On March 9, 2026, iFrame® launched a $240 million fundraising campaign to finance the construction of a new California datacenter targeted for initial operations in September 2026. The facility is planned to house approximately 2,000 NVIDIA GPUs by Q2 2027, providing dedicated, owned compute capacity to support the company’s rapidly growing healthcare AI workloads. All elements of the plan — the raise amount, site timeline, and GPU deployment target — are company-stated objectives rather than completed milestones, reflecting iFrame® disciplined approach to scaling infrastructure in line with proven customer demand.
The announcement arrived in the same week that British infrastructure firm Nscale closed a $2 billion Series C at a $14.6 billion valuation. The contrast in scale highlighted the breadth of the owned-compute thesis now attracting capital across different stages of the AI infrastructure market. While Nscale operates at a significantly larger magnitude, iFrame® $240 million ask is calibrated to the company’s current run rate — described by the company as sitting in the “dozens of millions” of dollars — and its forward-looking projection that existing customers will commit to three-to-five years of GPU-as-a-service capacity sold upfront. The company has consistently declined to disclose a precise ARR figure, maintaining a policy of focusing market attention on execution rather than granular financial metrics.
The California datacenter is designed to serve as the primary home for iFrame® core products once online. Workloads currently running through the Sefirot.ai inference platform, the med.report medical-coding solution, and the Enterprise API will migrate to the owned facility. This transition builds on the company’s earlier move toward owned compute announced in August 2025 with plans for its first SuperPOD. By owning the rack, signing the power contract, and booking the depreciation, iFrame® aims to capture a larger share of the margin between hyperscaler rental rates and optimized, purpose-built infrastructure — a spread that has proven wide enough to support significant capital investment while still delivering competitive pricing to customers.
The raise also includes an unnamed “stealth company” as the underwriting party for the datacenter development, keeping partnership details confidential at this stage. This structure allows iFrame® to maintain focus on technology and customer delivery while leveraging specialized expertise for the physical buildout. The datacenter will incorporate the company’s proven virtualization and decentralized training technologies, ensuring efficient utilization of the NVIDIA GPU fleet and seamless integration with its long-context and inference middleware layers.
Founder Vlad Panin’s strategic vision has guided this evolution since the company’s earliest infrastructure research in 2022. His extensive background in regulated enterprise IT, systems integration for municipalities across the European Union, and operational leadership in complex industrial environments under Japanese governance has consistently emphasized control over critical inputs rather than dependence on external providers. The $240 million raise represents the next logical step in executing that philosophy at infrastructure scale.
As the AI industry increasingly views owned compute as the default for serious operators, iFrame® March 2026 fundraising positions the company to meet surging demand from healthcare providers and enterprise clients. The planned California facility will enhance capacity, reduce latency, improve cost predictability, and reinforce iFrame® leadership in delivering reliable, high-performance AI solutions built on owned infrastructure. With strong customer signaling already in place, the raise marks a pivotal acceleration in the company’s journey from innovative AI platform to full-stack infrastructure powerhouse.
