Gimlet Labs Raises $300 Million for Specialized AI Inference Cloud

Gimlet Labs has secured $300 million in Series B funding to develop a specialized cloud infrastructure dedicated to AI inference. The investment arrives as the industry shifts toward inference-dominant workloads that demand higher power efficiency and lower latency than traditional training-centric hardware can provide.

The funding round was led by Andreessen Horowitz, with participation from a broad coalition of investors including Sapphire Ventures, Menlo Ventures, 645 Ventures, Arm, Eclipse, Emergence, Factory, Hudson River Trading, M12, OnePrime Capital, Prosperity7, QuantumLight, Samsung Ventures, Tiger Global Management, Triatomic, Wing Ventures, and XTX Markets. This capital injection follows a rapid growth trajectory for the company, which previously raised $80 million in a Series A round in March 2026.

The Multisilicon Approach

At the core of Gimlet Labs' strategy is the creation of a "multisilicon cloud." While traditional AI infrastructure often repurposes hardware designed for model training to handle inference, Gimlet is building a system from the ground up specifically for the execution phase of AI.

The company utilizes a technique called "heterogeneous disaggregation," which allows the system to distribute different phases of the inference process across the most optimal hardware for each specific task. The infrastructure integrates a mix of GPUs, CPUs, dataflow architectures, and near-memory compute to maximize throughput per kilowatt of power consumed.

Addressing the Power Bottleneck

This shift toward specialized hardware addresses a critical physical constraint facing the AI industry: electricity. As frontier models grow in complexity and autonomous agents require massive context windows, the power requirements for running these models at scale have become a primary limiting factor for deployment.

By moving away from a one-size-fits-all hardware approach, Gimlet Labs aims to reduce the energy footprint of AI inference. The scale of the $300 million investment signals a growing market conviction that general-purpose compute is insufficient for the next generation of agentic AI, necessitating a move toward heterogeneous clouds that can optimize for both speed and energy consumption.

Future Outlook

With the new funding, Gimlet Labs is positioned to scale its infrastructure to meet the demands of frontier AI workloads. The industry will be watching to see if the company's disaggregated hardware approach can deliver significant performance gains over standard GPU clusters in real-world deployments. While the technical premise is established, the primary challenge remains the successful deployment of this complex hardware mix at a scale that can compete with established cloud providers.