Expanse has secured seed funding to develop software that predicts the computing resources required by AI workloads before they run, with the aim of reducing wasted capacity and helping organisations make better use of existing infrastructure.

AI infrastructure company Expanse has raised $5.3 million in a seed funding round led by Crane Venture Partners, with participation from PXN Ventures and angel investors including former DeepMind researchers and leaders from AI infrastructure teams.

The company, which was founded by four University of Edinburgh graduates, develops software designed to predict the GPU, CPU, memory and runtime requirements of AI workloads before execution begins.

Expanse said the funding will support the expansion of its engineering team, accelerate product development and help bring its platform to more organisations across AI infrastructure, quantitative finance, life sciences, research and high-performance computing.

The investment comes as organisations face growing pressure to make more efficient use of computing infrastructure. GPU capacity remains constrained, while the time required to procure new hardware and the power, cooling and planning requirements associated with data centre expansion can make increasing capacity difficult.

According to industry estimates cited by Expanse, approximately 30% of cloud expenditure is lost through over-allocation. The company also pointed to research from Microsoft Research indicating that GPU utilisation sits at around 50% across many internal deep learning workloads.

These inefficiencies stem partly from the difficulty of estimating resource requirements before a workload runs. Engineers must determine how much processing power and memory a task will need, as well as how long it is likely to take, without knowing its precise requirements in advance.

Expanse aims to address this problem by analysing workloads before execution and recommending an appropriate resource configuration.

Unlike conventional monitoring and observability platforms, which provide information about resource consumption after workloads have run, the company’s software makes predictions before computing resources are committed.

Expanse said its approach could help organisations reduce over-allocation, prevent failed jobs and increase utilisation, allowing them to complete more work without necessarily investing in additional hardware.

The platform is installed directly within customers’ cloud or on-premises infrastructure, with workload analysis taking place inside their own environments. According to Expanse, this means code and telemetry do not need to leave the customer’s infrastructure.

“Today, every AI workload begins with an educated guess,” said Ismaeel Bashir, co-founder and CEO of Expanse. “Engineers shouldn’t have to predict exactly how much compute their code will need before they press run. The machine should carry the uncertainty, not the person.”

Expanse was founded by Bashir alongside fellow University of Edinburgh graduates Nikodem Bieniek, Eren Mendi and Yafet Melake.

The four engineers previously built and operated large-scale computing infrastructure across quantitative finance and national supercomputing facilities.

During his time at the Edinburgh Parallel Computing Centre (EPCC), part of the University of Edinburgh, Bashir developed a multimodal high-performance computing resource prediction system which, according to Expanse, established a new benchmark for predicting resource requirements across HPC workloads.

The founders identified an opportunity to apply their experience to the wider AI infrastructure market, where resource allocation continues to rely heavily on manual estimates, even in sophisticated computing environments.

Expanse said it has already demonstrated the potential value of its technology in a production deployment, identifying nearly $8 million worth of idle computing capacity within a single month. The company did not disclose the customer involved.

Scott Sage, co-founder and partner at Crane Venture Partners, said the ability to improve infrastructure utilisation would become increasingly important as demand for AI computing grows.

“The future of AI won’t be defined only by who builds the biggest clusters, but by who uses them most intelligently,” said Sage. “Expanse gives organisations something they don’t have today: certainty. By predicting exactly what every workload needs before it runs, customers recover existing capacity, reduce waste and extend the life of some of the world’s most valuable infrastructure. We believe that’s a new and important category in AI infrastructure.”

The round also includes investment from PXN Ventures, the venture capital arm of PXN Group, which was formed through the combination of Praetura Ventures and Par Equity.

Andy Barrow, managing director of PXN Ventures, said the growing adoption of AI had intensified the focus on the infrastructure needed to support it.

“The widespread adoption of AI by businesses and everyday users has led to an unprecedented focus on the infrastructure required to service this technology and many of the challenges that engineers face. As Expanse has identified, one of those challenges is optimisation and helping businesses to make better use of their compute resource, which can lower running costs and increase scalability. PXN is proud to be backing Ismaeel and the team, who are tackling this head on.”

The new funding will allow Expanse to develop its predictive software further and extend its reach into sectors where computing requirements can be substantial and difficult to anticipate.

The company’s proposition centres on recovering capacity that organisations already possess, rather than relying exclusively on further investment in computing hardware to meet growing demand.