Mount Thor, a startup based in the San Francisco Bay Area, has raised $45 million across Seed and Series A funding to build out cloud infrastructure powered by Apple silicon. The company intends to offer macOS instances—both virtual machines and bare-metal Macs—at data-center scale to support AI agents. It contends that while macOS is well-suited for agents, no provider currently leases it out at the scale that AI workloads demand.
What Mount Thor Offers
Its platform treats Macs as elastic compute resources, similar to how developers use Linux VMs now. Organizations can use the infrastructure for workloads that are macOS-specific: reinforcement learning environments, agent sandboxing, cross-device agent workflows tied to Apple IDs, and testing and building Apple software at scale. Renting works through two models: managed compute—where Mount Thor provisions Macs or macOS VMs via portal, API, Kubernetes, or CLI—and customer-managed compute, in which clients deploy their own Macs at Mount Thor facilities using their own MDM, while Mount Thor handles site operations and device reporting.
Rationale, Specs, and Team
Central to its case is that autonomous AI agents multiply the need for macOS compute—demand will grow with agent count, not merely human user count. With Apple’s newer chips like the M5 Ultra, Mount Thor argues that inference workloads are now feasible on single machines thanks to up to 512 GB of unified memory, all while offering better energy efficiency vs. GPU servers.
The company is newly founded—less than a year old—and already has its first data center up. The leadership team includes veterans from firms like Cerebras, CoreWeave, Fluidstack, Nebius, OpenAI, and xAI. To support growth, Mount Thor will use the new capital to increase compute capacity and scale the team rapidly.
Funding and Investors
The $45 million round was co-led by venture firms CRV and Conviction, with Chad Byers and Offline Ventures leading the Seed. Other contributors include CSC Leasing, Enea Capital, Long Journey Ventures, Upper90, and individual backers from OpenAI, Meta, Cognition, Together AI, and Vercel. In addition to equity funding, Mount Thor secured non-dilutive investment earmarked for acquiring hardware.
Space in the Market and What Sets It Apart
Macs in the cloud aren’t a new idea—AWS and MacStadium already lease out Mac hardware. What Mount Thor aims to do differently is focus primarily on AI agents as the end-users. While traditional cloud providers cater to developers and enterprises doing builds or testing, Mount Thor believes agents’ growth changes the calculus. It sees its niche in delivering macOS compute that scales with agent needs, not just development workloads.
The freshly raised funds will be directed toward growing physical infrastructure fast and bolstering the team to meet evolving demand.
Analytical Take: Mount Thor’s strategy targets a rising inflection point where AI agents—programs acting autonomously—demand native macOS environments at scale. By leaning into the efficiency gains of Apple silicon, and offering bare-metal Mac fleets and virtual macOS instances, it puts pressure on end-to-end GPU-driven cloud providers. The real test will be execution: rolling out data center locations, managing cooling and energy costs of Apple chips, and ensuring pricing competes with mainstream cloud. If successful, this could shift how organizations think about macOS—not merely as endpoint devices, but as first-class cloud infrastructure for AI workloads.
