👋 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, the DDVC Landscape Report 2026 just launched! Check out how 345 VC firms are using AI and automation to become more efficient & win against their peers
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.
Do VCs Love or Loathe Your Sector Right Now?
Peter Walker, Head of Insights at Carta, breaks down which sectors investors love and loathe at Series A right now.
1,082 Series A Rounds, AI Infra Off the Charts: Across 1,082 US Series A rounds (July 2025-June 2026), Walker flags AI Infrastructure, Semiconductors, Hardware and Cybersecurity as the sectors investors love most, against a median of $70.0M on a $14.1M median raise.
80% of Software Startups Are Now AI Companies: So far in 2026, about 80% of companies building software products were AI companies, making AI the default starting position for founders.
Food & Bev, Edtech and Medical Devices Lag: Medical Devices has the lowest median valuations of any sector, Edtech has stayed flat despite the AI boom, and Food & Bev rounds are steady in count but low in valuation.
✈️ KEY TAKEAWAYS
Sector rotation at Series A has become sharply concentrated: capital piles into AI Infra, Semiconductors, Hardware and Cybersecurity, while Edtech and Medical Devices stay starved regardless of overall fundraising heat. The 80% figure is a signal worth sitting with for any VC still running a non-AI software thesis.
Capital Allocation Is Dead
Kyle Harrison from Contrary argues that analysis-first capital allocation no longer works, and that investors now succeed by picking one of four aggregation strategies instead.
Four Strategies Replace Analysis-First Investing: Harrison, who has moved through Kickstart, TCV, Coatue, Index Ventures and now Contrary, argues the analyst's extrapolation-based playbook is dead, replaced by aggregating quality, narrative, leverage, or time.
Figma Beat Even the Bullish Case: In a 2018 case study, Harrison projected Figma could hit $200M ARR by 2022 from $4M ARR (valued ~$400M by Sequoia); an analyst dismissed it as too aggressive, yet Figma overshot even that projection.
Mohnish Pabrai's Estimated $2B Miss: Pabrai held roughly 77% of his fund in Micron since 2017, then sold in 2023 right before Micron ran 15x+ in the AI-memory boom, an estimated $2B miss.
✈️ KEY TAKEAWAYS
Harrison's framing puts defensibility at the center of picking winners. Each of the four allocator strategies demands an articulated worldview, something a spreadsheet can't supply on its own. Pabrai's miss shows the risk on the other side: even elite patient-capital investors get punished when analyst instincts override sitting still.
The Making of a Unicorn Founder Is Changing
Ilya Strebulaev from Stanford GSB maps the prior employers of 4,357 unicorn founders and finds the feeder pipeline has shifted hard toward Google, Facebook, and OpenAI.
68.6% of Founding Teams Share a Prior Workplace or School: Among 1,377 multi-founder US unicorns, 68.6% have co-founders who previously worked together or studied together, with shared employers (57.8%) more common than shared schools (33.3%).
OpenAI Alumni Carry the Highest Odds Ratio: Among employers with 10+ unicorn founders, OpenAI alumni are 14.7x overrepresented versus a random sample, ahead of Check Point (11.2x) and D.E. Shaw (9.8x).
The PayPal Mafia Effect Has Faded: PayPal's odds ratio as a founder-producing employer fell from 12.0x pre-2016 to 1.7x since, echoed by Apache Software Foundation (9.8x to 2.2x) and LinkedIn (6.0x to 1.1x), while Google, Facebook and MIT rose sharply over the same period.
✈️ KEY TAKEAWAYS
Sourcing strategy should track today's feeder pipeline, since the alumni networks that mattered a decade ago have shifted underneath it. OpenAI already shows the highest founder-overrepresentation of any employer, while once-dominant signals like PayPal alumni are fading toward baseline. For talent scouts and pre-seed investors, Google, Facebook and OpenAI are the new proxy worth watching.
Join 1,864+ investors in our free Slack group as we automate our VC job end-to-end with AI. Live experiment. Full transparency.
Stop Playing the Mark Up Game
Hemant Taneja, from General Catalyst, argues founders now need engineered capital structures more than advice, and that VC should look to 1980s Wall Street for a precedent.
Capital Is the New Edge: Taneja argues this generation of venture will win on a better-engineered capital stack, since infrastructure and hardware founders need financing that goes past the traditional Rule of 40.
Drexel Burnham Lambert as the Precedent: He points to Drexel's team, which engineered PIK bonds and the high-yield debt market in the 1970s-80s before founding Apollo, Ares and Cerberus, as a model for VCs to innovate on the asset class.
New Capital Types Beyond Growth Equity: He names specific financing gaps investors should fill: capital for sales and marketing, GPUs, energy and power-leasing contracts, advance customer deposits, factories, and public-private infrastructure deals.
✈️ KEY TAKEAWAYS
Taneja's piece pushes VCs to become "capital entrepreneurs" who engineer financing structures as their core value-add. For funds writing infrastructure and hardware checks, the old SaaS-era playbook of counsel and warm intros can't finance today's capital-intensive companies.
WTF Is Storytelling for VCs?
Laurie Owen, from Refinery Media, breaks down why VC storytelling works differently from company storytelling, and gives four repeatable frameworks funds can use to build one.
$274,000 Storyteller Roles and Doubling Job Postings: Owen cites Vanta hiring a head of storytelling for up to $274,000 and LinkedIn "storyteller" job postings doubling over the past year as proof the function is now a real hiring category.
Four Repeatable Story Structures: He outlines four frameworks: Contrast (Founders Fund's stance against incrementalism), Insight (Y Combinator's thesis that great startups don't look great early), Solution (Families Fund's framing around broken institutions), and Path (Indie.vc's alternative to raise-scale-exit).
Zoom In vs. Zoom Out: Company storytelling zooms into a specific problem and roadmap; VC storytelling has to zoom out to how value accrues across a decade, since fund performance is a lagging indicator.
✈️ KEY TAKEAWAYS
This reframes what fund marketing should target: LPs and founders respond to a specific character, conflict, stakes and resolution repeated consistently, well beyond partner bios or "we add value" claims. Committing to one of the four frames is the practical takeaway for any GP rewriting their positioning.
The Switch to Graph Engineering in AI Agents
Peter Steinberger's viral post sparked a wider essay on why AI agent architecture is shifting from single self-improvement loops to networks of loops, and what that shift does and doesn't fix.
A Single Tweet Captured a Field's Shift: Steinberger's post, "Are we still talking loops or did we shift to graphs yet?", gathered thousands of likes, capturing agent builders' recognized shift from single loops to networks of loops ("graphs").
Four Failure Modes of a Single Loop: The essay names four structural failures: Goodhart's law (an optimized metric stops measuring what it should), blindness to whether the target is correct, conflict between independent loops, and undetected measurement decay.
MLOps as the Working Example of a Graph: A mature deployment pipeline is a champion-challenger loop wired to drift-monitors and automatic rollback, with held-out eval sets the training loop is never allowed to see.
✈️ KEY TAKEAWAYS
For investors backing agent infrastructure and AI-ops tooling, this reframes what a defensible "self-improving" system needs: a governed network of loops with independent audits. The harder point, that even a good graph fails without ground-truth anchors, is a useful diligence question for any autonomous self-improvement pitch.
That’s it for today!
Stay driven,
Andre
PS: Try out Granola free for 1 month with code “DATA100OFF”
