AI companies are investing heavily in GPUs, but buying computing power is only part of the challenge. They also need to run models at a mix of speed and cost that suits their applications. Indian startup Simplismart says its software helps businesses make those trade-offs and use their existing infrastructure more effectively. Nvidia has backed that approach with an investment in the startup’s latest funding round.
A $14 million bet on inference software
Nvidia participated in Simplismart’s $14 million Series B alongside Accel, Shastra VC, and Micromax. Dallas Venture Capital led the round. The all-primary financing values the four-year-old startup at about $100 million, including the new funding.
Simplismart’s platform sits between AI models and the computing systems that run them. It adjusts deployment and operation to suit what a customer needs, including how quickly a model responds, how many requests it can handle, and how much computing power it uses. The aim is to automate choices that can otherwise require companies to configure and manage model operations themselves.
Co-founder and CEO Amritanshu Jain told TechCrunch that businesses do not simply need the fastest or cheapest inference. They need performance and cost levels that fit the specific application. That focus on balancing requirements is how Simplismart describes its role in a market where workloads call for different trade-offs.
Tools for cloud and on-premises systems
The Bengaluru-based company lets customers deploy and run AI models through its own platform or within their existing cloud and on-premises infrastructure. Its tools also support model fine-tuning, automatic adjustments to computing resources as demand changes, and performance monitoring.
Simplismart also works with cloud providers that want to build and operate AI inference services on GPU infrastructure. Indian cloud provider Yotta uses its technology to power the serverless AI inference service in its Shakti Studio platform.
The startup says its software helped AI video-generation platform InVideo reduce GPU serving costs by 56%. It also works with AI video companies HeyGen and Higgsfield, as well as enterprise AI startup Ema, among others. These examples illustrate the kinds of customers that may need to manage demanding workloads and computing expenses at the same time.
Growth, revenue and Nvidia’s role
Jain said Simplismart’s annualized revenue run rate grew 14-fold over the past six months to an eight-figure sum. The company said that figure reflects recent customer usage, rather than contracted recurring revenue, and excludes revenue passed through to infrastructure providers. Jain did not disclose the exact amount.
More than half of the company’s revenue comes from the U.S., even though it has more customers in India, Jain said. The new funding is intended to help Simplismart expand in the U.S., grow its team, and continue developing its inference software. Founded in 2022 by Jain, a former Oracle engineer, and Devansh Ghatak, a former Google engineer, the startup now employs 58 people, Jain said.
Nvidia’s involvement extends beyond financing. Jain told TechCrunch that the companies have worked on technical integrations, including around Nvidia’s Dynamo inference software. Simplismart engineers and Nvidia solutions architects are involved, and Nvidia has also helped introduce the startup to potential customers and partners.
Efficiency in a crowded field
Nvidia’s investment may seem surprising because the chip supplier is backing software that can help some customers run workloads with fewer GPUs. Jain’s argument is that more efficient inference can make Nvidia hardware more attractive and encourage broader use of AI applications. That is the company’s rationale for why reducing the computing required for some tasks can still support demand for AI systems.
Simplismart competes with inference offerings from Nvidia, cloud providers and hyperscalers such as Amazon Web Services, Google Cloud, and Microsoft Azure. Startups are also building tools to make models faster and less expensive to run as demand for AI applications grows and companies look for better use of costly GPU resources.
Jain said Simplismart tends to win customers with strict service-level agreements and total cost of ownership requirements. Rather than offering preset choices that prioritize speed or low cost, the company says it seeks a balance tailored to each customer.
The announcement follows a report by Indian English daily The Economic Times that Simplismart was in talks to raise $40 million to $50 million at a valuation of $300 million to $400 million. Jain declined to comment on that report, saying the company was not actively fundraising. He said the Series B was closed and would not be extended.