What happened
Mount Thor, a Bay Area startup that runs macOS on Apple silicon as rentable cloud compute, has raised $45 million across Seed and Series A rounds, 9to5Mac reports. CRV and Conviction co-led the Series A; Chad Byers and Offline Ventures co-led the Seed; CSC Leasing, Enea Capital, Long Journey Ventures, and Upper90 participated, along with angels from OpenAI, Meta, Cognition, Together AI, and Vercel. The company also raised separate non-dilutive financing for new hardware.
Mount Thor announced $45 million in combined Seed and Series A funding. 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, alongside angel investors from OpenAI, Meta, Cognition, Together AI, and Vercel. The company additionally raised separate non-dilutive financing to purchase new hardware.
The product rents macOS on Apple silicon as elastic compute, similar to how cloud providers rent Linux VMs. Reported use cases include training computer-use models in environments, running secure macOS sandboxes for agents, long-running instances that let an agent work across a user's devices under one Apple ID, and build-and-test farms for Apple software.
9to5Mac describes two access models. Managed compute lets customers pick a machine class and image and receive either a bare-metal Mac or an isolated macOS VM with administrator access, provisioned via a web portal, CLI, Kubernetes, or API, and billed per Mac on a 24-hour basis. Customer-managed compute lets an organization run its own Macs at a Mount Thor site under its own MDM while Mount Thor operates the facility and handles inventory by serial number, problem reporting, and event notifications. Specific per-Mac pricing figures were not disclosed in the report.
Founder and CEO Peter Voell said the funding will fund an expanded compute footprint and team, and the company plans to add capacity quickly to keep up with customer demand. He also said a focus of the round is building what the company calls the best inference lab for open-source models on Apple silicon, citing the M5 Ultra Mac Studio's up to 512GB of unified memory as putting large open-weight models within reach of a single machine.
The team comes from Cerebras, CoreWeave, Fluidstack, Nebius, OpenAI, and xAI, per 9to5Mac. The company was founded this year and says it deployed its first data center in under six months.
Why it matters
The round is a bet that AI agents, not human developers, will drive the next wave of demand for Macs in the cloud. Mount Thor's core claim is that agents increasingly need to operate macOS itself—training computer-use models, running secure macOS sandboxes, and persisting across a user's devices under one Apple ID—and that this demand scales with the number of agents rather than the number of people. If that thesis holds, it creates a new category of AI infrastructure built on Apple's platform, which has historically been under-served by cloud providers. The company also argues that Apple silicon, particularly the M5 Ultra Mac Studio with up to 512GB of unified memory, can run large open-weight models at a fraction of the power draw of traditional GPU servers, positioning Macs as an inference platform and not just a build-and-test tool.
The investment thesis is distinctive: Mac-in-the-cloud is not new—AWS has rented Mac instances for years and MacStadium longer—but Mount Thor's pitch is that the larger customer is an AI agent that needs a Mac to do its job, so demand grows with agent counts rather than developer headcount.
The company frames macOS as a strong environment for agents specifically, with Voell arguing that decades of accessibility and human-computer interaction investment make it a good place for agents to work and for humans to hold them accountable.
The inference angle matters for cost and power: if Apple silicon can serve large open-weight models with substantially lower power draw than GPU servers, it could open a cheaper production-inference path, though Mount Thor's efficiency claims are the company's own and not independently verified.
The round signals investor appetite for niche AI infrastructure plays that sit outside the conventional GPU cloud, following other recent funding for agent-focused infrastructure startups.
Interactive Mechanism: How It Actually Works
Explore the underlying technology behind this development interactively.
crm_get_transaction(id='4092').An agent must create a draft calendar event for Tuesday at 2 p.m. Which evidence would establish the requested result?
What to watch next
Whether agent-driven Mac demand materializes at data-center scale, and whether Mount Thor can differentiate from AWS Mac instances and MacStadium, which already rent Macs. Watch the buildout of its promised inference lab for open-source models on Apple silicon, how quickly it adds capacity, and whether customers actually shift agent workloads to macOS in production.
Evidence of real agent-driven demand: whether customers use managed Mac fleets for computer-use training and persistent agent sessions at meaningful scale, versus traditional Apple build-and-test farms.
The promised inference lab for open-source models on Apple silicon, and any published benchmarks comparing Mac-based inference cost and power draw to GPU servers.
Competition with AWS Mac instances and MacStadium, including pricing and enterprise features such as MDM integration and provisioning tooling.
Capacity expansion pace and any new data center deployments, plus whether pricing per Mac is publicly disclosed.