General Compute Secures Up to $400 Million to Scale the World’s Fastest Inference Neocloud

General Compute, the world’s fastest inference neocloud, today announced a committed debt facility of up to $400 million from Upper90 Capital Management, LP (“Upper90”). The financing begins at $100 million and scales with customer demand, allowing General Compute to quickly build one of the world’s largest inference neoclouds to deliver AI results 16x faster than current GPUs.

$400M Debt Facility to Build the Next Generation of AI Compute

The financing is one of the largest to date for an ASIC (non-GPU) cloud. General Compute holds no GPU allocation to protect. That frees it to build on the fastest independent inference silicon while GPU clouds stay tied to a single supplier for chips, financing, and roadmap. The company has more than $300 million of secured, price-protected supply and expects to be the first neocloud to deploy ASIC silicon at scale. Upper90 is also an equity investor in General Compute, aligning its debt and its equity behind the company.

“Upper90 has been early to the AI infrastructure opportunity while remaining selective about where we deploy capital. We focus on differentiated approaches that make compute cheaper, faster, and more accessible for customers. General Compute’s partnership with SambaNova gives it access to inference silicon that is up to six times more power efficient than traditional GPUs. Our role is to ensure the company has the capital it needs to scale alongside rapidly growing demand.” Billy Libby, Co-Founder & Chief Executive Officer, Upper90

The GPU Dilemma: Poor Inference Performance & Large Power Requirements

As AI usage skyrockets globally, Goldman Sachs predicts token consumption will grow 24x in the next 3.5 years alone. While the first generation of GPU neoclouds helped alleviate the AI training bottleneck, they have proven inefficient at inference, the process of actually delivering queries to end users. The newest GPU racks demand upwards of 120 kilowatts of power and require specialized liquid-cooling retrofits, trapping new capacity behind multi-year data center construction timelines.

General Compute’s Solution: The World’s Fastest Inference Neocloud

General Compute solves these problems by building an inference neocloud designed to deliver AI results as quickly and efficiently as possible. The cloud leverages specialized inference silicon from SambaNova (SN40 and SN50) that are more efficient than GPUs at delivering tokens, operating on only 20kW per rack, with no water-cooling required. The result is a better user experience with AI responses as quick as human conversation.

Key features include:

• Unmatched Speed: 16x faster inference performance and 7x faster time-to-first-token compared to standard GPU clouds
• Higher Bandwidth: 8.5x higher output throughput than standard GPU clouds, reaching 1,000 tokens/second
• Rapid Deployment: Air-cooled chips require no specialized data center infrastructure, allowing installation in weeks instead of years
• Power Efficiency: 6x less energy required than current GPUs, allowing radically more affordable inference
• Universal Compatibility: Switch to General Compute in less than 30 seconds and run frontier models from OpenAI, DeepSeek, MiniMax, and others
• Agentic-First: API & MCP compatible to allow agents to sign up on your behalf

“We are the only neocloud that can serve premium tokens: frontier-level intelligence on the largest models, served fast,” said Finn Puklowski, co-founder and CEO of General Compute. “That is where the workloads are moving. Because we are new, we are not handcuffed to a single chip supplier. This facility lets us deploy the fastest inference silicon into existing data centers and meet that demand now.”

Companies can switch their fast inference workloads to General Compute today at https://www.generalcompute.com.

About General Compute

General Compute is the fastest inference cloud for AI agents. It runs production AI workloads on specialized inference silicon from SambaNova, air-cooled hardware that deploys into existing colocation facilities in weeks rather than the years a liquid-cooled GPU build requires. Earlier this year, General Compute closed $15 million in seed and pre-seed funding from FUSE VC, Village Global, Carya Venture Partners, NZVC, Matterscale Ventures, Mana Ventures, and Upper90.

About Upper90

Upper90 is a hybrid investment firm providing credit and equity to Seed - Series B technology companies. Founded in 2018 by Billy Libby and Jason Finger, the firm manages more than $1.2 billion in assets and has developed innovative capital solutions for category defining companies including Crusoe, Stax Engineering, DeepInfra, Octane, PayJoy, Apptronik, and Clutch. By pairing flexible credit with equity for alignment, Upper90 helps founders scale faster with less dilution.

Contact Information

Riley Munks

PR Advisor

Activate PR - https://www.activate-pr.com/

riley@activate-pr.com