Nvidia backs Indian startup helping companies run AI models with fewer GPUs

Simplismart's Software

Simplismart, a Bengaluru-based startup, has developed software designed to sit between AI models and the underlying computing infrastructure, optimizing how these models are deployed and managed. The core value proposition, as articulated by co-founder and CEO Amritanshu Jain, is that companies don't necessarily need the fastest or cheapest inference, but rather speeds and costs that are appropriate for their specific applications. Simplismart's platform automates complex adjustments, allowing customers to get more out of their existing GPU stack by fine-tuning factors like response time, request handling capacity, and power consumption.

The software offers flexibility, enabling companies to deploy and run AI models either through Simplismart's own platform or within their existing cloud and on-premises infrastructure. Beyond basic deployment, it provides tools for model fine-tuning, automatic scaling of computing resources based on demand, and performance monitoring. The company has demonstrated tangible results, with AI video-generation platform InVideo reportedly cutting its GPU serving costs by 56% using Simplismart's technology. Other notable clients include HeyGen, Higgsfield, enterprise AI startup Ema, and Indian cloud provider Yotta, which uses Simplismart to power its serverless AI inference service within Shakti Studio.

Nvidia's Investment

Nvidia's participation in Simplismart's $14 million Series B funding round, alongside Accel, Shastra VC, and Micromax, with Dallas Venture Capital leading, is a notable strategic move. This investment values the four-year-old startup at approximately $100 million. The backing from Nvidia presents an interesting dynamic, as Nvidia is the dominant supplier of AI chips, and Simplismart's software aims to help companies run AI models with fewer GPUs. This apparent paradox is addressed by Jain, who argues that by making AI inference more efficient, Simplismart can ultimately benefit Nvidia by making its hardware more attractive and encouraging wider adoption of AI applications.

Beyond the financial investment, Nvidia is actively collaborating with Simplismart on technical integrations, including around its Dynamo inference software. This collaboration involves engineers and solutions architects from both companies, and Nvidia has also facilitated introductions to potential customers and partners for Simplismart. This suggests a deeper strategic alignment aimed at enhancing the overall AI ecosystem, rather than simply a financial stake, by ensuring that Nvidia's powerful GPUs are utilized as effectively as possible, thereby expanding the market for AI solutions.

Market Position and Future Plans

Simplismart operates in an increasingly competitive market for AI inference software, facing rivals that include Nvidia's own offerings, major cloud providers like Amazon Web Services, Google Cloud, and Microsoft Azure, and a growing number of other startups. Jain highlights that Simplismart tends to win in scenarios where customers have stringent service-level agreements (SLAs) and tight total cost of ownership (TCO) constraints. Unlike competitors that often provide preset options for faster or cheaper inference, Simplismart's approach focuses on finding a tailored balance for each customer's unique needs.

The startup, founded in 2022 by former Oracle engineer Amritanshu Jain and former Google engineer Devansh Ghatak, has seen significant growth, with its annualized revenue run rate increasing 14-fold over the past six months to an eight-figure sum. More than half of this revenue originates from the U.S., despite a larger customer base in India. The new funding will be used to expand Simplismart's presence in the U.S., grow its 58-person team, and further invest in developing its inference software, with the U.S. remaining a key market for continued customer base expansion.