Nvidia backs Indian startup helping companies run AI models with fewer GPUs

As the artificial intelligence boom strains computing capacity, chip giant Nvidia is backing a startup designed to help companies run AI models with fewer graphics processing units. Nvidia participated in a fourteen million dollar funding round for Bengaluru-based Simplismart, valuing the four-year-old firm at about one hundred million dollars.

Dallas Venture Capital led the round, with additional backing from Accel, Shastra VC, and Micromax. Simplismart offers software that sits between AI models and underlying computing infrastructure, automatically adjusting deployments to balance speed, cost, and performance. According to co-founder and CEO Amritanshu Jain, companies rarely need the absolute fastest or cheapest inference setup, but rather a configuration tailored to specific application requirements.

The platform automates tasks like scaling computing resources and fine-tuning models, and it previously helped video-generation platform InVideo cut its GPU serving costs by fifty-six percent. This investment highlights an apparent paradox, as the dominant supplier of AI hardware is funding technology that can reduce chip demand. Jain argues that making inference more efficient ultimately benefits Nvidia by making its hardware more attractive and driving wider AI adoption.

Our read is that hardware giants are moving to secure software layers that prevent infrastructure bottlenecks from stalling customer spending.

Simplismart operates in a crowded market alongside hyperscalers like Amazon, Google, and Microsoft, while also collaborating with Nvidia on technical integrations involving Dynamo inference software.

Jain notes that more than half of the startup's revenue comes from the United States.

The company plans to use the new capital to expand its U.S. presence, grow its team, and further develop its inference tools.