Nvidia backs Indian startup helping companies run AI models with fewer GPUs
TechCrunch Jagmeet Singh
Nvidia just backed Simplismart, an Indian startup that helps companies run AI models with fewer GPUs. That’s awkwardly clever: the chip giant is funding software that can use less of its own hardware.
Based on reporting by TechCrunch, Jagmeet Singh — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Nvidia has put money into Simplismart, an Indian startup that tries to make AI models cheaper and easier to run. The $14 million Series B was all primary funding, led by Dallas Venture Capital and joined by Accel, Shastra VC and Micromax. The round values the four-year-old company at about $100 million including the new cash.
The pitch is straightforward: most companies don’t actually need the fastest possible inference or the absolute cheapest one. They need something in the middle that fits their workload, budget and service commitments. Simplismart’s software sits between the model and the underlying infrastructure, then adjusts deployment and management so customers don’t have to hand-tune those settings themselves.
That matters because AI deployments get messy fast. Different applications want different trade-offs on speed, cost and throughput, and the same setup doesn’t work for everything. Simplismart says its platform can tune response times, request volume and compute use, and it can run on a company’s own cloud or on-premises infrastructure. It also offers tools for fine-tuning models, scaling resources as demand changes and monitoring performance. Cloud providers can use it too, including to build serverless AI inference services on GPU infrastructure.
Simplismart says it helped InVideo cut GPU serving costs by 56%. It also counts HeyGen, Higgsfield, Ema and Indian cloud provider Yotta among its customers. Jain said the company’s annualized revenue run rate has grown 14-fold in the past six months to an eight-figure sum, though that figure is based on recent usage rather than contracted recurring revenue and excludes money passed through to infrastructure providers. He also said more than half of revenue comes from the U.S., even though the startup has more customers in India.
The Nvidia tie-up goes beyond the cheque. Jain said the two companies have been working on technical integrations, including around Nvidia’s Dynamo inference software, and that Nvidia has helped introduce Simplismart to potential customers and partners. The irony is obvious enough: Nvidia is backing software that can help customers use fewer GPUs, but Jain argues that better inference efficiency should also make Nvidia’s hardware more attractive and push more AI adoption overall. Simplismart plans to use the funding to expand in the U.S., grow its team and keep building its inference software.
My take — AI-written commentary, not fact-checked reporting
This is the kind of deal that makes the AI market look less like a chip arms race and more like a bill-a-thon. Nvidia backing a tool that reduces GPU usage sounds perverse only until you remember the real enemy is waste, not thrift. Everyone else is still pretending brute force is a strategy.
Read more about this at: TechCrunch