each series has a zero-based scale.

Two of the top six are companies, not funds.

## NVIDIA funds its own customers

[NVIDIA](/investors/nvidia/) backs 42 of the funded companies [[3]](#ref-3). Where it puts that money is the interesting bit.

Share of the funded companies in a category with NVIDIA or AMD among their investors, per cent

| Category | NVIDIA | AMD |
| --- | --- | --- |
| Inference providers, 16 funded | 50 | 6 |
| Open-weight model labs, 12 funded | 50 | 33 |

Each series has a zero-based scale.

Half the funded inference providers have NVIDIA money, Anthropic, Cohere, Groq, Deep Infra, OpenRouter and Reflection among them. So do half the funded open-weight labs, Mistral, Hugging Face, Thinking Machines, Poolside, Nous Research and Recursive [[3]](#ref-3). [AMD](/investors/amd/) is in four of those same open-weight labs [[4]](#ref-4). These are the companies that buy the most GPUs. The chip makers are on their cap tables.

I don't think that's sinister. It's how a supplier makes sure there's demand. But it means that when an agent picks an inference provider, odds are even that the chip supplier is also a shareholder.

NVIDIA and AMD sell rival chips, and they're on nine of the same cap tables, Cohere, Fireworks AI, Hugging Face, Luma AI, Runway, Thinking Machines and World Labs among them [[1]](#ref-1). And NVIDIA's most frequent partner in a round isn't a fund. NVIDIA and Salesforce back 13 of the same companies, the fourth most common pairing in the whole graph, ahead of most venture pairs [[1]](#ref-1).

## The same money on both sides

I expected investors to back one horse per race. Instead I found the same names on rivals, over and over.

| Rivals | Investors they share | Some of them |
| --- | --- | --- |
| Anthropic and SambaNova | 8 | BlackRock, GIC, Samsung, Temasek |
| Anthropic and Groq | 7 | NVIDIA, Altimeter, Cisco, Samsung |
| ElevenLabs and OpenAI | 6 | Andreessen Horowitz, Sequoia, NVIDIA |
| Anthropic and Reflection | 6 | NVIDIA, Sequoia, Lightspeed |
| Thinking Machines and Mistral | 4 | Andreessen Horowitz, NVIDIA, ServiceNow |
| Supabase and ClickHouse | 4 | Coatue, GIC, Lightspeed |

The table is pairs of companies whose main category is the same [[1]](#ref-1). Accelerators do it by design. [Y Combinator](/investors/y-combinator/) has 7 of the 16 funded voice-agent companies (Vapi, Retell AI, Bland AI, Daily, Bolna, Vogent and VOYGR), 6 of the 10 in document extraction and 6 of the 15 browser-agent companies, including Browser Use, Kernel and Skyvern [[2]](#ref-2).

Sometimes it's one person. Emily Zhao at Salesforce Ventures led the fund's investments in Anthropic, Cohere, Hugging Face, Together AI, fal and Runway [[5]](#ref-5). That's two model labs, a model hub, two inference platforms and a video model, from one desk.

Sometimes it's two people who invest as one. [Daniel Gross](/investors/daniel-gross/) has nine listed investments, and [Nat Friedman](/investors/nat-friedman/) is in eight of them, AssemblyAI, ElevenLabs, Hume AI, Suno, Pika, ggml.ai, Weights & Biases and Shippo [[16]](#ref-16). The one Gross made without him is Railway. Four of the eight are voice or music.

## Founders fund each other

174 founders of listed tools also invest in other listed tools [[1]](#ref-1). [Guillermo Rauch](/investors/guillermo-rauch/), who founded Vercel, is in 32 of them on his own [[6]](#ref-6). The people building the agent stack are also paying for it.

Rauch's list has more than one horse in several races. Supabase and Neon both sell hosted Postgres. Chroma and Meilisearch are both search databases. Runway, fal and FLORA all make or serve generative media. And he backs Railway, which deploys web apps the way Vercel does [[6]](#ref-6). It reads less like picking winners and more like owning a slice of whatever developers build with next.

Only one pair goes both ways. Karri Saarinen of Linear invested in Vercel [[7]](#ref-7), and Guillermo Rauch invested in Linear [[8]](#ref-8).

The ones I didn't expect are founders backing a company in their own category.

- [Eoghan McCabe](/investors/eoghan-mccabe/), co-founder of Intercom, is an investor in [Front](/tools/front) [[9]](#ref-9). Both are support platforms.
- [Mati Staniszewski](/investors/mati-staniszewski/), co-founder of ElevenLabs, is an investor in [LiveKit](/tools/livekit-agents) [[10]](#ref-10). Both sell to people building voice agents.
- [Matias Woloski](/investors/matias-woloski/), co-founder of Auth0, is an investor in [Keycard](/tools/keycard) [[11]](#ref-11), which handles identity for agents.
- [Shin Kim](/investors/shin-kim/), founder of Eraser, is an investor in [tldraw](/tools/tldraw) [[12]](#ref-12). Both draw diagrams.

