General Compute has opened its production inference cluster to developers building agent applications, running SambaNova SN40 and SN50 dataflow silicon that posts the fastest independently benchmarked speeds on the MiniMax M2.7 model family.
General Compute today announced the general availability of General Compute Cloud, the first ASIC-native neocloud purpose-built for the next generation of autonomous AI development tools. Where existing neoclouds rack incumbent GPUs, General Compute has designed its serving stack around its inference-optimized ASICs.
Unlike traditional cloud platforms designed around human operators clicking through dashboards, General Compute Cloud is also the first major cloud to treat AI agents as first-class users. Coding agents can complete the entire onboarding flow themselves, creating an account, claiming the launch credit, and retrieving a working API key, without requiring a developer to step in. The result: a developer can ask their AI agent to “switch its inference to General Compute,” and minutes later be looking at running infrastructure they never had to provision by hand.
Also Read: AIThority Interview With Rohit Agarwal, Founder & CEO of Portkey
Agentic workloads issue dozens or hundreds of model calls per task, which means even small per-token gains compound into dramatic differences in developer experience and unit economics. By optimizing the silicon, runtime, and API surface for inference rather than retrofitting general-purpose accelerators, General Compute aims to push down both first-token latency and sustained token throughput on the open and frontier models developers use most.
Agent-native signup is supported out of the box. When invoked through OpenCode, OpenClaw, or any compatible client, the agent can complete account creation, verify the workspace, claim launch credit, and return a scoped API key back to the developer’s environment – turning what was historically a multi-step onboarding into a single natural-language instruction.
“Our goal is simple: we want General Compute to be the fastest inference provider on the market, and we want to ship the fastest inference API any developer or AI agent can call,” said Jason Goodison, CTO and co-founder of General Compute. “Optimizing the silicon is how we get there. The $200 in launch credit is our way of inviting builders, and their agents, to put us up against anyone else and see the numbers for themselves.”
General Compute Cloud is available immediately to customers globally at generalcompute.com. The launch credit is automatically applied to new accounts created between May 20 and May 27, 2026. OpenCode and OpenClaw users can begin a General Compute session directly from within their agent by asking it to “sign me up for GeneralCompute.com”
Also Read: AI-Driven Risk Intelligence: How FIs Are Predicting Systemic Shocks
[To share your insights with us, please write to psen@itechseries.com]
Comments are closed.