Bessemer Venture Partners has 519 portfolio companies tracked on the network, and all 519 publish APIs indexed here. Founded in 1911 out of the Phipps steel fortune, more than $20B under management, investing globally across seed, early and growth, with Shopify, Twilio, LinkedIn, DocuSign, Toast, PagerDuty, Auth0, Box and Canva in the book.
The distribution
| Band | Count | Share |
|---|---|---|
| Exemplar | 13 | 2.5% |
| Strong | 34 | 6.6% |
| Developing | 67 | 12.9% |
| Thin | 54 | 10.4% |
| Emerging | 144 | 27.7% |
| Minimal | 206 | 39.7% |
| Unrated | 1 | 0.2% |
Forty-seven companies out of 519 — 9.1% — score strong or better.
The top of the book
| Company | Band | Score |
|---|---|---|
| Zapier | exemplar | 79.3 |
| SendGrid | exemplar | 77.0 |
| Fal.ai | exemplar | 75.1 |
| Anthropic | exemplar | 75.0 |
| Storwize | exemplar | 74.6 |
Zapier at 79.3 is a fitting leader for this portfolio. A company whose entire product is connecting other companies’ APIs has an unusually direct incentive to publish a good one.
Against Lightspeed
Lightspeed, covered on Friday, makes the useful comparison. Two large multi-stage funds, similar vintages of activity, tracked the same way:
| Bessemer | Lightspeed | |
|---|---|---|
| Companies tracked | 519 | 603 |
| Exemplar | 13 | 13 |
| Strong | 34 | 27 |
| Strong-or-better share | 9.1% | 6.6% |
| Minimal | 206 (39.7%) | 278 (46.1%) |
Identical exemplar counts. Thirteen each, from books of different sizes.
That is the more interesting number than either fund’s total. The very top of a venture portfolio — the companies that have built genuinely excellent, agent-ready API surfaces — seems to cap out around a dozen regardless of whether you backed 519 companies or 603. It is a small population, and no amount of portfolio construction manufactures more of them.
Where the funds actually differ is the middle. Bessemer carries 34 strong against Lightspeed’s 27, and a shorter minimal tail — 39.7% against 46.1%. On this measure Bessemer’s book is modestly healthier below the top.
What to do with this
For an LP, the distribution is a diligence input that no pitch deck contains: it is an independent read on whether portfolio companies have shipped something real and describable.
For the fund, the actionable band is developing — 67 companies at Bessemer. These have working APIs and inadequate descriptions of them. It is the cheapest score movement available anywhere in a portfolio, and it is a platform-team play rather than a per-company engineering project.
For a founder, the useful framing is that 2.5% of a top-tier portfolio reaches exemplar. Getting there is a differentiator against your own investor’s book.
Takeaway
519 companies, 13 exemplars, 9.1% strong or better, and a top band identical in size to a fund with 84 more companies in it. Excellence at the top of a venture portfolio is a fixed, small number. The difference between funds shows up in the middle.
See the full portfolio at apis.io/vcs/bessemer-venture-partners/.