Facebook
Britain's News Portal
Around The Clock
BREAKING
Loading latest headlines…

AI Inference Chips Secure £310m Loan, Signalling Next Tech Investment Wave

A £310 million loan has been secured by AI inference cloud startup General Compute, backed by specialized chips. This deal highlights a growing trend in AI infrastructure financing, moving beyond traditional GPU investments.

  • General Compute secured a $400 million (approx. £310 million) loan from Upper90, a tech investment firm.
  • The loan is reportedly the first to use AI inference-specific chips as collateral, rather than the more common training GPUs.
  • This move signals a shift in AI financing towards more cost-efficient infrastructure for running open-source AI models.
  • Inference chips, like SambaNova's SN50s, are designed for power efficiency and faster deployment, offering an alternative to Nvidia's ecosystem.
  • The deal could mark a fragmentation of Nvidia's market dominance as capital organises around new chip technologies.

A significant $400 million (approximately £310 million) loan has been granted to General Compute, an AI inference cloud startup, by investment firm Upper90. This financing deal is notable as it is believed to be the first of its kind to use inference-specific chips as collateral. These chips are designed to run already-trained artificial intelligence models quickly and efficiently, contrasting with the more expensive chips primarily used for developing and training AI models.

This substantial investment signals a potential shift in the AI infrastructure market. As concerns grow over the high cost of advanced AI tools and tokens, investors are increasingly looking towards infrastructure that can run open-source models more affordably than the latest large language models (LLMs) from leading AI labs. General Compute, founded by CEO Finn Puklowski, previously secured a $15 million seed round in May to develop an inference 'neocloud' utilising silicon from SambaNova, an Intel-backed chipmaker. Neoclouds are specialised data centres built specifically for AI workloads, differing from the general-purpose infrastructure offered by major hyperscalers like AWS or Azure.

The SN50 chips, central to General Compute's offering, are engineered for inference tasks. They boast power efficiency and do not necessitate expensive water-cooling systems, enabling quicker deployment across a wider array of data centres. General Compute claims these new chips can deliver inference speeds up to 16 times faster than current GPU-based cloud solutions. Securing a large volume of these advanced chips presents a challenge for a new company, but Upper90's involvement provides a strategic solution.

Upper90's co-founder and CEO, Billy Libby, has a history in this niche financing sector. His firm previously financed GPU purchases for Crusoe, an energy-focused data centre startup, in 2021, a deal he believes was the initial loan backed by advanced chips. Traditional lenders were initially hesitant due to the perceived risks and uncertainties surrounding GPU depreciation. However, as companies like CoreWeave successfully established chip-backed loans as a viable business model, eventually leading to a major IPO, this form of financing has gained wider acceptance.

With GPUs now more widely understood and potentially over-purchased, Upper90 is now focusing on companies like General Compute to capitalise on the next phase of the AI boom. Libby stated that the firm believes open-source models will play a crucial role and that not every entity requires a supercomputer, but all will need inference capabilities. This thesis is supported by the increasing valuations of companies providing access to open models, and new chips from manufacturers like Groq and Cerebras attracting significant market interest, further indicating a diversifying landscape beyond Nvidia's ecosystem.

Why this matters: This deal indicates a maturing and diversifying AI market, moving beyond a sole focus on expensive training chips. It could lead to more affordable and accessible AI services for businesses and consumers in the UK.

What this means for you: What this means for you: This shift could lead to more efficient and potentially cheaper AI-powered services across various industries, from customer service bots to advanced analytics, making AI more integrated into daily life and business operations in the UK.

Related Articles

Get the news that matters.

Join thousands of readers getting the best of British news straight to their inbox.