Yoni Rechtman

SandHill has indexed 165 posts from Yoni Rechtman, most recently in 13 Nov 2026.

Where they publish

99% Derisible ↗

What they write about

Venture Capital & Startups · AI (Horizontal) · Data & Analytics · AI Infrastructure · Robotics

Recent posts
  • Note to the team↗ 13 Nov 2026
    As Nvidia's dominance faces pressure and the AI capex boom unwinds, a VC argues investors must reposition from financial frenzy to deployment phase—choosing between aggressive down-round buying, steady dollar-cost averaging, or strategic thesis work and retrenchment ahead of application-layer opportunities.
  • Data to deployment↗ 6 Nov 2026
    The data market is immature but structurally important; raw data has little value, but transformed data—tasks, RL environments, traces from deployed AI—will become a durable commodity-like market bifurcating into synthetic-at-scale and scarce out-of-distribution segments, reshaping how firms accumulate and finance data as capital.
  • Our 2026 LP Meeting↗ 30 Oct 2026
    A VC firm's annual LP presentation distills their 2026 investment thesis: neutral on AI itself, long on outcomes-based software, second-order effects of AI (trust, security, attention), supply-chain bottlenecks, and post-AI founders. Notably skeptical of raw capability plays and infinite-capital-consumer businesses.
  • Data-native capital is coming↗ 23 Oct 2026
    Data locked in corporate balance sheets is newly monetizable for AI training, creating a trillion-dollar arbitrage opportunity. Capital markets will financialize this hidden asset through new structures—from data-backed credit to PE strategies—reshaping both business models and financial engineering.
  • Venture’s chivalric myths↗ 16 Oct 2026
    Venture capital's informal "chivalric" norms that govern deal-making are breaking down as markets globalize, rounds financialize, and pro rata rights become legally enforced rather than socially enforced—a clash between Silicon Valley's relationship-based culture and global capitalism's rules-based system.
  • Data beyond brokerage↗ 9 Oct 2026
    Data sales to AI labs are a lucrative but temporary funding opportunity; founders should treat data revenue as capital to fund durable businesses rather than as an end goal, with the rare exception of companies positioned to generate proprietary data as operational exhaust.
  • Do you know a guy?↗ 2 Oct 2026
    As AI agents proliferate and enable ruthless optimization at scale, the economy will shift from efficiency gains to zero-sum competition—deals vanish, subsidies disappear, and advantage consolidates around proprietary relationships ('knowing a guy') rather than exploitable hacks.
  • Safetymogging and marginmaxxing the accelecels ↗ 18 Sept 2026
    AI safety rhetoric from labs like Anthropic rings hollow when coupled with financial incentives to slow competitors rather than genuinely de-risk development; the field must build credibility through affirmative cases for physical AI deployment, not doom narratives, while creating markets around safety and compliance infrastructure.
  • Be dumb. Waste Money. Win.↗ 11 Sept 2026
    Early-stage founders should embrace high-risk experiments and
  • Our 2Q letter to LPs↗ 21 Aug 2026
    A venture fund's LP letter argues that AI's real value will accrue to users and builders of AI-enabled businesses, not infrastructure vendors—and that the best investments target newly possible outcome-based businesses rather than trades on AI hype.
  • What capital wants↗ 14 Aug 2026
    Fundraising success now hinges on a specific gate: demonstrating a path to a $20B+ market where you can win. Vertical expertise matters less than finding true believers or attaching to category-agnostic mega-trends; complex narratives are liabilities against clear metrics.
  • No revenue is enough revenue↗ 7 Aug 2026
    Series A funding thresholds have become undefined and dynamic. Instead of hitting a specific revenue number, founders need rapid growth, simplicity of narrative, and obvious category potential—and even that varies by investor and market cycle.
  • Get rid of 90 day windows 31 Jul 2026
  • Open models for open America 24 Jul 2026
  • Back in the saddle 17 Jul 2026
  • Make a bet 26 Jun 2026
  • Eyes on the End Game 19 Jun 2026
  • We had it coming 12 Jun 2026
  • We're all short intelligence 5 Jun 2026
  • Slow Takeoff: 2026 29 May 2026
  • AI Accenture, Not Accenture for AI 22 May 2026
  • Take the dive 15 May 2026
  • Slop 'em up 8 May 2026
  • I don't want UBI 1 May 2026
  • Context is the fifth primitive 24 Apr 2026
  • All you need is terminal value 17 Apr 2026
  • Don't ask the group chat for permission 10 Apr 2026
  • There will only be four jobs 3 Apr 2026
  • Agency, not agents. 20 Mar 2026
  • Don't call it a moat 13 Mar 2026

All writers · all sources · by topic