61% of VC theses do not match the deals they actually fund

I scraped every active VC fund website I could find. 8,665 of them. I pulled their thesis pages, their about sections, their blog posts where partners explain what they back. Then I pulled their actual portfolio from the last 18 months.

61% of funds had a meaningful mismatch between what they say they invest in and what they actually write checks for.

Not a small drift. Not "we say fintech but we also do a little payments." I mean a climate fund with 7 logistics deals and 2 climate deals. A "pre-seed generalist" fund that has written 14 of its last 16 checks at Series A. A "B2B SaaS specialist" whose recent portfolio is 40% consumer hardware.

This is not about catching VCs in a lie. Most investors believe their own thesis when they write it. The problem is simpler: thesis drift happens faster than website updates, and founders waste months pitching the wrong story to the wrong fund at the wrong stage.

The climate fund that stopped doing climate

One fund's site said "we back founders decarbonizing the built environment." Clean thesis. Specific vertical. You would pitch them if you were doing industrial heat pumps or low-carbon concrete.

Their last 12 deals:

4 logistics and supply chain software companies
3 manufacturing ops tools
2 EV charging networks (this counts as climate)
2 financial infrastructure plays
1 actual decarbonization technology

What happened? I talked to a founder who pitched them in month 6 of her fundraise. She had a legitimately good industrial emissions play. Perfect fit for the thesis on the site. They passed in the first call. The partner told her, off the record, "we have been seeing better opportunities in the picks-and-shovels layer. Software that helps climate companies scale, not the climate technology itself."

That is a reasonable pivot. Plenty of funds made that move between 2022 and 2024 when hard-tech timelines stretched and early software exits started closing. But the website still said "decarbonizing the built environment." The thesis page had not been touched in 19 months.

She spent 11 days researching that fund, tailoring her deck, writing a custom cold email. All of it based on a thesis that no longer represented what they actually funded.

The four most common mismatches

I tagged every mismatch by type. Four patterns made up 83% of cases:

1. Stage drift (31% of mismatches)

Fund says "pre-seed and seed." Portfolio shows Series A and B. This happens when a fund raises a bigger vehicle or when later-stage deals start converting better. The economics change, the thesis page does not.

2. Sector drift (28%)

Fund says "fintech." Portfolio shows logistics, proptech, and infrastructure software with a payments angle. The real thesis became "software with transaction revenue" but the site still says fintech.

3. Geography drift (14%)

Fund says "West Coast and Mountain West." Portfolio shows 9 of the last 11 deals in Texas and Florida. Migration follows the founders, the website does not follow the migration.

4. Model drift (10%)

Fund says "we lead rounds and take board seats." Portfolio shows 8 of 10 recent deals were follow-on participation, no board seat. This happens when partnership bandwidth tightens or when the fund decides small checks into hot rounds beat leading cold ones.

The remaining 17% were a mix: stage + sector double-drift, thesis abandonment (pivoted entirely to a new strategy mid-fund), and a few cases where the website was just stale and wrong.

Why this happens (and why it is not malicious)

VCs do not wake up and decide to mislead founders. Thesis drift happens because:

Deal flow is uneven. A fund says "we do B2B SaaS" because that is the plan. Then three incredible consumer deals walk in through the network, and they are too good to pass. After 18 months, the portfolio is 40% consumer. The thesis was real when written. The market just sent better consumer deals.

Funds follow their winners. A generalist seed fund backs 20 companies. Three of them are in logistics. Those three do well, the partners start taking more logistics calls, the network starts sending more logistics deals. Two years later they are a logistics-heavy fund. The thesis page still says "generalist."

Websites are not managed by the people writing checks. A junior associate updates the site when the fund launches. The partners are in back-to-back meetings. The thesis page gets touched once a year, if that. The portfolio page sometimes does not reflect the last fund at all.

Pivots happen quietly. A fund decides to move upstream or change focus. They tell their LP base, they tell their close network. They do not always announce it publicly, and the old messaging stays live because no one remembered to delete it.

None of this is fraud. It is just operational drift. But if you are a founder building a list of 60 funds to pitch, and 37 of them are drifted, you just wasted weeks of research time.

What founders should do instead

Stop starting with the thesis page. Start with the portfolio.

Here is the process I have watched work:

Pull the last 12-18 months of deals, not the full portfolio. A fund's 2019 deals do not tell you what they are backing in 2025. The recent deals do. If you are using our investor directory, filter by "deals closed in last 18 months." If you are using Crunchbase or another tool, do the same.
Tag the deals by stage, sector, and geography. Not what the fund says they do. What they actually did. If 9 of the last 11 deals were Series A, they are a Series A fund right now, regardless of what the site says.
Look for repeat patterns, not one-offs. One consumer deal in a B2B portfolio is an outlier. Four consumer deals is a signal. Six is a new thesis, whether they have written it down or not.
Check the date of the thesis page. If you are on their site, view the page source and search for "date" or "modified." If the thesis page has not been touched in 18+ months and the portfolio shows drift, trust the portfolio.
If there is a mismatch, adjust your pitch or skip the fund. Do not try to pitch the written thesis if the portfolio shows they moved. Either pitch what they actually fund now, or move on. Pitching old thesis to a drifted fund is like applying for a job listing that is 2 years old. Technically still posted, but no longer real.

One founder I worked with had a list of 80 funds. We ran this filter. 52 of them had meaningful drift. She re-pitched 19 of them with messaging that matched the actual portfolio, not the website. She skipped 33 entirely. Her response rate went from 11% to 34% and she closed her round in 8 weeks.

The funds that do not drift (and why)

Not every fund drifts. About 39% had tight thesis-to-portfolio matching. The common traits:

Specialist funds with narrow mandates. If you only do AI infra or only do femtech, drift is hard. The lane is too narrow to wander.
Funds that update their site every quarter. A few firms treat their website like a living document. Thesis pages get revisited every 90 days. Portfolio pages update in real time.
Solo GPs and small teams. When one person is both writing checks and managing the web presence, the message stays tight. Drift happens more in larger platforms where communication overhead is higher.
Funds that publish deal memos or investment theses regularly. If a partner is writing public posts about what they are seeing and funding, the thesis stays current because it is constantly being articulated.

If you find a fund with tight alignment, that is a good sign. It usually means they have clarity about what they do and they communicate it well. Those funds are easier to pitch because you know what they want.

The one-line fix for your outreach

If you are mid-raise right now and you have been pitching thesis pages, stop. Spend 90 minutes this week pulling recent portfolio data for every fund on your list. Compare it to the thesis. If there is drift, either adjust your angle or cut the fund.

The highest-converting cold emails I have seen do not mention the thesis at all. They say: "I saw you backed [Company A] and [Company B] in the last year. We are building something similar but for [adjacent market]. Here is the deck."

You are not trying to fit their written thesis. You are showing them you understand what they actually fund. That is the signal that gets you the call.

If you want to skip the manual research, score your deck and get a matched list in 30 seconds. The tool pulls recent portfolio data, not thesis pages, and surfaces funds that actually wrote checks in your category in the last 18 months. It is free for the first score, and it will save you the 11 days that founder spent pitching a climate thesis that no longer existed.

<<>>