Infinity banks $15M for AI inference infrastructure at $100M valuation

Infinity, an AI infrastructure startup focused on inference, has raised $15 million in new capital at a reported $100 million valuation. The round brings in institutional investors Touring Capital and Principal VC, alongside individual backers who work as researchers at major AI labs including OpenAI and Anthropic.

While model training has attracted much of the fanfare in the last few years, inference is where AI systems actually serve users and generate revenue. Infinity operates in this less glamorous, but increasingly critical, layer of the stack. The company’s focus is the infrastructure that runs models in production—managing compute, latency, throughput and cost so that applications can scale to millions of requests without exploding cloud bills or degrading performance.

For founders building AI products, this layer matters because it often becomes the hidden bottleneck. As more companies embed large models into customer workflows, the cost and reliability of inference can determine whether a product is viable. A new generation of infrastructure companies like Infinity is targeting that pain point, aiming to make it easier for teams to deploy and run sophisticated models without owning all the low-level systems expertise themselves.

Infinity has not publicly detailed its full product lineup, but positioning itself as an "AI infrastructure" business focused on inference suggests it is addressing topics such as model serving, scaling, and resource utilization. This may include everything from optimized runtimes and routing across hardware types, to orchestration logic that chooses the right model or configuration per request. That kind of tooling is becoming essential as enterprises move from experimentation to real workloads with strict SLAs.

On the financing side, Touring Capital and Principal VC lead the institutional participation in this $15 million round. The company also attracted checks from researchers affiliated with OpenAI and Anthropic, two of the most visible players in frontier model development. For a young infrastructure company, that mix of backers signals strong technical credibility in the eyes of practitioners who work with state-of-the-art systems every day.

The reported $100 million valuation puts Infinity in the club of early AI infra companies that investors are pricing as potential category leaders. For comparison, many applied-AI application startups raising similar dollar amounts are still valued below nine figures. Infra plays can command richer multiples because their success is not tied to a single end-user use case; they can serve multiple verticals as long as they solve horizontal scaling and cost issues for model deployment.

The investor roster is also notable for founders elsewhere in the AI stack. Touring Capital and Principal VC are clearly signaling that inference is not a solved problem, and that there is room for new platforms even as hyperscalers offer their own managed services. The participation of researchers from OpenAI and Anthropic suggests that independent tooling around productionizing large models is still valued by people who sit inside the leading labs themselves.

For startup operators, this round highlights several trends. First, the center of gravity in AI funding is expanding from model training and data pipelines to the operational layer: serving, monitoring and optimizing live systems. Second, investors are increasingly looking for infrastructure that can work across multiple model providers, not just tied to a single ecosystem. Third, credibility with advanced model practitioners—whether as angels or advisors—is becoming a key differentiator when raising capital in crowded infra categories.

Looking ahead, the main question for Infinity will be how quickly it can turn technical promise into visible traction. With a $15 million war chest, the company has room to hire more engineers, build a go-to-market function and deepen integrations with model providers and cloud platforms. Near-term milestones will likely revolve around landing early lighthouse customers, publishing performance benchmarks, and clarifying where its platform sits relative to offerings from major clouds and model APIs.

As enterprises shift more experimental AI pilots into production environments, demand for reliable inference infrastructure is expected to climb. If Infinity can demonstrate that it meaningfully cuts costs or improves reliability versus DIY stacks and incumbent platforms, this raise could be a launchpad to a much larger presence in the AI operations ecosystem. For now, the company joins a select group of inference-focused startups that investors are valuing at nine figures before the category has fully matured.