Mount Thor Secures $45 Million to Deploy Apple Silicon Data Centers for AI Agents
Mount Thor Secures $45 Million to Deploy Apple Silicon Data Centers for AI Agents
Bay Area startup Mount Thor has raised $45 million across Seed and Series A funding rounds to build cloud infrastructure dedicated to Apple silicon. The company aims to provide elastic macOS compute instances specifically tailored to meet the growing demands of autonomous AI agents.
Mount Thor, a newly founded Bay Area startup, has announced a $45 million capital raise across Seed and Series A rounds to build cloud infrastructure powered by Apple silicon. Co-led by prominent venture firms including CRV, Conviction, Chad Byers, and Offline Ventures, the funding will expand the company's compute footprint and team. The startup's core premise is that autonomous AI agents are increasingly functioning as macOS users, creating a massive untapped opportunity to scale Apple operating systems and hardware at data-center levels to support modern reinforcement learning and agentic workflows.
The Rise of Agentic Mac Infrastructure
Traditional cloud providers have long focused on Linux virtual machines for enterprise workloads, leaving Apple's ecosystem largely unoptimized for massive elastic cloud deployments. Mount Thor addresses this gap by running macOS on Apple silicon as scalable compute instances designed explicitly for autonomous software agents.
As artificial intelligence shifts toward complex, multi-step agentic behaviors, the demand for macOS environments is scaling with the volume of software agents rather than human developers. This infrastructure shift enables automated workloads that require native Apple operating system accessibility and human-computer interaction standards.
Elastic Compute Instances
Providing on-demand fleets of macOS VMs and bare-metal servers for specialized agentic workloads.
Agentic Workflows
Supporting autonomous agents that require native macOS sandboxes and cross-device Apple ID access.
Managed and Customer-Run Solutions
The startup offers two distinct operational models to accommodate varying enterprise requirements. Customers can choose managed compute options to provision full bare-metal Macs or isolated virtual machines via web portals, APIs, or Kubernetes toolchains.
Alternatively, organizations can utilize customer-managed compute by operating their own hardware at a Mount Thor facility while leveraging the startup for serial number inventory, physical facility operations, and event notifications.
Hardware Efficiency and Inference Labs
Apple silicon hardware, particularly configurations like the M5 Ultra Mac Studio, delivers high performance combined with significantly lower power draw compared to traditional GPU servers. With up to 512GB of unified memory, single machines can efficiently host large open-weight models.
A primary focus of the recent funding round is building out a specialized inference laboratory dedicated to optimizing open-source models on Apple hardware. This approach offers a power-efficient alternative for production inference and reinforcement learning environments.
Unified Memory Capacity
Leveraging up to 512GB of unified memory to support large open-weight models locally.
Power Efficiency
Achieving high-performance model inference at a fraction of the power consumption of standard GPU servers.
Investor Backing and Industry Context
The Series A round was co-led by CRV and Conviction, with the Seed round led by Chad Byers and Offline Ventures. Additional participation came from institutional investors and notable angel investors from OpenAI, Meta, and Vercel.
While AWS and MacStadium have historically offered cloud-based Mac instances, Mount Thor distinguishes itself by targeting AI agents as its primary customer base, positioning Apple silicon at the core of next-generation enterprise AI infrastructure.
Why It Matters
Mount Thor's successful funding round highlights a pragmatic evolution in AI infrastructure, moving beyond generic GPU clusters to specialized hardware architectures. By matching the operational demands of autonomous software agents with the robust security and accessibility of macOS, the startup bridges a critical gap in enterprise cloud computing. As autonomous agents become pervasive across user devices, providing scalable, power-efficient Apple silicon environments at data-center scale could redefine how modern artificial intelligence applications are built, tested, and deployed in production.
Mount Thor raises $45 million to build Mac data centers for AI agents (9to5mac.com) · OpenAI and Georgia county leaders hit with Mafia-style RICO racketeering lawsuit over $30 billion AI data center (tomshardware.com)
Reporting reviewed through November 12, 2026.
