Modular has raised $250M in its third financing round to continue its mission to build AI’s unified compute layer – a hypervisor for AI. The round was led by Thomas Tull’s US Innovative Technology fund, with DFJ Growth joining and with participation from all existing investors including GV (Google Ventures), General Catalyst and Greylock Ventures. This brings its total capital raised to $380M across three rounds since its founding in 2022 and values Modular at $1.6 billion – almost tripling its valuation from its last raise. The investment reflects Modular’s incredible momentum and reinforces its position as the world’s only truly unified AI infrastructure platform to power the future of AI superintelligence.

“Strategic AI implementation is the most important competitive factor in today’s economy, and as the public and private sectors ramp up their efforts to remain competitive, the demand for compute power to handle these heavy workloads is greater than ever. Modular is foundational in this new era of diverse AI infrastructure, providing a unified AI compute layer that maximizes efficiency, resilience, and cost reduction - and their platform is already in high demand from enterprises, clouds and developers. We are proud to support the team and their vision for portable AI that will power both the U.S. and global economies.” 
‍
- Thomas Tull, Chairman of USIT

‍

"Modular is addressing the most urgent challenge in AI: unifying the compute layer by enabling diversified processing hardware and software to operate cohesively. Modular's platform is poised to become a defining pillar of AI systems, unlocking portability, performance, and efficiency that will accelerate the path to superintelligence."

- Sam Fort, DFJ Growth partner

‍

Building the future of AI infrastructure for everyone

The world's appetite for compute is insatiable. CPUs yield to GPUs and ASICs as AI transforms everything, while data centers rise at unprecedented pace to feed the demand. Superintelligence won't just live in server farms - it's coming to every device, every chip becoming an AI-enabled agent. Inference costs plummet as reasoning models drive explosive usage, yet training costs climb relentlessly higher. The paradox deepens: amid this computational renaissance, massive underutilization haunts our existing capacity, fragmented by every hardware vendor's insistence on proprietary software stacks. The imperative is elegant but unforgiving: chase every flop, and make every one count - because software, not silicon, will determine whether this revolution soars or stalls.

Modular has spent the last 3+ years building foundational infrastructure to solve this for the world. Modular reinvented the world's accelerated compute programming model from the ground up, and is rapidly scaling to meet the enormous demand they are seeing from advanced enterprises and hardware partners. Modular has grown to more than 130 people today with their main headquarters in San Francisco Bay Area, along with a global footprint in North America, United Kingdom and Europe.

Since launching their platform in 2023, Modular has redefined what is possible with heterogeneous programming and AI across the world's CPU and GPU silicon. Its platform is being downloaded 10K’s of times per month and is growing at 75% m/m, has earned 24K+ GitHub stars, powers trillions of tokens served daily in production, and has 100K’s of developers in their ecosystem across more than 100 countries. Modular has now released 600K+ lines of open-source code, has thousands of contributions from developers globally, achieved state-of-the-art performance across NVIDIA and AMD on a single, unified stack, and delivered up to 70% latency reduction and 80% cost reductions for their partners and customers. Modular is building the future of AI alongside an emerging alliance of co-architects - including enterprises like Inworld, SF Compute and more, research teams like Jane Street; cloud providers like Oracle, AWS, Lambda Labs and TensorWave; and hardware leaders like AMD and NVIDIA - each rallying and driving towards the vision of a simpler, more open and innovative AI hardware ecosystem.

The Modular Platform

The Modular Platform is the first enterprise-grade AI inference stack that abstracts away hardware complexity. By replacing vendor-specific runtimes like CUDA and ROCm with a unified low-level layer, it eliminates the fragmentation that holds back existing AI frameworks that were never developed for modern Generative AI inference. The Platform includes the following components, that span from cloud orchestration layer down to hardware programming model (more details):

• 🦣Mammoth: A Kubernetes-native control plane, router, and substrate specially-designed for large-scale distributed AI serving. It supports multi-model management, prefill-aware routing, disaggregated compute and cache, and other advanced, at-scale, AI optimizations.
• 🧑🏻‍🚀MAX: A high-performance GenAI serving framework that delivers state-of-the-art optimizations – like speculative decoding and operator-level fusions – out of the box. It exposes an OpenAI-compatible endpoint, runs both MAX-native and PyTorch models seamlessly across GPUs and CPUs, and offers deep customization at the model and kernel level for maximum performance and flexibility.
• 🔥Mojo: A kernel-focused systems programming language that enables high-performance GPU and CPU programming, blending Pythonic syntax with the performance of C/C++ and the safety of Rust. All the kernels in MAX are written with Mojo and it can be used to extend MAX Models with novel algorithms.

And with the latest release, 25.6, Modular delivers 20–50% performance gains over the latest vLLM and SGLang on next-generation hardware like NVIDIA’s B200 and AMD’s MI355. At the same time, Modular is expanding support to new silicon such as Apple GPUs and many other consumer grade GPUs, along with many upcoming ASIC’s.

Mission and the future

This next round of funding will enable Modular to aggressively scale the Modular Platform natively in the cloud, extend support across cloud and edge hardware platforms, and power the world’s most advanced AI workloads – delivering throughput, latency, cost, and accuracy gains unmatched by any other inference company. All of this is in service of empowering developers with a single, unified AI infrastructure they can build on and trust.

“When we founded Modular, we believed that the world needed a unified platform for AI, and today, that vision is more important than ever. This funding will enable us to realize that vision for developers, enterprises and hardware companies around the world.” said Chris Lattner, CEO of Modular.

Modular is hiring in North America and Europe across a wide range of roles. If you want to shape the future of AI infrastructure and collaborate with some of the brightest minds in the field, visit Modular’s career page. The future of AI infrastructure is bright! Download, contribute and help Modular power the future of AI in the open today.

‍