👋 Hi, I’m Andre and welcome to my newsletter Data Driven VC which is all about becoming a better investor with data and AI.
ICYMI, check out some of our most read episodes:
Brought to you by Vessel - Agentic fund operations for VC and PE firms
LPs aren't just evaluating your returns. They're evaluating whether you're in control of your own numbers.
If your data lives in disparate, fragmented systems with your fund admin, scattered across PDFs, board decks, spreadsheets, and legacy software, you are at a disadvantage.
Vessel's AI agents give you control of your fund operations and portfolio data in real time, so your team can spend its time on value-add activities.
Welcome to another Data Driven VC “Insights” episode where we cover the most interesting research and reports about startups, VCs, LPs, AI & automation.
The Power Law Doesn't Stop at Early Stage
Thomas Laffont from Coatue presented this data at the All-In Podcast Liquidity Summit: 8% of unicorns ($1-10B) go on to reach decacorn status, 13% of decacorns ($10-100B) reach $100B, and 31% of $100B+ companies have gone on to 10x. The higher you are in the valuation stack, the more likely you are to keep compounding.
31% vs. 8%: A company already valued above $100B is nearly four times more likely to 10x than a unicorn is to reach decacorn status. Examples in the centacorn tier include Tesla, Nvidia, Broadcom, Meta, Amazon, and Apple.
Scale Compounds: At large scale, network effects, pricing power, and platform expansion offset the smaller theoretical multiple ceiling that makes late-stage investing look less attractive on paper.
Survivorship Bias: The dataset only includes companies that reached each tier. The 8% unicorn figure excludes the majority of $1B+ companies that never graduated to the next bracket.
✈️ KEY TAKEAWAYS
The data challenges the assumption that 10x upside lives exclusively at early stages. For investors with late-stage access, the power law may extend further than the conventional model suggests. Survivorship bias warrants a discount, but the directional finding is large enough to affect how growth and crossover funds should be priced.
The Enterprise Software Playbook Is Dead
Mike Vernal (Conviction) argues in a recent post that the wedge-to-suite-to-platform progression no longer holds. Cursor, Clay, Harvey, Cognition, Sierra, Baseten, Fireworks, and Lovable all went from roughly $0 to $100M ARR in the past two years, compressing what used to be a 6-10 year journey.
Wedge as Liability: The traditional Act I harbor (a niche to reach $10-50M ARR before building adjacencies) now functions as a constraint. As software engineering costs drop, building Act I and Act II simultaneously is no longer unrealistic.
Cursor as the Reference Case: Vernal initially thought Cursor's plan to replace VS Code at seed stage was too aggressive. He was wrong, and now says replacing VS Code feels under-ambitious. Ideas that look too aggressive on day one may already be too conservative by fund close.
Ambition Over Wedge: Vernal writes his investment filter has shifted from "what is the protective wedge" to "unreasonable, unrelenting ambition," because the compressed timeline rewards founders who plan to build the full stack from the start.
✈️ KEY TAKEAWAYS
Investors still screening primarily for a defensible wedge may be systematically underweighting companies going for the full platform from day one. The open question is whether the cohort that compressed this timeline reflects a durable structural shift or a specific AI wave that will eventually plateau.
VC Fund Performance Benchmarks
Peter Walker at Carta published Net TVPI benchmarks across 2,689 US venture funds ($10M to $1B+) as of Q1 2026. The 2021 vintage five-year distribution: 50th percentile 1.04x, 75th 1.29x, 90th 1.54x, 95th 1.93x.
What the distribution means: The median 2021 fund has barely returned invested capital at five years. Only the top 10% of funds have reached 1.54x or better. The entire cohort's distribution is compressed toward 1x in a way no prior vintage shows at the same age.
Against the 2017 benchmark: The 2017 vintage at nine years shows a median of 1.68x and a 90th percentile of 3.47x. The 2021 cohort's 95th percentile (1.93x) is below the 2017 cohort's 75th percentile (2.21x).
Walker's four headwinds: Record investor entry, peak entry valuations, the 2022 rate reversal, and the AI platform shift arriving 14 months after peak 2021 deployment. The 2022 vintage (4 years) is already tracking better: 90th percentile 1.70x, 95th 2.02x.
✈️ KEY TAKEAWAYS
The 2021 vintage's entire return distribution sits below every comparable prior vintage at the same age. For LPs evaluating current AI-era entry prices, this is the clearest available data point on what happens when elevated valuations, a rate reversal, and a platform shift compound on the same cohort.
