AI Infrastructure Strain Spurs Singularity Compute to Deploy GPU Cluster in Sweden

The rapid surge in AI adoption over recent months has made one thing abundantly clear: there simply isn't enough computational power to go around. This shortage has become painfully obvious, with cloud providers accumulating months-long waitlists for access to high-end GPU instances. Unlike the fleeting GPU craze driven by cryptocurrency mining a few years ago, today's crunch is fueled by genuine, sustained demand from AI research and real-world deployments.

To put this in perspective, Amazon Web Services has been charging around $98 per hour for an 8-GPU server equipped with NVIDIA's top-tier H100 chips. In stark contrast, some decentralized GPU platforms offer comparable hardware for as little as $3 per hour. Amid this dramatic 30x price gap, Singularity Compute—the infrastructure division of decentralized AI pioneer SingularityNET—has announced the Phase I deployment of its first enterprise-grade NVIDIA GPU cluster. This state-of-the-art installation is located in a cutting-edge data center in Sweden.

Through a partnership with Swedish operator Conapto, Singularity's cluster leverages next-generation NVIDIA hardware, including the H200 and L40S GPUs. The Stockholm-based facility is powered entirely by renewable energy, aligning computational power with sustainable practices.

What exactly is being offered?

Designed as a high-density solution, this cluster forms the foundation for both traditional enterprise workloads and projects within the Artificial Superintelligence (ASI) Alliance—a decentralized AI ecosystem led by SingularityNET. It provides flexible access modes tailored to modern AI developers. Companies can rent entire bare-metal machines, launch GPU-powered virtual machines, or utilize dedicated API endpoints for AI inference tasks.

In practical terms, this means an organization can train large machine learning models from the ground up, fine-tune existing models on custom datasets, or run intensive inference workloads for generative AI applications—all powered by Singularity's infrastructure.

On the operational side, the partnership will be managed by Cudo Compute, a well-known cloud provider and NVIDIA partner. Cudo will ensure the cluster delivers the enterprise-grade reliability and timely support required by mission-critical AI projects. Commenting on this development, Dr. Ben Goertzel, founder of SingularityNET and co-chair of the ASI Alliance, stated:

"As AI accelerates toward AGI and beyond, access to high-performance, ethically aligned compute is becoming a defining factor in who shapes the future. We need powerful compute configured to interoperate with decentralized networks running a rich variety of AI algorithms serving diverse populations. Our new GPU deployment in Sweden is a significant milestone on the path to a truly open, global Artificial Superintelligence."

Singularity Compute CEO Joe Honan echoed this sentiment, emphasizing that the launch represents more than just additional compute capacity. It's a step toward a new paradigm in AI infrastructure. He noted that the cluster's NVIDIA GPUs will deliver the performance and reliability modern AI demands while upholding principles of openness, security, and sovereignty in how computational resources are provisioned.

In this broader context, the Swedish cluster is set to serve as the backbone for ASI:Cloud—Singularity's new AI model inference service developed in collaboration with Cudo. ASI:Cloud provides developers with wallet-based access to an OpenAI-compatible API for model inference, offering a seamless path to scale from serverless functions up to dedicated GPU servers.

Early customers are already being onboarded to the Swedish cluster, with the team indicating this is just the beginning of planned hardware expansions and new geographic locations. For a community consistently at the forefront of the AI and blockchain revolution, this deployment represents a tangible step toward the long-held vision of a decentralized, globally distributed AI infrastructure.

The race for AI compute is accelerating rapidly

Since the start of this decade, the tech sector has invested heavily in AI infrastructure, with 2025 alone witnessing over $1 trillion in new AI-focused data center projects. Even nation-states are entering the fray; France, for example, recently unveiled a surprise €100+ billion plan to bolster its AI infrastructure.

However, not every entity can spend billions to address the current compute shortage. This reality has spurred the emergence of alternative approaches, such as decentralized or distributed GPU networks that can aggregate hardware spread across numerous locations and operators.

In essence, if the 2010s rewarded those who amassed data, the 2020s appear poised to reward those who control compute power. Within this emerging future, initiatives like Singularity Compute's new GPU cluster reflect a growing determination to democratize who gets to shape AI's next chapter—primarily by expanding and diversifying the sources of computational power. The road ahead promises to be fascinating.