Mount Thor, a San Francisco-based startup that rents out Apple hardware at data-center scale, has raised $45 million across Seed and Series A rounds to build Mac data centers for AI agents. CRV and Conviction co-led the Series A, while Chad Byers and Offline Ventures co-led the Seed. CSC Leasing, Enea Capital, Long Journey Ventures and Upper90 also participated, along with angel investors from OpenAI, Meta, Cognition, Together AI and Vercel. The company said it also raised separate non-dilutive financing to pay for new hardware.

The company's central bet is that AI agents increasingly need to use real computers, and many of the tasks they handle only run on macOS. Mount Thor argues that demand for Apple's platform now scales with the number of agents rather than the number of people. Founder and CEO Peter Voell called making Apple's operating system and silicon available at data-center scale "a massive untapped opportunity in AI infrastructure," pointing to decades of investment in macOS accessibility and human-computer interaction as a strong foundation for agents.S1S2

Mount Thor runs macOS on Apple silicon as elastic cloud instances, much as providers rent Linux virtual machines. Target uses include training computer-use models in reinforcement learning environments, running secure macOS sandboxes where agents can test new skills, keeping instances alive for weeks so an agent can work across a user's devices under one Apple ID, and building and testing software for Apple devices at scale. Customers can rent either a full bare-metal Mac or an isolated macOS virtual machine with administrator access, billed per Mac on a 24-hour basis and provisioned through a web portal, command-line tool, Kubernetes or API.S1S2

Organizations can also place their own Macs in a Mount Thor facility and manage them with their own mobile device management tools, while Mount Thor operates the site and provides inventory by serial number, problem reporting and event notifications. The company also aims to run AI models directly, and says a key goal for the new funding is building a top inference lab for open-source models on Apple silicon. The M5 Ultra Mac Studio, which shipped in August, offers up to 512GB of unified memory, enough to fit large open-weight models on one machine.S1

Mount Thor was founded in 2026 and deployed its first data center in less than six months. Its team previously built infrastructure at Cerebras, CoreWeave, Fluidstack, Nebius, OpenAI and xAI.S1