The Series A bar for AI just tripled — and somehow the loudest rounds this week didn't seem to feel it. This is Startup Fundraising. Today, we're putting a number on what Series A and B are actually buying — and asking who's clearing it versus who's getting waved through. Adam's got receipts. Nearly half of all US venture dollars are chasing AI, the proof bar just tripled — and a stack of rounds this week have zero named customers. Let's figure out which bucket each one belongs in. If AI is hoovering up nearly half of all US venture dollars right now, what are Series A and B investors actually underwriting? Durable ARR, compute-adjusted margins, proprietary data moats — or are they buying a lottery ticket on category ownership before anyone really knows what the pricing model looks like? Honestly, it's all four — but the weighting has swung hard toward proof over promise. The numbers are pretty clear. On the Product Market Fit Show, Carta's Head of Insights cited data showing the median ARR threshold for a Series A is now around $3.5 million, up from about $1 million not long ago. Median raise sizes have moved from $8–10 million to $13–15 million, with post-money valuations in the $75–85 million range. Foundra's 2026 benchmarks stack on top of that: investors want net revenue retention above 100 percent — sometimes 120 percent — gross margins north of 75 percent for software, and month-over-month growth above 15 percent. Otherwise, you're probably not getting the meeting. AI makes the margin piece messier. VC Cafe put it well: AI coding tools make software cheaper to build, but building a fundable company is harder, because GPU costs and inference spend can quietly wreck a margin profile that looks clean from the outside. CRV's investment criteria framework lands in the same place. AI SaaS gets evaluated differently from traditional SaaS because compute isn't fixed overhead; it rises with usage. So investors are now stress-testing whether gross margins still hold when the product scales. Proprietary data and category position can get the conversation started. You still need unit economics that survive diligence. So if the ARR bar has tripled but AI companies are still commanding the lion's share of capital, are non-AI SaaS founders basically being told the old benchmarks don't even get them in the room anymore? Yeah, that's the read — and Carta's data is pretty blunt: outside AI infrastructure, founders face a harder, more expensive path than the headlines make it sound, even when they hit the old benchmarks. AI is capturing roughly 40 to 45 percent of all US VC investment, per the 2026 state-of-the-market data, and the seed-to-Series A graduation rate is sitting around 20 percent even for funded companies. So yes, the squeeze is real. The next pressure point is Series B and the Rule of 40. Investors are already asking which margin definition founders are using to calculate it, which tells you compute costs are the hidden variable everyone's trying to get in front of. If Startup Fundraising helps you stay sharp, take a second to subscribe wherever you're listening. And if you can leave a quick review, it really helps other founders and operators find the show.
You'll find links to every story we covered in the show notes. So if one of today's items is useful for your next raise or your next board conversation, it's there to dig into. That's Startup Fundraising for today. This is a Lantern Podcast.