Simplismart sells software that lets customers wring more work out of the GPUs they already own. On Tuesday it closed a $14 million all-primary Series B with Nvidia among the backers, an infusion that puts the four-year-old Bengaluru company at roughly $100 million including the new money.
Dallas Venture Capital led the check. Accel, Shastra VC and Micromax joined, as did the chip supplier whose business efficiency software supposedly undermines. Chief executive Amritanshu Jain sees no contradiction: tighter inference makes Nvidia’s hardware more appealing and lets more AI applications exist.
The product is a layer between a model and the machines running it. Rather than hand over presets, Simplismart retunes deployment to each customer’s balance of response time, request volume and power draw, inside its platform or on customer cloud and on-premises gear. Fine-tuning, autoscaling and monitoring come with it, and cloud providers can stand up their own inference services on the same stack.
Proof points are concrete. InVideo reports a 56 percent cut in GPU serving costs. HeyGen, Higgsfield and Ema are also customers, and Yotta, an Indian cloud provider, runs serverless inference inside Shakti Studio on Simplismart’s technology.
Revenue is climbing fast. The annualized run rate rose 14-fold in six months to an eight-figure sum, Jain told TechCrunch, tracked on recent consumption rather than signed contracts, and over half of it arrives from the U.S.
Nvidia’s involvement goes beyond the wire: the two work on integrations around Nvidia’s Dynamo inference software, and the chipmaker has opened doors to customers. Jain and Devansh Ghatak, veterans of Oracle and Google, started the firm in 2022 after meeting at BITS Pilani; it now has 58 employees.
