Lead the seed at $4M, $24M post-money. 18 of 24 diligence items closed; remaining items are commercial loss-reason analysis (Marc), open-source contracts review (Counsel), and a hiring-plan doc from the founders. None are blockers, all expected to close before IC May 22.
AI-infra cohort, vertical 4 (inference orchestration). Aligns with our December 2025 thematic note. Memory recall surfaces two prior pass decisions in this category for traction reasons. Vertex meets the bar that Helio Labs missed in November 2025.
AI inference layer projected at $2.4-3.8B in 2026, growing into $8B+ by 2028 if model-provider passthrough persists. Three direct comps (Portkey, Helicone, Martian Labs); each occupies a slightly different wedge. Vertex's wedge is heuristic-driven model selection per request, which we believe is the highest-leverage cut.
Two founders. CEO Aanya Iyer (Stripe payments routing, 5 years) is the systems thinker. CTO Devi Patel (Anthropic infra, 3 years) is the model-runtime expert. Six prior co-workers backchanneled, all positive. One concern surfaced: founder velocity has been 'frenetic' the last 60 days; recommend addressing in next board call.
ARR ~$140k from a $0 base in January 2026. 22% MoM in last 3 months. 3 production customers on annual contracts, 1 of which is a Veridian portco (Stratus Cloud). Net dollar retention not yet calculable at this size.
1. Model-provider concentration: 71% of routed requests go through one provider. Mitigation: Vertex is in active integration with two more by Q3. 2. Margin compression risk if model providers drop wholesale prices below Vertex's blended rate. Mitigation: Vertex has audited cost model; resilience scenario analyzed. 3. Founder bandwidth: see Team note.
Memo language has been checked for: securities-recommendation framing, IP-claim accuracy, third-party-source attribution. Status: pre-flagged clean; counsel approval auto-requested on send.