Now is the Time for the App Layer | OpenAI & Anthropic Won't Win the App Layer | Mike Mignano, USV
TL;DR
Mike Mignano argues that the AI infrastructure buildout is complete, making now the ideal time to build at the application layer, while predicting the model landscape will either consolidate through recursive self-improvement or commoditize into an S-curve plateau favoring open weights and cost optimization.
🏗️ 🏗️ The Application Layer Inflection 3 insights
Infrastructure buildout is complete
Billions have been spent building foundational models (OpenAI, Anthropic, xAI), creating a 'broadband moment' where the technology is now accessible for application development, similar to the internet post-fiber.
Thesis-driven investing wins
With thousands of potential AI applications emerging, venture success requires having a specific thesis and knowing what you're looking for, rather than consensus-driven bets.
Always-on context is coming
Products will increasingly push toward persistent listening and contextual awareness, moving beyond wake words to continuous ambient intelligence in meetings and workflows.
🧠 🧠 Model Landscape Predictions 3 insights
Recursive self-improvement vs. S-curve plateau
The industry faces two futures: labs achieving runaway recursive improvement (winner-takes-all superintelligence) or hitting architectural/data limits that cause plateau and commoditization.
Open weights win in plateau scenarios
If models plateau, enterprises will optimize for cost over capability, driving adoption of open-weight models, routing layers for token optimization, and distributed compute.
Harnesses capture the value
Winning applications will be 'harnesses'—tightly coupled interfaces like Claude Code or Hermes that create flywheels between specific models and user workflows, tightly coupling product and model.
🚀 🚀 Founder & Investment Strategy 3 insights
Constraints drive excellence
Mignano cites doing Anchor's best work when three months from running out of cash, arguing that fear of failure combined with ambitious missions produces the best outcomes.
Ship your thesis publicly
Effective seed investors must publicly share their theses (like USV's 'Rebel Alliance' for open-weight models) to act as bat signals for the right founders, even if ideas are imperfect or incomplete.
Maximize token spend now
Startups should currently prioritize maximizing token spend to gain speed and capability advantages, rather than prematurely optimizing for cost during this critical build phase.
Bottom Line
Founders should aggressively build AI applications now that leverage the completed infrastructure layer, focusing on creating 'harnesses' that tightly couple models to specific workflows while maximizing token spend to move fast before the market commoditizes.
More from 20VC with Harry Stebbings
View all
The $100,000 token budget EVERY engineer will need | Sierra Co-Founder
Sierra Co-Founder Clay Bavor explains why the future of enterprise AI involves mixing frontier and open-weight models, predicts unbounded demand for frontier intelligence despite cost pressures, and reveals how AI-native engineering teams are achieving 3-20x productivity gains.
Coinbase Cuts AI Spend by 50% | Kalshi's $40B Valuation & Impending IPO | The Year for SaaS Roll-Ups
Coinbase's reduction of AI spend by 50% through open-source adoption signals a broader enterprise shift from experimental 'token maxing' to rigorous cost discipline, raising urgent questions about the revenue sustainability of frontier AI models facing cheaper commoditized alternatives.
Bloom Energy CEO: Why Electricity, Not AI Models, Will Decide the Winners of the AI Race
Bloom Energy CEO KR Sridhar argues that electricity infrastructure, not AI models, will determine the winners of the AI revolution, sharing how his NASA background and Andy Grove's 'walk the floor' mentorship shaped his 25-year mission to power the digital age.
How We Got Fred Wilson, Benchmark and Index to Invest $94M | Why Robinhood's Strategy is Wrong
FOMO CEO Paul Erlang reveals how the mobile trading app raised $94M by rejecting traditional venture capital until post-product-market fit, instead leveraging 140 angel investors for distribution while motivating a lean 17-person team with founder-level equity stakes and no internal hierarchy.