Nscale seeks $3.5B pre-IPO financing as Anthropic deal doubles backlog

This massive capital injection signals a shift toward specialized AI hyperscalers capable of absorbing high-density compute demands that traditional cloud giants may bypass due to risk or capacity constraints. For the streaming and media ecosystem, Nscale’s rapid scaling of GPU campuses in Texas and North Carolina provides the necessary infrastructure for generative video and real-time AI processing at scale. Nvidia’s participation highlights a growing trend of circular financing where hardware providers fund the very platforms purchasing their silicon. Watch for the final IPO pricing to see if public markets validate the $30 billion valuation cap set by private noteholders.

Nscale's fundraising effort arrives amid a broader wave of GPU cloud providers racing to secure capital before public listings. In June 2026, Ericsson launched its AI in RAN commercial software subscription claiming up to 20% higher downlink throughput across more than 15 live deployments, demonstrating how AI infrastructure demand is spreading beyond hyperscale cloud into telecom networks. Nscale's $103 billion contract backlog, anchored by its Anthropic compute agreement, positions it as one of the few independent GPU cloud providers with the scale to serve frontier model developers. The company's planned U.S. listing would place it alongside a growing cohort of AI infrastructure firms seeking public market validation after years of private capital accumulation.

Nvidia's $2 billion commitment to Nscale reflects a pattern of circular financing in the AI hardware supply chain. Nokia announced work with AWS and Databricks to build the data, cloud, and control layers for autonomous networks at DTW Ignite in June 2026, illustrating how hardware and platform vendors are stacking partnerships to lock in ecosystem demand. For Nscale, Nvidia's participation as both supplier and investor creates alignment on Vera Rubin GPU deployment timelines while raising questions about whether the chipmaker's capital is effectively subsidizing demand for its own products. Third Point's $1.5 billion convertible note tranche, led by Daniel Loeb, signals that activist investors see the AI infrastructure buildout as a durable theme rather than a speculative bubble.

The technical architecture Nscale is building targets high-density GPU clusters optimized for training and inference workloads that strain conventional cloud providers. Ericsson and Nokia are diverging on AI-RAN strategy, with Nokia running all Layer 1 functions on Nvidia GPUs while Ericsson limits GPU use to forward error correction, showing how GPU-accelerated compute is reshaping infrastructure decisions across industries beyond AI model training. Nscale's planned campuses in Texas and North Carolina will house Blackwell-series and next-generation Vera Rubin GPUs, with the company's contract backlog implying utilization rates that exceed what most traditional data center operators have committed to publicly. The $30 billion valuation cap embedded in the convertible notes will face scrutiny when Nscale files for its IPO, particularly if public market investors apply the same discount to circular financing structures that they have applied to other GPU cloud providers.