Modular, a startup specializing in AI infrastructure, announced today it has secured $100 million in a funding. This brings its total raised to date to $130 million. The round was led by General Catalyst, with participation from existing investors including GV, SV Angel, Greylock, and Factory.
The new capital will support Modular's mission to provide next-generation AI infrastructure and tools to empower developers. Founded in 2022 by former Google engineers Chris Lattner and Tim Davis, Modular aims to address the barriers that slow down AI development. "We’re attacking the complexity that slows AI development today by solving the fragmentation issues that plague the AI stack," said Lattner.
Modular's AI engine, which currently operates in a closed preview, promises to improve inferencing performance of AI models on CPUs and eventually GPUs. It claims to speed up models by up to 7.5 times compared to their performance on native frameworks. It is compatible with existing machine learning frameworks like Google’s TensorFlow and Meta’s PyTorch, offering flexibility and efficiency to its clients.
But it's not just about the engine. Modular has also stirred excitement within the developer community with its Mojo programming language. Touted as being 35,000 times faster than Python, Mojo aims to bring all the functionalities of Python and C under one roof. This has a dual benefit: it offers a common interface for AI hardware while also slashing performance bottlenecks that are typical with traditional programming languages.
In just four months since their product keynote, Modular has managed to galvanize a community of more than 120,000 developers. They've also hinted at some major updates to come. "The ability to download Mojo and a whole developer toolset locally to your desktop machine is the most requested feature from Mojo developers," said the company. Developers can expect a general release in early September, with a standard library becoming open source by year-end.
The round comes amid surging enterprise demand for AI tools, though barriers remain in deployment and infrastructure. Microsoft warned its AI server shortages could disrupt services, while NVIDIA's top AI chips are backordered until 2024.
Deep Nishar, Managing Director at General Catalyst, noted, "Our conviction in Modular stems from the team's impressive expertise, particularly the co-founders' pivotal role in shaping the compilation landscape over the past two decades and fueling this new wave of AI through their work on TensorFlow."
Despite Python’s current stronghold in the AI and data science sector—being used by 87% of data scientists as of 2020—Modular believes that Mojo's undeniable performance advantages will usher in a new era of AI development. In a world where compute demand is fast approaching sustainability limits, Modular’s promise of affordability and scalability has never seemed more pertinent.
The new funding brings more than just financial heft; it adds a potent mix of validation and resources for Modular to accelerate towards its ambitious vision. As AI models grow larger and more complex, there is pressure on companies to optimize performance and efficiency. By tackling fragmentation in the AI stack, Modular aims to drive the next wave of AI innovation for enterprises worldwide.