Modular

Modular was founded in 2022 by Chris Lattner and Tim Davis, who met while working together at Google. Lattner is one of the most influential figures in modern compiler and programming-language engineering: he created LLVM and the Clang compiler as a graduate student at the University of Illinois at Urbana-Champaign, later joined Apple where he designed and led development of the Swift programming language, then moved to Tesla (Autopilot) and Google, where as a Director of Engineering he led TensorFlow infrastructure and co-created MLIR (Multi-Level Intermediate Representation), before a brief stint as SiFive's Chief Platform Officer. Tim Davis spent years on Google Brain and Google's core ML systems teams (TensorFlow APIs, the XLA compiler, MLIR, and TensorFlow Lite for mobile/edge), giving the founding team deep, firsthand experience with the fragmentation and performance bottlenecks plaguing AI infrastructure at hyperscale. Modular was conceived (per its own materials) in Seattle before establishing its headquarters in Silicon Valley (Palo Alto), later adding offices in San Francisco, Los Altos, Boston, and Edinburgh.

Modular's core thesis is that AI's 'compute layer' is fractured across incompatible hardware and software silos (CUDA for NVIDIA, ROCm for AMD, various vendor-specific toolchains), forcing developers to rewrite performance-critical code for every accelerator. The company's answer is a full-stack, open, hardware-agnostic platform: Mojo, a new systems programming language that looks and feels like Python (and interoperates directly with the Python/CPython ecosystem) but compiles via LLVM/MLIR to near-C/CUDA-level performance and can target CPUs, GPUs, and other accelerators from a single codebase; and MAX, a GenAI-native compiler/runtime and model-serving framework that packages models (from PyTorch/TorchScript, ONNX, or native Mojo graphs) into an OpenAI-API-compatible inference server, with a library of 500+ pre-optimized open models (Llama, Mistral, Whisper, FLUX, and others). Mojo launched publicly via a hosted Playground on May 2, 2023, shipped a local SDK for Linux that September, and was progressively open-sourced — the standard library went Apache 2.0 in March 2024, and the full Mojo compiler was open-sourced in August 2026, one week after Mojo reached its 1.0 stable release (August 11, 2026; a 1.0 beta had shipped in May 2026 alongside the mojolang.org site). Modular also built 'Modular Cloud' (shared/dedicated hosted inference endpoints) and, per third-party analysis, a Kubernetes-native GPU cluster orchestrator ('Mammoth').

Modular raised venture financing in three primary rounds before being acquired. It closed a $30 million Seed round on June 30, 2022, led by GV (Google Ventures) with participation from Greylock, The Factory, and SV Angel. On August 24, 2023, it raised a $100 million Series B led by General Catalyst (with GV, SV Angel, Greylock, and Factory participating), reportedly valuing the company at roughly $600 million and bringing total funding to $130 million. On September 24-25, 2025, Modular announced a $250 million Series C led by Thomas Tull's US Innovative Technology Fund, joined by new investor DFJ Growth and existing investors GV, General Catalyst, and Greylock — valuing the company at approximately $1.6 billion post-money and bringing total funding to $380 million since inception. Greylock, led by partner Saam Motamedi, was Modular's first institutional investor and participated across all three rounds.

Business model

B2B AI infrastructure software; consumption/usage-based pricing for inference compute via the MAX platform and Modular Cloud (shared endpoints, dedicated endpoints, custom models), plus self-hosted/enterprise licensing of the open-source Mojo/MAX stack and revenue-sharing arrangements with cloud/hardware partners (e.g., AWS Marketplace).

Usage-based pricing for hosted inference (pay-by-usage cloud), enterprise/self-hosted deployment agreements, and marketplace/partner revenue share (e.g., AWS Marketplace listing for MAX); open-sources core Mojo/MAX technology while monetizing managed cloud, dedicated capacity, and enterprise support.