Join 1630+ investors in our free Slack group as we automate our VC job end-to-end with AI. Live experiment. Full transparency.
The Agent Loop That Actually Works
Shann Holmberg published a breakdown of agent looping distinguishing single-agent loops, fleet loops, and the open vs. closed loop split that determines whether an autonomous workflow is viable in production.
Single vs. Fleet: A single-agent loop runs one agent through discovery, planning, execution, verification, and iteration until output meets a standard. A fleet loop adds an orchestrator that breaks the goal into components, delegates to specialist agents, who delegate further to subagents. Every node runs the same five-stage cycle.
Open Loop = Token Burn: Open loops give the agent wide exploratory space. Holmberg notes they burn an "insane amount of tokens" and without tight standards become a "slop machine." Currently viable only for teams with uncapped compute budgets.
Closed Loop = Production Viable: A human designs the path first: clear goal, defined steps, an eval gate at each step, a handoff point. The agent loops within that structure, and each run feeds performance data into the next. Runs on a normal budget and improves over time.
✈️ KEY TAKEAWAYS
Most "autonomous agent" products being sold today are open loops that few customers can afford to run at scale. The near-term infrastructure winners are the tools that make closed loop construction faster. For investors evaluating agent companies, the key question is not whether the agent can loop but whether the loop's cost structure is viable for the target buyer.
The Portfolio Math Most VCs Get Wrong
Clayton Petty at Gradient Ventures ran Monte Carlo simulations across four deployment strategies for $100M and $30M funds, varying ownership targets, portfolio size, and follow-on reserves. Core finding: extreme concentration destroys median returns without proportionally improving top-decile upside.
$100M Fund Sweet Spot: A follow strategy targeting ~5% ownership returns 2.8x at the 75th percentile. A lead strategy at ~10% ownership produces a 90th percentile of 6.1x. Maximum concentration (22 companies, 15% ownership) has a median TVPI of 0.9x and a 90th percentile of 5.5x, below the less-concentrated lead strategy at the same percentile.
$30M Fund: Concentration Backfires: At $30M, high-concentration strategies shrink the portfolio to 9-13 companies. Top quartile falls to 1.4-1.5x and top decile to 3.5x, below what more diversified strategies produce. Half of simulated maximum-concentration funds lose LP money.
What Simulations Cannot Fix: Petty notes the model distributes strategy outcomes, not individual fund results. No construction strategy compensates for poor deal selection or insufficient ownership in the eventual outlier.
✈️ KEY TAKEAWAYS
Below a certain portfolio size, the probability of hitting a meaningful winner drops faster than the theoretical upside increases. For LPs evaluating managers, the key diagnostic is whether the portfolio construction math is internally consistent across fund size, follow-on reserves, and realistic deal velocity.
Why Venture Capital Should Scale Out, Not Up
Dan Gray from The Odin Times argues in Survival of the Fittest that scaling venture through platform expansion erodes the selection mechanism that makes venture work. Drawing on Lerner and Gompers (HBS) and Ewens and Rhodes-Kropf (SSRN), he makes the case for horizontal scaling (more small firms) over vertical scaling (fewer, larger platforms).
20x Capital Growth, No Proportional Outcome Growth: The VC industry has grown approximately 20x over two decades without a proportional increase in high-quality exits. Capital surges tend to intensify competition for the same technologies rather than fund genuinely new categories.
Value Is Partner-Level, Not Firm-Level: Ewens and Rhodes-Kropf found that value in VC firms resides in individual partners, not firm brand or process. Larger firms dilute individual partner influence and appear to scale primarily to reduce fundraising friction, not to improve returns.
First-Time Funds and Exit Value: Excluding 2021, years with a larger share of capital going to first-time funds show a positive correlation with subsequent exit value, consistent with the thesis that emerging managers source outliers that incumbent firms miss.
✈️ KEY TAKEAWAYS
If VC alpha is partner-level, concentrated LP allocations to large platform firms optimise for career safety rather than returns. The early-stage market appears chronically undercapitalised relative to its contribution to eventual exit value. AI-driven reductions in LP reporting friction may be the structural fix that makes a broader portfolio of smaller fund commitments economical.
That’s it for today!
Stay driven,
Andre
PS: Check out Vessel to automate your fund operations
