The nature of computing is changing at an unprecedented speed, driving an explosion of new hardware architectures to support applications powered by artificial intelligence (AI). Both hardware and software development are at an inflection point, particularly in the infrastructure layers supporting machine learning and AI.
Today's platforms haven't yet scaled to meet the market where there is clear demand for a seamless software layer enabling a heterogeneous set of compute options. The software driving today's hardware is fragmented, difficult to use, and largely either built by research groups inside universities and large corporations, or custom-made and proprietary for the particular hardware stack being created.
We believe there's an opportunity for a more broadly-available software layer to emerge, which is why we are leading the first round of funding for Modular, investing alongside Greylock and Factory. Founded by top AI infrastructure visionaries with deep experience building for high-performance AI systems, Modular focuses on solutions that abstract hardware and mask software complexities. They've been around long enough to know that "fixing" AI as it stands would just serve as a patch; at this point, a modular rewrite of the underlying software layer is necessary.
Modular's co-founders are responsible for the development of the compilers, languages, and projects that underpin devices carried by billions of people. I've gotten to know Chris Lattner over the past few years. He's an expert in compiler and runtimes, founding and scaling systems like LLVM, Clang, and the Swift programming language. He led Apple's developer tools team, played a meaningful role in driving Google's Cloud TPUs to market, and helped accelerate SiFive's business. Co-founder Tim Davis is an experienced entrepreneur and product leader who founded TensorFlow Lite and scaled on-device ML to billions of devices, setting the vision, strategy, and roadmaps for Google product teams spanning small research groups to production ML systems across thousands of engineers and a large global compute footprint spanning edge and data centers.
Together, the two built MLIR, a compiler framework for ML and heterogeneous compute, an emerging standard for heterogeneous compilers. Chris and Tim have solid reputations and networks, and their product success will be driven by shared experience building usable ML developer tools and infrastructure.
The vision Chris, Tim and the Modular team have for enterprise-grade artificial intelligence is essentially to become a core technology that will fundamentally transform businesses — something we've seen in previous cycles for innovations like software, the internet, and mobile computing. The hardware and software ecosystems supporting AI have reached a familiar inflection point, where enthusiasm for new capabilities has spawned complexity and fragmentation, and is ripe for simplification. Modular's founders are headed down a familiar road: reimagining foundational infrastructure with a modular software layer to prioritize simplicity and usability.
Throughout tech history, modularity has always displaced monolithic architectures. Minicomputers supplanted mainframes, PCs supplanted minicomputers, mobile devices overtook PCs – and now those devices inferring data at the edge could become more critical to users than the cloud itself. Each of these previous technical shifts spurred profound accelerations in innovation through greater ease of use, lower cost and democratization of access. We think it's the right time for Modular, a multi-cloud, multi-framework, multi-hardware approach focused on usability. Modular accomplishes this without compromising performance or efficiency, to enable AI to flow seamlessly to production.
This approach will provide the heat for a new wave of innovation, growth and solutions to previously intractable issues. Modular is uniquely qualified to meet the moment. Chris and Tim have proven themselves to be prescient in predicting the winters and shaping the springs of AI for many years.
We're incredibly excited to partner with Chris, Tim, and the Modular team as they execute on a promising vision for AI's future.