Spectro Cloud raises $100 million to manage AI infrastructure

Spectro Cloud has raised $100 million in a Series D round to expand its Kubernetes-based infrastructure management software for production AI deployments. The company did not disclose a valuation, revenue, customer count or headcount, but said the new financing brings total funding to more than $260 million.

Growth Equity at Goldman Sachs Alternatives led the round. AMD Ventures, Ericsson, LG Technology Ventures and Maximus also participated. Spectro Cloud said the round was oversubscribed, a common financing label that does not by itself say much about pricing or investor demand without valuation terms.

Founded in 2019 by Chief Executive Tenry Fu and Chief Technology Officer Saad Malik, Spectro Cloud sells software for deploying, governing and operating infrastructure across public cloud, private data centers, bare-metal servers and edge sites. Its core product, Palette, gives IT teams a control plane for Kubernetes environments that may span several physical locations and operating models.

The company is now pitching that control plane more directly at AI infrastructure, where enterprises are trying to run GPU-heavy workloads across mixed environments. Spectro Cloud says customers in areas such as healthcare, defense, manufacturing and telecommunications face added complexity from security, governance and regulatory requirements. Those claims fit a broader market pattern: AI hardware spending has accelerated, while many organizations are still trying to standardize the software layer between chips and applications.

Gartner expects global AI spending to reach $2.59 trillion this year, up 47% from a year earlier, according to figures cited by Spectro Cloud. The company argues that much of that spending is going into hardware, leaving a gap in the software needed to manage GPUs, storage, networking and policy controls at production scale.

PaletteAI moves to the center of the pitch

Spectro Cloud launched PaletteAI in October as an extension of Palette for AI infrastructure use cases, including GPU management and distributed inference. The company says the product is meant to help enterprises use GPUs more effectively and control the cost of running models as deployments scale.

Fu said in the company’s announcement that customers are approaching the problem from different starting points, including legacy modernization, edge operations, Kubernetes scaling, AI factories and neocloud services. He said Spectro Cloud’s platform is intended to give those customers a consistent way to manage that complexity while maintaining control.

The Series D proceeds are earmarked for three areas. Spectro Cloud plans to add more PaletteAI functionality, expand geographically and into more markets, and build tighter integrations with hardware vendors, server makers and systems integrators. The company specifically pointed to neocloud and sovereign cloud providers, which sell regionally focused infrastructure for organizations that need to keep data and AI workloads local.

AMD’s role signals a multi-chip bet

AMD Ventures’ participation is relevant because Spectro Cloud is positioning itself for environments that will not run only on Nvidia GPUs. As inference becomes a larger share of enterprise AI demand, infrastructure teams are under pressure to support different accelerators and server configurations rather than standardizing on a single supplier.

Patrick Rundell of AMD Ventures said in the funding announcement that inference is becoming a major driver of infrastructure demand as AI moves into production, and that Spectro Cloud’s platform approach addresses challenges for enterprises deploying those workloads at scale.

The company did not say how much of the round came from strategic investors versus financial backers, nor did it provide adoption metrics for PaletteAI. Without those numbers, the financing mainly shows that investors are still willing to fund software layers around AI infrastructure, especially where Kubernetes, edge deployments and hardware diversity intersect.

This story draws on original reporting from SiliconANGLE.