## Public-market money arrived

The rounds in the graph used to be venture money. In the last two years the names changed.

| Investor | Funded companies backed |
| --- | --- |
| [GIC](/investors/gic/), Singapore's sovereign fund | 18 |
| [T. Rowe Price](/investors/t-rowe-price/) | 18 |
| [BlackRock](/investors/blackrock/) | 15 |
| [Fidelity](/investors/fidelity-investments/) | 12 |
| [Temasek](/investors/temasek/) | 8 |
| [Qatar Investment Authority](/investors/qatar-investment-authority/) | 3 |
| [MGX](/investors/mgx/), Abu Dhabi | 3 |

T. Rowe Price alone backs Anthropic, OpenAI, ElevenLabs, Cognition, Crusoe, Lambda, ClickHouse and Temporal [[13]](#ref-13). And 2026 already has more rounds in the graph than 2021, the last peak, did.

Round mentions by year, one per investor per round

| Year | Rounds |
| --- | --- |
| 2019 | 371 |
| 2020 | 546 |
| 2021 | 1127 |
| 2022 | 872 |
| 2023 | 683 |
| 2024 | 881 |
| 2025 | 1156 |
| 2026 | 1262 |

Each series has a zero-based scale.

The dip after 2021 and the climb back are the market everyone remembers. The 2026 figure covers ten and a half months.

## From OpenAI's fund to SpaceX

The strangest chain in the graph runs through Cursor. OpenAI's startup fund led Anysphere's $8 million seed round in October 2023 [[14]](#ref-14), and OpenAI is named again in the $60 million Series A in August 2024 [[17]](#ref-17). It's the only listed company where [OpenAI](/investors/openai/) appears as an investor [[1]](#ref-1). Then SpaceX agreed to buy Anysphere in June 2026 at an implied equity value of $60 billion, and the merger closed on 14 August 2026, leaving Cursor a wholly owned SpaceX subsidiary [[18]](#ref-18).

So one line in the graph starts at a fund with OpenAI's name on it and ends inside SpaceX. It's worth being precise about that line. The OpenAI Startup Fund raises its money from outside limited partners, Microsoft among them, and neither OpenAI nor Sam Altman has a financial interest in it [[19]](#ref-19). The link is to the fund, not to OpenAI the company.

## Does money buy a better grade?

This is the question only our data can answer, because the graph sits next to a grade for every listing. So I matched each funded company to its best-graded listing and grouped them by how many investors they have.

Median best score of the company's listings, by number of investors

| Investors | Median score |
| --- | --- |
| 1 | 57 |
| 2 to 4 | 60 |
| 5 to 9 | 61 |
| 10 to 19 | 62 |
| 20 or more | 65 |

Each series has a zero-based scale.

It goes up, a little. Eight points between a company with one investor and one with twenty [[15]](#ref-15). But the correlation between investor count and score is 0.16, which is close to nothing. The middle half of each group spans 16 to 20 points, about twice the gap between the groups [[15]](#ref-15).

The other comparison is even flatter. Listings from funded companies have a median score of 61.0. Listings from everyone else have 59.2 [[15]](#ref-15). "Everyone else" mixes two very different groups, the public giants whose products aren't startups at all and the open-source projects that never raised. The best-graded listings without a round in the graph are [Vertex AI](/tools/vertex-ai), the [MongoDB MCP server](/tools/mongodb-mcp), [Backblaze B2](/tools/backblaze-b2) and [Strands Agents](/tools/strands-agents) [[15]](#ref-15).

Mem0 is the case that makes it concrete. Its backers include the chief executives of Datadog, Supabase, PostHog and Weights & Biases, GitHub's former chief executive and HubSpot co-founder Dharmesh Shah [[20]](#ref-20). That's about as experienced an infrastructure cap table as you'll find. Its [listing](/tools/mem0) grades E, 42.6. [Infisical](/tools/infisical), a secrets manager with Datadog's, Supabase's and PostHog's founders and Y Combinator among its backers too, grades AA, 86.7 [[15]](#ref-15). Same people, opposite ends of the scale.

What the grade measures is whether an agent can use the thing. Published prices, a status page, an API that tells you what went wrong. None of that costs a Series C. A team of four can ship it, and plenty of very well funded teams haven't.

## Go and look

The page is at [/graph/](/graph/), and the whole thing is [one JSON file](/graph/graph.json) you can load into anything. Every investor has a page with each company and the source that names them, and every founder has one too.

Two things I'd like to do next. Join this to the use cases we've been mapping for every category, to see whether the jobs nobody does well are the ones nobody funded. And look at who backs the tools that agents actually call most, which is a different list from the ones that raised most.

