LONDON: Vaultz Capital plc, an Aquis Exchange-listed investment company, said it has invested 3 million pounds ($3.8 million) in a newly incorporated software company whose technology allows computer hardware to process artificial intelligence workloads beyond conventional memory limits.

The company funded the investment primarily by selling 47 bitcoin at an average price of 62,914 pounds, raising about 2.96 million pounds. Vaultz said it now holds 86 bitcoin worth approximately 5.41 million pounds as of Oct. 1, 2026, along with about 400,000 pounds in cash.

Vaultz subscribed for 449,101 new ordinary shares at 6.68 pounds per share, giving it roughly 2.3% of the investee company’s enlarged share capital. The deal values the startup at 90 million pounds before the investment and 95 million pounds after, and forms part of a 5 million pound funding round.

The investee company has developed a patented software layer that sits between AI and high-performance computing workloads and the graphics processing units, or GPUs, that run them. Every GPU has a hard limit on the size of a problem it can hold in its own memory; beyond that, the industry typically adds more GPUs connected through expensive specialist networking.

The startup’s technology addresses that constraint at the software level, allowing a single GPU — including older, lower-cost hardware — to process workloads that would otherwise require a large cluster. The company says the technology is hardware-agnostic and works within standard AI training and serving systems without modifying underlying hardware or AI models.

Independent benchmarking has shown increases in usable compute per GPU, reduced time and cost for model fine-tuning, faster response times for long-document inference, and lower energy consumption per unit of useful work, according to Vaultz. The technology has also been demonstrated on non-AI workloads, including running a full-resolution operational weather model on a single desk-side device.

“This technology is designed to overcome that limitation, allowing larger workloads to run on existing hardware and potentially reducing the number of chips, infrastructure and energy required,” Vaultz Chairman Charlie Wood said. “With independent testing already demonstrating the technology across real workloads and hardware, we see considerable commercial potential.”

Vaultz said the addressable market covers the global installed base of GPU compute across AI training and inference, cloud infrastructure, enterprise and sovereign computing, and scientific and financial modeling.

The identity of the investee company was not disclosed, and Vaultz said certain terms of the investment remain confidential for commercial reasons.

Directors Charlie Wood and Fungai Ndoro co-invested 140,000 pounds alongside the company on the same terms, a related-party transaction under Aquis Growth Market rules. Director Ian Burns, who is independent of the co-investment, considered it fair and reasonable to shareholders.

Vaultz described the investment as the first deployment of part of its treasury into a high-growth technology position that the board believes could be transformational for shareholder value. The company’s modified net asset value per share, including the investment at cost, is approximately 3.41 pence.