NewsAI

Mount Thor raises $45M to build Mac data centers for AI agents

Bay Area startup Mount Thor has secured $45 million in funding to deploy cloud infrastructure on Apple silicon, targeting the growing demand for macOS environments for AI agents and inference workloads.

3 min read

SCROLL

What happened

Mount Thor, a startup based in the Bay Area, has announced that it has raised $45 million across Seed and Series A funding rounds. The company specializes in building cloud infrastructure specifically for Apple silicon, aiming to provide data-center-scale access to macOS environments. According to 9to5Mac, the company’s core argument is that AI agents perform exceptionally well on Apple silicon, yet there is currently no provider offering this capability at a large scale.

The Series A round was co-led by CRV and Conviction, while the Seed round was co-led by Chad Byers and Offline Ventures. Other participants in the funding rounds include CSC Leasing, Enea Capital, Long Journey Ventures, and Upper90. The company also received investment from angel investors associated with OpenAI, Meta, Cognition, Together AI, and Vercel. Additionally, Mount Thor secured separate non-dilutive financing to acquire new hardware.

The details

Mount Thor operates by running macOS on Apple silicon as elastic compute instances, similar to how traditional cloud providers rent out Linux virtual machines. The service offers customers on-demand fleets of Macs for tasks that are exclusive to the macOS ecosystem. These use cases include training computer-use models in reinforcement learning environments, running secure macOS sandboxes for agents, and maintaining long-running instances that allow an agent to operate across multiple devices under a single Apple ID. The platform also supports build and test farms for Apple software development.

The company offers two primary models for accessing its infrastructure. In the managed compute model, customers select a specific machine class and image to receive either a full bare-metal Mac or an isolated macOS virtual machine with administrator access. Provisioning can be handled through a web portal, command-line tools, Kubernetes, or an API, with billing calculated per Mac on a 24-hour basis. In the customer-managed model, an organization runs its own Macs at a Mount Thor site using its own Mobile Device Management (MDM) system, while Mount Thor handles the facility operations, inventory tracking by serial number, and event notifications.

Background

The startup was founded this year and deployed its first data center in under six months. The team includes members with previous experience at Cerebras, CoreWeave, Fluidstack, Nebius, OpenAI, and xAI. Peter Voell, the Founder and CEO of Mount Thor, stated that making Apple’s operating system and silicon available at data-center scale represents a significant untapped opportunity in AI infrastructure. He noted that decades of investment in accessibility and human-computer interaction make macOS an ideal environment for agents to work and for humans to maintain accountability.

Voell also highlighted that Apple silicon is becoming a high-performance, power-efficient platform for production inference, particularly with the introduction of chips like the M5 Ultra. The M5 Ultra Mac Studio, which shipped in August, features up to 512GB of unified memory, allowing large open-weight models to run on a single machine with significantly lower power consumption than traditional GPU-based servers. Mount Thor plans to use part of the new funding to build what it describes as the best inference lab for open-source models on Apple silicon.

What it means

The concept of renting Macs in the cloud is not new; AWS has offered Mac instances for several years, and MacStadium has provided similar services for a longer period. However, Mount Thor differentiates itself by targeting a different customer base. While traditional providers focus on developers, Mount Thor argues that the primary customer is now the AI agent itself. The company posits that demand for Apple’s platform is scaling with the number of autonomous agents rather than the number of human users.

Brittany Walker, General Partner at CRV, described Mount Thor as sitting at the intersection of expanding autonomous agent capabilities and the increasing demand for compute resources that support both local and cloud workloads. The company stated that the new funding will be used to expand its compute footprint, build out its team, and rapidly add capacity to meet growing customer demand. This move suggests a shift in how cloud infrastructure is perceived, with Apple hardware becoming a critical component of the AI agent economy.