April 29, 2024

Run fast

Hi there,

Nvidia is snapping up AI startups.

The company has agreed to acquire Run:ai for a reported $700M. The startup had raised $118M in total funding.

Nvidia is also said to be acquiring Deci AI ($55M total funding), per The Information.

Both startups aim to optimize AI workloads, as can be seen in our MLOps market map.

They were also featured on our 2021 AI 100 list of the most promising AI startups.

For Nvidia, the acquisitions fit a pattern.

Run:ai and Deci previously partnered with the chipmaker.

For example, in 2022, Run:ai announced a collaboration with Nvidia to simplify ML deployment as highlighted by CB Insights business relationship insights.

run-ai-business-relationship-insights_04262024

Note: partnerships are a key predictor of M&A. See where the competition is heading with CB Insights business relationship data.

In February 2023, Nvida also acquired OmniML ($10M in funding) — which shrinks models so they can run on the edge, similar to Deci's approach.

Notably, Run:ai would mark Nvidia's second largest acquisition ($700M) since it acquired networking company Mellanox ($6.9B) in 2019. This further underscores the strategic importance Nvidia is putting on AI optimization.

The bottom line:

Nvidia is looking to make using its pricey chips more cost-effective and efficient.

Customers of both Run:ai and Deci we spoke with expressed goals around optimization:

"Our partnership with Deci was around optimizing a machine learning model that we had and making it much faster to run on specific hardware that we were interested in." — Senior Manager, Fortune 500 company
"The biggest strength of the Run:ai platform is its flexibility and customization to fit our needs. The algorithm behind the platform is fixed, but we can configure projects, users, hardware allocation, and preemption and bin packing mechanisms to ensure high utilization of the system." — C-level executive, Public university

Nvidia is also shoring up against smaller chipmakers, which are emerging to compete with Nvidia's expensive GPUs by focusing on building new architectures that could boost AI efficiency.

We mapped out companies developing novel processor approaches here.