Huge thanks to everyone who joined us for our first gathering in SF this week. From initial idea to full execution, we had a very tight timeline. I am deeply grateful to @ariachen_ai, @js_epi01, and John for their incredible support and for bringing such sharp insights on short notice. Special thanks to @PixVerse_ , @mulerun_ai, @Beta_ucb, @allscaleio for your incredible support of @commonstack_ai 😊 See you next time:)))

Jensen Huang's speech quick summary: -Vertically integrated but horizontally open strategy enables broad ecosystem reach -Computing demand increased 1M times in last 2 years - Current demand visibility: $1T through 2027 (vs $500B through 2026 last year) - Grace Blackwell delivers 35-50x performance improvement over Hopper (vs typical 1.5x Moore’s Law progression) - Vera Rubin platform: 7 chips, 5 rack-scale computers, integrated AI - Tokens are new commodity - data centers becoming “token factories” - Groq Integration & Specialized Computing - OpenClaw & Nvidia Nemo Claw: enterprise-secure version with policy engines, network guardrails - Every enterprise needs “OpenClaw strategy” - SaaS companies becoming “agentic as a service” - Confidential computing enables secure model deployment across regions - Pattern: NVIDIA brings customers to cloud partners through integrated acceleration - Expanding beyond data centers: space computing (Vera Rubin Space One) - Six frontier model families: Nemotron (language), Cosmos (world models), Groot (robotics), others ...

GPT 5.4 Pro is live on Commonstack. It's OpenAI's most capable model. It also costs $30/$180 per million tokens. Not every task needs frontier reasoning. Route the hard stuff to 5.4, and the rest to models that cost 100x less. 1 API. 35+ models. Let routing do the math.

Gemini 3 Flash ranked #1# on PinchBench on OpenClaw agent tasks success rate, followed by MiniMax M2.1 and Kimi K2.5. Don't forget Gemini 3 Flash Lite: same family, built for high-volume workloads at even lower prices. All on

GPT-5.4 & 5.4 Pro are live on > built for AI agents & automation: > 1M token context window > stronger multi-step reasoning > better tool & function calling > more efficient token usage Build your agents with commonstack today