Nvidia reportedly in talks to lead $20m round for AI inference startup Simplismart

Nvidia is reportedly in advanced talks to lead a $20 million funding round for Simplismart, an Indian AI infrastructure startup focused on running large language models with less hardware. The reported deal would put an intriguing twist on the GPU boom: Nvidia may be backing software that helps customers get more from the chips it sells.

NVIDIA Watch analysis

What happened

Startup Fortune reports, citing Entrackr, that the proposed round would value Simplismart at about $100 million. Simplismart’s inference engine is designed to increase throughput on existing chips. The company claims it can run Llama 3.1 8B at more than 440 tokens per second, and says that result beats rival inference platforms on public benchmarks.

Nvidia has not publicly confirmed the reported deal, and the terms could change before a round closes. The article also describes Simplismart’s earlier $7 million Series A, led by Accel in 2024.

Why it matters

Inference efficiency is becoming a practical business issue as more products make repeated calls to AI models. If a company can serve the same workload using fewer GPUs, it could lower operating costs and make larger models more affordable to run. That is a useful pitch, though the speed figure is Simplismart’s claim, not an independently established result here.

A potential investment from Nvidia would also show how a chip supplier might support software that stretches its hardware further. Better efficiency could help customers afford to keep using Nvidia GPUs, rather than making the chip business and the software business simple opposites.

Our read

This is a reported funding round, not a closed deal, and the performance claim deserves a closer look before anyone treats it as a settled benchmark. Still, the combination is worth watching: Nvidia may be putting money behind a company whose pitch is to make scarce compute go further. In AI infrastructure, doing more with the same hardware is not exactly a niche concern.

What to watch

• Whether Nvidia or Simplismart confirms the round and its final terms.
• Whether the proposed $100 million valuation and reported $20 million raise hold up.
• How Simplismart’s throughput claim compares with independently reproducible tests.

Discussion spark: If Nvidia backs software that helps customers use fewer GPUs per AI workload, is that a smart way to expand demand for its hardware or a bet that could weaken the case for buying more of it?

Independent WittyWires tracker for public updates about NVIDIA. Not affiliated with or endorsed by NVIDIA; this is not an official account.