Nvidia Invests $14M in Indian Startup Simplismart to Optimize AI Model Deployment

Image: TechCrunch · Source

Nvidia led a $14 million investment in Bengaluru-based Simplismart, whose software enables companies to run AI models more efficiently with tailored GPU usage, boosting performance while reducing costs.

As demand for AI applications soars, companies face increasing challenges in deploying AI models cost-effectively at scale. Nvidia, a leading supplier of AI chips, has taken a significant stake in Indian startup Simplismart, contributing to a $14 million Series B funding round valuing the four-year-old firm at approximately $100 million. The round also included investors Accel, Shastra VC, Micromax, and Dallas Venture Capital.

Simplismart specializes in software that automates and optimizes how AI models run on existing GPU infrastructure. Unlike generic solutions focused solely on speed or cost, Simplismart’s platform customizes deployment configurations based on each customer’s specific performance and cost requirements. Its software manages trade-offs between inference speed, request volume, and GPU consumption, enabling organizations to use their current GPU resources more efficiently.

The Bengaluru-based company offers its platform as a standalone service or integrates with customers' cloud and on-premises setups. It also provides tools for model fine-tuning, automated resource scaling, and real-time performance monitoring. Additionally, Simplismart licenses technology to cloud providers, such as India’s Yotta, to support serverless AI inference services.

Among Simplismart’s clients are AI video-generation firms InVideo, HeyGen, and Higgsfield, as well as enterprise AI startup Ema. InVideo reported a 56% reduction in GPU serving expenses after adopting Simplismart’s software. While the startup has a greater number of customers in India, over half of its revenue comes from the United States, reflecting its growing international footprint.

Co-founder and CEO Amritanshu Jain emphasized that organizations do not necessarily need the fastest or cheapest inference but require solutions tailored to their unique application contexts. The startup’s annualized revenue run rate has increased 14-fold in six months to an eight-figure sum, based on recent usage rather than contracts, underscoring strong market traction.

Nvidia’s investment comes alongside ongoing technical collaboration, including integration work with Nvidia’s Dynamo inference software. Jain noted that although Simplismart’s software can reduce the number of GPUs needed for some workloads, this efficiency helps broaden adoption of AI applications and ultimately benefits Nvidia’s hardware business.

Simplismart operates in a competitive landscape, facing off against inference tools from Nvidia itself, major cloud providers such as AWS, Google Cloud, and Microsoft Azure, as well as emerging startups. The company differentiates itself by offering finely tuned trade-offs to meet stringent service-level and cost requirements, rather than providing preset speed or cost modes.

Founded in 2022 by former Oracle and Google engineers Amritanshu Jain and Devansh Ghatak, Simplismart now employs 58 people. The new capital will support expanding its U.S. presence, growing the team, and further developing its inference software platform. Despite reports of ongoing fundraising, the company stated its Series B round is closed and is not presently seeking additional investment.