👋 Hi, I’m Andre and welcome to my newsletter Data Driven VC which is all about becoming a better investor with data and AI.
Upcoming events:
DDVC Investor Summit at Bits & Pretzels 2026 Sep 28-30
DDVC x Bits & Pretzels Breakfast 30th Sept in Munich
Brought to you by Granola – AI Copilot for Back2Back Meetings
You know those meetings where a random bot joins and suddenly everyone’s distracted? Granola works differently.
There are no meeting bots. Nothing weird joins your call.
Granola transcribes directly from your computer or phone audio. It works across any meeting tool: Zoom, Google Meet, Microsoft Teams. And even for in-person conversations just click to start.
Our team has been using it for years and couldn’t live without it anymore. Now you can test it 1 month for free with code "DATA100OFF"
Welcome to another Data Driven VC “Insights” episode where we cover the most interesting research and reports about startups, VCs, LPs, AI & automation.
Picking Your AI GTM Strategy
a16z lays out two enterprise AI sales playbooks: the "Lighthouse" strategy, where a few marquee logos unlock a market, and the "Landgrab" strategy, where speed and math win at volume.
The Lighthouse Strategy: Fits high-exposure, status-driven markets where a few credible customers going first unlock the whole market; sales cycles often run three to six months or longer.
The Landgrab Strategy: Fits low-exposure markets where proof doesn't travel and buyers just need the math to work; unit economics have to hold at volume, since coverage wins the market.
The Decision Map: Two questions place a market on the grid: buyer exposure and whether proof travels. High exposure plus proof that travels means lighthouse; low exposure plus proof that doesn't travel means landgrab. When they conflict, exposure wins.
✈️ KEY TAKEAWAYS
Buyer exposure and whether proof travels are the primary map for picking a GTM motion, ahead of product quality or team pedigree. Budget and ROI math are tiebreakers once exposure is already answered, and sales cycle length is a separate gut check for which territory you're in.
Chamath's Six-Layer AI Investing Stack
Chamath Palihapitiya laid out his current AI investing framework on X, splitting the market into six layers and explaining where he is putting capital and where he is staying out.
Bullish on Land, Power, and Shell, Bearish on Chips: He's acquired almost 6GW of grid power and behind-the-meter capacity through 2029, his fastest path to returns. He no longer invests in silicon since chip performance and supply-chain demands are now out of reach for new startups.
Skeptical on Clouds and Models: As AI safety concerns grow, he expects clouds will need to prove who's using their compute, and also questions how much model-layer revenue today reflects real usage versus inflated token consumption.
Building at the Harness and Application Layer: He started 8090 Solutions around the "harness," an enterprise's own data and workflows that sit on top of any AI model and keep switching costs low. He expects applications to win the same way, as companies embed their own harness into custom software instead of buying off-the-shelf tools.
✈️ KEY TAKEAWAYS
Chamath's picks track the chart's own margin logic: bullish on land, power, and shell, bearish on silicon and clouds, and building at harness and applications for tomorrow's margin. For VCs, "harness" is a useful lens for judging AI startup defensibility.
Public Markets Reward Quality Over Scale
Dan Gray (Odin) charted alpha versus the Nasdaq-100 for every US IPO since 2020 that raised over $3 billion in venture or private equity funding, and only one company comes out ahead.
Robinhood Is the Only Company With Positive Alpha: Robinhood is the sole company in the cohort beating the Nasdaq-100 since its IPO, with +61% lifetime alpha.
Uber and Airbnb Trail the Index Despite Positive Returns: They're the only other two companies in the cohort with positive lifetime stock returns, but both still underperform the Nasdaq-100, at -203% and -123% lifetime alpha respectively.
SpaceX Investors Down Nearly $18B in Six Weeks: SpaceX has raised comparatively less relative to its scale than peers, but a separate report cited in the thread put its IPO investors down close to $18 billion on paper six weeks after listing; SpaceX's own lifetime alpha stands at -28% so far.
✈️ KEY TAKEAWAYS
Dan Gray argues the more a company raises pre-IPO, the wider its post-IPO gap with public markets tends to run. For VCs backing late-stage mega-rounds, that's a repricing risk no amount of pre-IPO buzz offsets.
Join 1,897+ investors in our free Slack group as we automate our VC job end-to-end with AI. Live experiment. Full transparency.
One Engineer for Every Five Investors
Our Data Driven VC Landscape Report 2026 benchmarked team composition at 345 venture firms by assets under management (AUM), showing how staffing scales as funds grow.
7x Scaling from Smallest to Largest Funds: Full-time investor headcount grows from 5.9 at firms managing under $500M to 40.0 at firms managing $5B+, while engineering headcount scales from 1.2 to 7.9 over the same range.
A Ratio That Holds at Every Size: Firms maintain roughly one full-time engineer for every five full-time investors across all AUM tiers, a ratio that holds steady regardless of fund size.
Smaller Firms Cut Teams 25% Since 2025: Firms managing under $500M now run investment teams 25% smaller than in 2025, while firms above $5B grew their teams 20% larger.
✈️ KEY TAKEAWAYS
Engineering headcount scales with investor headcount, holding at roughly one engineer per five investors regardless of fund size. A firm that strays far from that ratio, or cut its team the way smaller funds did this year, is worth a direct question in diligence.
Startups Get Half Their Employee Equity Back
Peter Walker shared Carta data on how much vested startup equity employees never exercise, and the rate has barely moved despite a wave of AI secondary sales.
70%+ of Vested Options Went Unexercised in 2025: Over 70% of vested startup equity was not exercised when an employee left by layoff or by choice, ticking down only slightly in 2026.
AI Company Rate Bottomed at 42%, Now Back to 70%: Non-exercise rates at AI companies fell to a low of 42% during the 2021 to 2022 funding peak, then climbed back to 70% by Q2 2025, close to the 72% rate at non-AI companies.
Four Reasons Employees Skip Exercising: Carta cites not knowing they have to, lacking cash for the exercise cost, doubting future equity value, and uncertainty about ever getting a chance to sell before an M&A or IPO.
✈️ KEY TAKEAWAYS
Walker's real advice is simpler than fixing the non-exercise pattern: with leaner teams today, each hire carries more weight, so grant equity generously since your team is the most valuable asset.
AI Blurring the Lines Between Jobs
A field experiment with 791 P&G professionals, led by researchers at Harvard and Wharton, and a new OpenAI report on ChatGPT use at work both find AI is letting workers take on tasks outside their formal role.
Individuals with AI Matched Team Output: In a preregistered experiment with 791 P&G professionals, individuals working with AI matched the performance of two-person teams working without AI on real product innovation challenges.
AI Balanced Out Departmental Bias: R&D professionals typically proposed technical solutions and commercial professionals proposed commercially-oriented ones. Adding AI moved both groups toward more balanced proposals, regardless of background.
43.5% of Task-Specific ChatGPT Use Crosses Job Lines: OpenAI's analysis of over 800,000 U.S. ChatGPT messages found 43.5% of occupation-specific use involved tasks tied to a different occupation, rising to 77% for customer experience workers.
✈️ KEY TAKEAWAYS
Both studies point to the same shift: generalists are using AI to take on specialist tasks that once required a handoff. That crossover is itself a demand signal for founders building AI tools around a single function, since people are already trying to do that work without one.
That’s it for today!
Stay driven,
Andre
PS: Join the DDVC Investor Summit at Bits & Pretzels virtually or physically in Munich during Oktoberfest Sep 28-30
