The financing brings Gruve’s total funding to $87.5 million and was led by Xora Innovation (backed by Temasek), with participation from Mayfield, Cisco Investments, Acclimate Ventures, AI Space and other strategic investors.
The capital accelerates Gruve’s ability to make low-latency AI inference capacity immediately available across Tier 1 and Tier 2 U.S. cities and scale efficiently as demand grows—without multi-year data center buildouts.
The Execution Gap in AI
As inference becomes the dominant AI workload, infrastructure has emerged as the industry’s primary constraint. While models, agents, and hardware continue to see breakthroughs, the systems running them have not kept pace.
Most production inference today relies on infrastructure that was never designed for low-latency, high-throughput, cost-sensitive AI, resulting in unsustainable costs, mounting technical debt, and weak unit economics.
Gruve’s Inference Infrastructure Fabric was built to close this gap.
Inference Infrastructure Services Purpose-built for Production AI Workloads
Gruve’s Inference Infrastructure Fabric is a distributed platform engineered specifically for production-grade AI inference, delivering predictable latency, scalable throughput and industry leading economics.
Key capabilities include:
- 500MW+ of expandable U.S. capacity, leveraging excess power and existing infrastructure near Tier 1 and Tier 2 cities, enabled by long-term partnerships with Lineage, Inc. and other major colocation providers
- Modular, high-density, rack-scale inference capacity, engineered for cost efficiency in inference-heavy workloads and rapid deployment
- A distributed, low-latency edge fabric for seamless connectivity and workload orchestration across sites
- Full-stack operations, including a 24×7 AI-powered SOC, network services, and cluster management to meet enterprise-grade reliability and performance standards
Gruve is bringing 30MW live today across four U.S. sites, with additional capacity under development and further near-term expansions in Japan and Western Europe. This unique approach bypasses multi-year data center build cycles and delivers AI-ready capacity in months instead of years.
“Gruve’s Inference Infrastructure Fabric combines modular state-of-the-art pods with a distributed network architecture to enable rapid capacity deployment in power- available locations today — without compromising on latency. As demand for inference accelerates, scalable, low-latency infrastructure with strong unit economics is increasingly critical, and Gruve is well position to meet that need as it scales in 2026.”
— Phil Inagaki, Managing Partner and Chief Investment Officer, Xora Innovation
“We’re launching our Inference Infrastructure with 30MW across four U.S. sites, immediate capacity available nationwide, and near-term expansions in Japan and Western Europe. Combined with our 24×7 AI-powered SOC, inference fabric and infrastructure operations, Gruve is ready to support customers at true production scale.”
— Tanuj Mohan, GM & SVP, AI Platform Services, Gruve
