AI Infrastructure Funding Becomes the New Center of Gravity
AI infrastructure funding refers to capital flowing into platforms that provide the computing, data, and orchestration layers needed to run AI systems, rather than the consumer or enterprise applications that sit on top. In the current cycle, this layer has become the center of gravity for startup funding rounds, reflecting a shift in enterprise AI investment priorities. Instead of chasing the next chatbot or niche app, investors are backing the rails that make those products possible: unified payments infrastructure with AI, real-time AI models for video and simulation, and vertically integrated inference stacks. These startups are building the core plumbing that others will depend on, giving them pricing power and long-term relevance. The result is larger, more concentrated funding rounds flowing into fewer, highly technical companies that aim to own critical parts of the AI backend.
Primer Shows Why Payments Infrastructure Needs an AI Core
Primer’s USD 100 million (approx. RM460 million) Series C round highlights how payments infrastructure AI has become a priority for enterprises. The company offers a unified payments infrastructure platform that sits across a merchant’s entire lifecycle, from checkout to payout, capturing over 400 data points per transaction and managing more than 95% of customer payment volume on average. That complete data layer underpins its AI agent, Primer Companion, which is evolving from an insights tool into one that can run experiments, optimise performance, and execute decisions autonomously within merchant-defined parameters. “In the next few years, every payment decision in a large business will be initiated, optimized or audited by AI,” said CEO Gabriel Le Roux. As merchants consolidate fragmented processors and fraud tools, unified infrastructure becomes the natural place to embed AI and drive reliable decision-making.
Decart and the Race for Real-Time AI and World Models
Decart’s USD 300 million (approx. RM1.38 billion) raise signals strong investor conviction that real-time AI models will underpin the next wave of products. The company runs a vertically integrated AI research lab focused on real-time video and world models, built around its core infrastructure product, DOS 2.0. DOS delivers over 1,600 tokens per second for agentic inference—around eight times the industry average—and streams full-HD video at up to 100 frames per second across major cloud hardware providers. On top of this stack, Decart maintains two model lines: Lucy, tuned for real-time immersive experiences in gaming, e-commerce, and advertising with sub-30-millisecond response times; and Oasis, aimed at physical AI, including robotics and autonomous vehicles, through physically accurate real-time simulation. This blend of infrastructure and models positions Decart as a platform rather than a single-application company.
Why Infrastructure Rounds Outsize Application-Layer AI Funding
Both Primer and Decart show why infrastructure-focused startups are securing larger funding rounds than many application-layer AI companies. These platforms sit at critical chokepoints: payments routing and decisioning for Primer, and low-latency inference plus real-time simulation for Decart. They demand heavy up-front investment in specialised AI stacks, spanning data pipelines, model training, and inference optimisation across different hardware. Investors see that once these systems are in place, they can support many use cases and customers, turning each company into a multiplier for downstream applications. Vertical integration—owning the infrastructure, models, and sometimes the agent layer—reduces dependency on third parties and supports better performance and margins. In a crowded market of AI apps, owning the infrastructure means owning the rails others must use, making these companies more attractive targets for large-scale enterprise AI investment.
Enterprise Demand Is Pulling Capital Down to the Backend Layer
The funding momentum around AI infrastructure reflects a clear enterprise trend: demand is strongest for reliable backend systems that make AI operational, measurable, and safe to automate. Payments teams want AI-powered decisioning grounded in complete, contextual data; engineering and product teams want real-time AI models that can handle video, simulation, and agentic workloads at scale. These priorities are pulling capital toward platforms that can standardise and automate core functions—billing, payments routing, inference, and world modelling—rather than one-off application experiences. As more businesses seek to embed AI across operations, they treat infrastructure choices as long-term strategic bets, not tactical tools. This dynamic is likely to keep AI infrastructure funding in the spotlight, with investors backing specialised stacks that integrate tightly from data to model to execution, and enterprises building on top of those foundations.