The dominant chip supplier is funding a company whose selling point is that you may not need as many of its chips. Nvidia joined Simplismart’s $14M Series B, led by Dallas Venture Capital, alongside Accel, Shastra VC and Micromax. The round puts the four-year-old Bengaluru business at roughly $100M.
Simplismart sells a layer that sits between a model and the hardware underneath it. Rather than asking customers to tune deployments by hand, its software adjusts latency, throughput and resource consumption to match whatever an application actually requires.
Chief executive Amritanshu Jain argues the market has been chasing the wrong target. Very few companies need the fastest or cheapest inference, he said. They need the speed and cost that fit their specific workload.
Proof in a customer’s bills
The startup says it cut GPU serving costs by 56 percent for AI video platform InVideo. It also counts HeyGen, Higgsfield and enterprise AI firm Ema as customers, while Indian cloud provider Yotta runs its serverless inference service on the software.
Jain told TechCrunch that annualized revenue grew 14-fold in six months to an eight-figure sum, measured on usage rather than signed contracts. More than half comes from the United States, despite India holding the larger customer count.
Why Nvidia showed up
The apparent contradiction resolves if efficiency expands the market. Cheaper inference invites more applications, and more applications consume more silicon eventually.
Nvidia has also been working with Simplismart’s engineers on integrations around its Dynamo inference stack and introducing the startup to customers and partners.
The capital funds go-to-market expansion and deeper platform work. For cloud providers, the appeal is a shortcut to selling inference without building the serving stack from scratch.