If a link is wrong, or you'd like your page taken down, write to privacy@anchorterminal.com and we'll fix it.

---

More in [From the data](https://www.anchorterminal.com/blog/from-the-data/index.md) · Previous from the data [Who gets paid in the agent stack. Seats, meters and the year the labs went shopping](https://www.anchorterminal.com/blog/who-gets-paid-in-the-agent-stack.md)

- Older: [A machine that proves Fermat in 29 hours, and the behemoth that's back](https://www.anchorterminal.com/blog/a-machine-that-proves-fermat.md)
- Newer: [The Log, issue 6: the ground stop, the outage, the proof](https://www.anchorterminal.com/blog/the-log/issue-6-the-ground-stop-the-outage-the-proof.md)
- All posts: https://www.anchorterminal.com/blog/index.md

Anchor Terminal grades the tools AI agents run on, with a source for every score. [The top tools](https://www.anchorterminal.com/top/index.md) · [Categories](https://www.anchorterminal.com/categories/index.md) · [Compare two tools](https://www.anchorterminal.com/compare/index.md) · [Prices](https://www.anchorterminal.com/prices/index.md) · [Sunsets](https://www.anchorterminal.com/sunsets/index.md) · [API, MCP server and CLI](https://www.anchorterminal.com/agents/index.md)

## References

Sources read on 15 November 2026 unless stated.

[1] Anchor Terminal, Who backs whom, the investor graph as JSON, 15 November 2026. Source for every count, pair and connection in this post, worked out from the file. https://www.anchorterminal.com/graph/graph.json

[2] Anchor Terminal, Y Combinator investor page. Source for the companies Y Combinator backs, by category. https://www.anchorterminal.com/investors/y-combinator/

[3] Anchor Terminal, NVIDIA investor page. Source for the companies NVIDIA backs. https://www.anchorterminal.com/investors/nvidia/

[4] Anchor Terminal, AMD investor page. Source for the open-weight labs AMD backs. https://www.anchorterminal.com/investors/amd/

[5] Salesforce Ventures, Emily Zhao. Source for the investments she led. https://salesforceventures.com/people/emily-zhao/

[6] Anchor Terminal, Guillermo Rauch investor page. Source for the listed companies he backs. https://www.anchorterminal.com/investors/guillermo-rauch/

[7] Vercel, About. Source for Karri Saarinen as an investor in Vercel. https://vercel.com/about

[8] Linear, Series B. Source for Guillermo Rauch as an investor in Linear. https://linear.app/now/series-b

[9] Front, Announcing our Series A. Source for Eoghan McCabe as an investor in Front. https://front.com/blog/announcing-our-series-a

[10] LiveKit, Series A. Source for Mati Staniszewski as an investor in LiveKit. https://livekit.com/blog/livekit-series-a

[11] Keycard, About. Source for Matias Woloski as an investor in Keycard. https://www.keycard.ai/about/

[12] tldraw, Announcing tldraw's seed round. Source for Shin Kim as an investor in tldraw. https://tldraw.dev/blog/announcing-tldraws-seed-round

[13] Anchor Terminal, T. Rowe Price investor page. Source for the companies T. Rowe Price backs. https://www.anchorterminal.com/investors/t-rowe-price/

[14] TechCrunch, Anysphere raises $8M from OpenAI to build an AI-powered IDE, 11 October 2023. Source for the OpenAI Startup Fund leading the seed round. https://techcrunch.com/2023/10/11/anysphere-raises-8m-from-openai-to-build-an-ai-powered-ide/

[15] Anchor Terminal, rankings, 15 November 2026, and the benchmark methodology. Source for every score and grade in this post. https://www.anchorterminal.com/api/v1/rankings.json

[16] Anchor Terminal, Daniel Gross investor page. Source for his listed investments and the ones shared with Nat Friedman. https://www.anchorterminal.com/investors/daniel-gross/

[17] Cursor, Series A, 22 August 2024. Source for the investors named in the Series A, OpenAI among them. https://cursor.com/blog/series-a

[18] Investing.com, SpaceX completes acquisition of Anysphere in $60 billion all-stock deal, 14 August 2026, from the filing with the SEC. Source for the closing and Anysphere's status as a wholly owned subsidiary. https://www.investing.com/news/sec-filings/spacex-completes-acquisition-of-anysphere-in-60-billion-allstock-deal-93CH-4860682

[19] TechCrunch, OpenAI Startup Fund quietly raises $15M, 26 April 2024. Source for the fund's outside limited partners and that neither OpenAI nor Altman has a financial interest in it. https://techcrunch.com/2024/04/26/openai-startup-fund-quietly-raises-15m

[20] Mem0, Series A. Source for the named backers. https://mem0.ai/series-a

---

Query this catalogue directly over MCP: https://www.anchorterminal.com/mcp (Streamable HTTP, no key). How: https://www.anchorterminal.com/agents/#mcp
