$4.5M · Metal + 2 more

What they're building

Metal is fundraising software built by someone who raised $120M and hated every minute of the process. The core observation is one Usman Gul makes about his own Airlift rounds: roughly 80% of the investors he spoke to were never a fit on stage, sector, geography or check size. That is not a hustle problem, it is an information problem, and the industry's answer to it has been spreadsheets and warm-intro folklore. The product attacks it with a proprietary investor memory layer underneath an agent. It identifies and ranks investor prospects against your actual stage and sector, researches the firms, surfaces warm introduction paths through your network, drafts the intro requests, analyzes your pipeline, and captures meeting intelligence after calls. Autopilot handles the back-office side: deck optimization, call prep, round strategy, and reading leading indicators on which VCs are actually deploying right now. The long-term framing is more interesting than the tool. Metal describes its mission as building the rails of the venture industry by bringing efficiency to venture rounds. That is a bid to be infrastructure for capital formation rather than a CRM for founders, and it explains the distribution strategy: Metal is embedded with Techstars, where 10,000+ founders have access, and has 100+ YC founders as paying customers post-Demo Day. It is also worth naming what is unusual here. This is a company headquartered in Lahore, selling AI software to North American venture-backed founders, hiring on-site engineers in Karachi, and co-led by a16z speedrun and YC.

Why this matters

The numbers are the argument and they are the company's own reported figures: 1,000+ founders, $542M raised through the platform in FY2026, six straight quarters of 30-80% QoQ growth. Very few seed-stage companies in this category have compounding usage attached to a dollar outcome that customers can measure. A founder either got the round or did not. The structural case is that venture matching is genuinely mispriced information. Investor stage, sector focus, geography and check size all change quarterly, and the public record of them is stale by construction. Gul's 80%-mismatch figure from his own $120M of rounds is the kind of number that only a repeat founder would bother to quantify, and it is the wedge. If Metal is right that this is rails rather than a tool, then the thing being built is a graph of who deploys into what, refreshed continuously, and that compounds in a way a database of VC email addresses never has.