👋 Hi, I’m Andre and welcome to my newsletter Data Driven VC which is all about becoming a better investor with data and AI.
👉 Join the DDVC x Slush Breakfast in Helsinki Nov 18 with Goodwin & Harmonic
Brought to you by Affinity – The AI-first private capital CRM
What the most data-driven funds got right (and wrong) about AI
AI has made investment workflows cheap to automate. The hard part is organizational: who owns scoring, who sees which deals, and which IC memo sections need a human.
Join Andre Retterath of Data Driven VC and Lindsay Bayuk, Affinity’s CMO, on Nov 17 at 1 PM ET to learn why messy-data scoring models lose trust, why 80%+ of IC papers are recycled, and how permission-aware MCPs protect sensitive deals.
Welcome to another Data Driven VC “Insights” episode where we cover the most interesting research and reports about startups, GPs, LPs, AI & automation.
Is Private Equity Cooked?
Jen Kha at a16z asks whether private equity is cooked, arguing that tech buyouts face structural pressure even as LPs keep re-upping with the same managers.
Same Sector, Very Different Marks: In H1 2026, venture funds marked their tech holdings up 16% while buyout funds marked theirs down just 4%. Kha argues PE owners set their own marks behind closed doors, so the real stress may not show yet.
2021-22 Deals Drive the Distress: Many buyouts struck near the 2021-22 peak can no longer pay interest in cash, so it gets added to the loan. Those two vintages account for 70% of the debt lenders took over in 2025 and 2026.
Deals Now Need 10-12% EBITDA Growth: Citing Bain, Kha says a typical 2015 buyout needed about 5% annual EBITDA growth to reach 2.5x in five years. With pricier debt and flat entry multiples, the same deal now needs 10 to 12%.
✈️ KEY TAKEAWAYS
Kha advises LPs to ask managers how much interest on 2021-22 deals is being added to the loan, then plan an exit route. Her case for venture rests on access to the top 5% of managers, and it comes from a firm that raises venture capital.
The Corpus Is the Economy
Yoni Rechtman at Slow Ventures draws on 60+ calls across the data supply chain to argue that AI training data is becoming a priced asset, bought by labs, AI-native firms and large enterprises.
No Units, No Prices: Training data has no standard unit or price, and deals carry wide spreads and high transaction costs. Company data sits unvalued on balance sheets until it's sold, usually as an entire dataset from a failing company.
Raw Data Is the Smallest Slice: Most value is added after collection, as data is turned into tasks, training environments and deployed models. If tasks sell for 2 to 5x the raw data, Rechtman argues sellers should build them, and the largest vendors are already moving into training and deploying models.
Every Business Becomes a Data Source: He expects synthetic data to flood the cheap end of the market, leaving the value in novel real-world data. Companies deploying AI generate it constantly, as each deployment hits edge cases models have never seen.
✈️ KEY TAKEAWAYS
For investors, the opportunity sits in businesses that originate novel data through their own operations. Rechtman also expects data to reach balance sheets as a financeable asset, opening room for data-backed rollups and making what to keep or share a core CEO decision.
Upgrade your subscription to access our premium content & join the Data Driven VC community
How to Engineer Your Fundraise
Jamie Gull, founder and GP at Wave Function Ventures and a former SpaceX engineer, lays out how to build an investor pipeline before the first pitch.
Every Investor Must Earn a Spot: Gull's CRM targets 200+ names, each with a named partner and a reason to care, ideally shown by an adjacent, non-competing portfolio company. Check size filters fastest, since a fund writing $10M Series A checks won't lead a $3M pre-seed.
Practice on Investors You Can Afford to Lose: A pitch needs 10 to 50 meetings to lose its kinks, so he starts with real but lower-priority investors and friendlies before his hot list. Practicing for two months, though, risks letting the round go stale.
Line Up Intros Before Launch: For each target, Gull asks roughly three people who genuinely know the investor whether they'd forward a note at some point, so every intro can go out at once on launch day.
✈️ KEY TAKEAWAYS
Founders planning a January raise should be building this list now, with a warm intro path mapped for every priority investor before launch. For investors, the framework shows the diligence founders run on them, from portfolio fit and reserves to which partner will champion the deal.
Venture's Black Box Problem
Dan Gray at Odin argues in The Black Box that large, liquidity-insensitive LPs have turned venture into a confidence game where shared markups count for more than exits.
Marks That Never Get Tested: When LPs stop demanding liquidity on a 10-year horizon, Gray argues, funds can hold assets at reported values that are never reconciled. He cites a 2020 study by Gornall and Strebulaev that put reported unicorn valuations 48% above fair value on average.
LPs Get Back Less Than They Put In: Odin's analysis of Cambridge Associates data shows US venture returning just $0.59 to $0.91 per $1 called each year since 2022, against $1.26 to $1.47 in 2016-20. Capital-call growth slowed from at least 30% a year to 7% at most.
Small Firms Face a Hard Choice: Emerging managers can ride alongside large firms for markups and accept a fragile portfolio, or avoid their categories and accept slower markups, possibly in exchange for healthier and earlier exits.
✈️ KEY TAKEAWAYS
For LPs, the piece is a reason to weigh distributions and outside-led rounds above reported value when judging managers. Gray expects capital to drift back toward small specialist firms eventually, though he admits it's unclear what would trigger that shift.
Find new LPs in our database of 30k+ active family offices, fund of funds, pensions & more via Final Close
How Menlo Got Hot Again
Laurie Owen at Refinery Media unpacks how Menlo Ventures rebuilt its brand around its Anthropic bet, and what other firms can take from it.
A Winner Alone Isn't a Brand: Menlo's roughly $1B Anthropic stake could be worth about $21.8B at a $1.5T valuation, yet Owen points out that earlier hits like Roku and Siri never built Menlo's brand. This time, the firm deliberately turned the win into an identity.
4 Parts to the Story: Menlo paired the bet with (1) a recurring enterprise AI survey the press kept citing, (2) a $100M fund for Anthropic-ecosystem startups that backed 45+ companies within about a year, (3) technical partners with their own followings and (4) builder-focused content.
The Combination Is the Point: None of the parts was new, but together they made Menlo the default enterprise AI firm. Spark and Sound are set to earn about $2B and $1.5B in carry from the same winner, yet Owen's anecdotal read is that their brands gained far less.
✈️ KEY TAKEAWAYS
For GPs with a breakout winner or a strategy shift, Owen's advice is to tie the firm to the category the company is winning, since a single company fades as a brand asset within 10 to 15 years. The association is far easier to build while the win is still news.
12 Questions LPs Should Ask VCs
Alex Menn at Begin Capital lists the questions every LP should ask a VC fund, after around 300 meetings, calls and events for Fund II in which almost no one asked them.
Open the CRM: LPs should ask how many deals a fund reviews each year, how sourcing splits across channels and how each channel converts into investments.
Ask About Lost Deals: Menn treats losses to tier-1 funds as a healthy signal for emerging managers, since a fund claiming to win every deal it wants is likely seeing weak deals or overpaying.
Test the Marks: LPs should ask exactly how TVPI was calculated, whether each markup is backed by a round with an external lead, and whether those valuations still hold today.
✈️ KEY TAKEAWAYS
GPs raising a Fund II should prepare answers to all 12 before the first meeting, including an introduction to a founder whose company was written off. For LPs, the list sets a diligence baseline that goes well beyond headline TVPI.
That’s it for today!
Stay driven,
Andre
