Nvidia has invested in Simplismart, an Indian startup whose software helps companies run AI models more efficiently. In some cases, that means running them on fewer GPUs.
The investment is part of a $14 million funding round. It is a slightly ironic bet from a company that sells the very chips Simplismart helps customers use less of.
The details of the round
Dallas Venture Capital led the investment. Nvidia joined alongside Accel, Shastra VC, and Micromax.
The deal values the four-year-old startup at about $100 million, a figure that includes the new capital.
Simplismart, also known as Verute Technologies Pvt Ltd, was founded in 2022 in Bengaluru. It also has a presence in San Francisco.
Accel is a repeat backer. It led Simplismart’s $7 million Series A in October 2024.
What Simplismart actually does
Simplismart focuses on inference. Its software tunes how models run in production so they use compute more efficiently.
The company pitches its platform as hardware-agnostic. Customers can deploy models across different infrastructure setups, including lower-cost GPUs, rather than being locked into one configuration.
It covers several kinds of AI workloads. These include large language models, visual recognition, and speech processing. The platform handles orchestration, optimization, and deployment.
Simplismart says it has dozens of enterprise customers. Named clients include Tata 1mg, an online health platform, and InVideo, a video creation tool. Its revenue has been scaling from the low single-digit millions toward larger annual recurring revenue targets.
Why Nvidia is funding efficiency
The relationship did not start with this check. Simplismart was an early participant in the Nvidia Inception Program, the chipmaker’s support network for startups. The two companies have also collaborated on inference microservices.
India is an interesting backdrop for this. A Bengaluru-founded company selling compute efficiency to a global market fits neatly with the country’s long track record in enterprise software.
Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.