Spectral Compute, the company behind SCALE, a breakthrough software framework that enables CUDA (Compute Unified Device Architecture) applications to run seamlessly on any GPU, today announced a £4.5 million seed round led by Costanoa, with participation from Crucible and prominent angel investors. The funding will accelerate Spectral’s go-to-market expansion and deepen product development as it brings true hardware flexibility to the high-performance computing (HPC) and AI ecosystems.
The world of AI and HPC has been dominated by Nvidia’s software stack and CUDA programming model. This dominance has created what many call the “Nvidia Tax”, premium hardware costs and vendor lock-in that make entire supply chains vulnerable to disruption. Now, new tariffs and export restrictions have only heightened this exposure, leaving businesses across research, finance, and AI highly-dependent on GPUs with 18-month lead times.
Spectral’s flagship solution SCALE is a software platform that allows CUDA-based code, the foundation of most modern AI and HPC applications, to run natively on any AI accelerated chip, starting with AMD GPUs. This means organisations can use the same codebase across multiple hardware vendors and allows organisations to finally make vendor agnostic cluster procurement decisions. Spectral’s technology ensures customers remain future-proofed and can tap into whichever hardware offers the best performance, efficiency, or availability at any given time.
Unlike traditional translation or porting tools, which can take 18–24 months and introduce bugs or degraded performance, SCALE enables near-instant interoperability. Early customers in microprocessor design and high-performance motorsports are already running workloads using SCALE to diversify their hardware stacks.
Founded in 2018 by engineers Michael Søndergaard, Chris Kitching, Nicholas Tomlinson, and Francois Souchay, Spectral brings over 60 years of combined experience in HPC, GPU programming, and compiler design. The team’s roots in performance optimisation span industries from computational fluid dynamics and high-frequency trading to AI and digital broadcasting, all unified by a mission to make high-performance computing open, portable, and efficient.
Today, SCALE lets CUDA work seamlessly across the entire AMD chip ecosystem with additional chip architecture support to follow soon. Spectral has 19 full-time employees and continues to grow as it expands into AI, scientific research, and cloud computing sectors. The company has plans to grow its engineering team to meet demand.
