Nvidia backs Simplismart to cut GPU demand for AI inference

Nvidia has joined a $14 million Series B round for Bengaluru‑based Simplismart, valuing the four‑year‑old startup at roughly $100 million. The round was led by Dallas Venture Capital with participation from Accel, Shastra VC and Micromax.

As demand for AI compute outpaces GPU supply, enterprises face rising costs and operational complexity when serving models at scale. Simplismart addresses this by inserting a software layer between models and the underlying hardware, automatically tuning deployment parameters such as latency, throughput and power draw.

The platform can be run on a customer’s own cloud or on‑premises infrastructure, and it also enables cloud providers to build serverless inference services on top of their GPU clusters. Early adopters report significant savings; video‑generation service InVideo reduced its GPU serving spend by 56 percent, while HeyGen, Higgsfield and enterprise AI firm Ema also rely on the technology.

Revenue has surged 14‑fold in the last six months, with more than half of the run‑rate coming from U.S. customers despite a larger Indian client base. Nvidia’s involvement goes beyond capital – its solutions architects are collaborating with Simplismart engineers on integration with the Dynamo inference stack and are opening doors to new partners.

Why it matters for GPU infrastructure

Software that extracts more work from each GPU delays the need for fresh silicon purchases, improves return on existing clusters, and gives cloud operators a lever to offer cheaper inference tiers. For AiGpu customers, this trend underscores the value of pairing high‑density GPU capacity with intelligent orchestration layers.