DVC, a San Francisco–based VC focused on early-stage AI startups, has launched its State of AI Report — a continuously updated analysis of the artificial intelligence landscape, built on over a decade of AI investing experience by co-founders Nick Davidov and Marina Davidova.
The report combines meta-analysis of leading research with proprietary insights from a community of 200+ operators and investors. Unlike traditional static reports, it functions as a live product: updates are generated by AI and manually reviewed, reflecting real-time shifts across infrastructure, models, and application layers.
Why AI Is Being Mispriced
One of the report’s key insights focuses on how the market evaluates AI companies. Traditional SaaS valuation models assume high long-term margins (up to ~80%), driven by scalable software economics. AI startups operate differently. They face a persistent “inference tax,” where 40–60% of revenue is consumed by infrastructure costs, including compute, foundation models, cloud, and energy. As a result, margins across much of the AI stack resemble industrial businesses: thinner and dependent on scale. Value is shifting elsewhere — toward control over user behavior and distribution, interface ownership, and speed of market expansion — while upstream layers commoditize.
“NVIDIA’s Jensen Huang put it nicely: think of a datacenter as a modern-day factory. So your ‘AI agent for banking AML and compliance’ is not SaaS — it’s closer to a silverware customization service trying to build a brand. The spoons are mass-produced in a factory from sheet metal, sourced from mills that process ore from mines. Metal and electricity go in — consumer emotions (and some utility) come out. However, this doesn’t mean that paying a SaaS multiple for an application-layer or agentic AI startup today is wrong. But let’s be honest — we’re not paying a premium for long-term terminal value (as with SaaS), but for much higher revenue growth potential,” says Nick Davidov.
A New Scaling Dynamic
The report also highlights a shift in how quickly companies scale. Small teams are reaching hundreds of millions in revenue and competing with incumbents much earlier. Examples include Perplexity AI reportedly adding 50% to $300M ARR in a month, and Higgsfield scaling from zero to $300M within a year.
The full report is available at https://state-of-ai.dvc.ai/
Alongside the launch, DVC is announcing The AI Rabbit Hole — an invite-only conference on June 5 in San Francisco, where many of the report’s themes will be explored in depth with founders, researchers, and investors shaping the field. This year’s theme is AI Bubble Tea Party.
