Main highlights:
- ASI:Cloud is now live, giving developers unauthorized access to enterprise-grade GPU computing without the need for KYC.
- Developers can run inference across leading open source models for as low as $0.07 per million input tokens.
- Built by SingularityNET and CUDOS, ASI:Cloud addresses cloud capacity constraints and cost barriers in the AI market.
ASI: Cloud goes live with permissionless AI access
ASI:Cloud, the decentralized AI computing platform developed by the Artificial Super Intelligence (ASI) Alliance, has officially exited beta and is now processing live workloads. The platform provides developers building AI applications with a permissionless gateway to high-performance GPU infrastructure without the need for KYC or traditional cloud accounts.
The launch of ASI:Cloud comes at a critical moment as enterprise demand for production-grade AI is surging and Nvidia H100 availability is locked in through most of 2026. Unlike centralized providers, this platform consolidates GPU access, model inference, and billing into a single permissionless environment.
“Every GPU cycle brings us closer to sustainable distributed intelligence. ASI:Cloud represents a vision of becoming operational infrastructure, not an aspiration.”
—Ben Goertzel, CEO of Singularity
Cost-effective inference with FET and open source models
Built through a partnership between decentralized AI project SingularityNET and global computing provider CUDOS, ASI:Cloud supports inference across top open source models including Llama 3.3 70B, Qwen 3 32B, and Gemma 3 27B. Pricing starts at $0.07 per million input tokens, significantly lower than those offered by AWS and Azure. Comparable GPU instances can often cost up to $6.98 per hour on AWS and Azure, not including additional storage and bandwidth charges.
Developers can use Web3 wallets to access the platform and pay fees in FET and stablecoins. Fiat payment options will also be added soon. ASI:Cloud supports crypto-native billing to simplify access in line with broader goals for decentralized infrastructure.
“Enterprises face real constraints in the centralized cloud market, with capacity shortages and vendor lock-in holding back AI innovation at the very moment when demand is exploding. ASI:Cloud combines CUDOS' enterprise-grade compute infrastructure with SingularityNET's AI backend. This ensures builders have access to high-performance computing without the bottlenecks that define traditional providers.”
—Luke Gniwecki, Head of CUDOS AI Computing Products
conclusion
As GPU access becomes increasingly scarce in the traditional cloud ecosystem, ASI:Cloud introduces a decentralized, cost-effective option tailored for the AI development community. Powered by SingularityNET's model infrastructure and CUDOS' global computing power, the platform provides a scalable alternative to centralized cloud services with support for widely used models and predictable, competitive pricing. As distributed AI gains momentum, ASI:Cloud could be a turning point in how AI workloads are deployed at scale.
