Andreessen Horowitz general partner Jennifer Li is pushing back against the ARR arms race gripping AI startups. In a TechCrunch Equity podcast appearance, Li - who oversees some of the firm's fastest-growing AI companies including Cursor and ElevenLabs - warned founders to stop stressing over the eye-popping revenue numbers flooding social media. The issue isn't just inflated claims, it's that many founders are confusing revenue run rate with actual annual recurring revenue, creating what Li calls "a lot of anxiety" among inexperienced builders who think they need to hit $100 million before raising a Series A.
Silicon Valley has seen gold rushes before - the AI investing boom is just the latest Big New Thing attracting mountains of venture capital. But there's something completely new about this cycle: startups rocketing from zero to $100 million in annual recurring revenue in a matter of months.
Andreessen Horowitz general partner Jennifer Li thinks it's time for a reality check. Li, who helps steer the firm's infrastructure team and oversees portfolio companies like Cursor, ElevenLabs, and Fal.ai, is seeing too many founders spiral into anxiety over Twitter bragging that doesn't match reality.
"Not all ARR is created equal, and not all growth is equal either," Li told TechCrunch's Equity podcast. She's especially skeptical of founders announcing spectacular numbers in tweets - and she's got good reason to be.
Here's the thing: there's a legitimate accounting term called annual recurring revenue. It measures the annualized value of contracted, recurring subscription revenue - money that's essentially guaranteed because customers are locked into agreements. But what many founders are actually tweeting about is revenue run rate - taking whatever cash came in during a hot month or quarter and multiplying it out to look like an annual figure.
Those two things aren't remotely the same. "There's a lot of missing nuances of the business quality, retention, and durability that's missing in that conversation," Li explained. A startup might have crushed it in January with pilot program signups, but that doesn't mean February will repeat the performance. And pilot customers? They're not exactly committed long-term.












