As the surge in artificial intelligence (AI) technology places increasing demands on computing resources, Nvidia has invested in the Indian startup Simplismart through a $14 million funding round. This investment is aimed at supporting software that enables companies to operate AI models more efficiently, often requiring fewer GPUs. Nvidia participated in this all-primary Series B funding round alongside Accel, Shastra VC, and Micromax, with Dallas Venture Capital leading the investment. This deal values the four-year-old startup at approximately $100 million, inclusive of the new funding.
In the current landscape, companies are significantly investing in GPUs to support AI applications; however, the reliable and cost-effective operation of these models at scale continues to pose challenges. Different workloads necessitate varying trade-offs concerning speed, cost, and performance. The deployment and management of AI models to fulfill business needs often involve intricate adjustments. Simplismart claims that its software automates much of this process, optimizing how models operate on existing computing infrastructure, thereby alleviating the need for customers to manage these configurations independently.
“No one actually requires the fastest or the cheapest inference,” stated Amritanshu Jain, co-founder and CEO of Simplismart, in an interview. He emphasized that companies need to execute AI models at speeds and costs that align with their specific applications. Simplismart’s software functions as an intermediary between AI models and the computing infrastructure, adjusting deployment and management to satisfy customer requirements. The company asserts that it can enhance various factors, including response time, request handling capacity, and overall computing power consumption, enabling businesses to maximize their existing GPU resources.
Based in Bengaluru, Simplismart provides companies with the capability to deploy and manage AI models through its platform or within their existing cloud and on-premises infrastructure. The startup also offers tools for model fine-tuning, automatic scaling of computing resources in response to demand fluctuations, and performance monitoring. Furthermore, Simplismart enables cloud providers to develop and operate their own AI inference services on GPU infrastructure utilizing its software. The company claims to have assisted AI video-generation platform InVideo in reducing its GPU serving costs by 56% through its software. Additionally, Simplismart collaborates with AI video firms HeyGen and Higgsfield, as well as enterprise AI startup Ema, among others.
Moreover, Indian cloud provider Yotta employs Simplismart’s technology to enhance the serverless AI inference service within its Shakti Studio platform. Jain disclosed that Simplismart’s annualized revenue run rate has surged 14-fold over the past six months, reaching an eight-figure sum. This figure is derived from recent customer usage rather than contracted recurring revenue and excludes revenue passed through to infrastructure providers. However, he did not reveal the precise amount. Jain noted that over half of Simplismart’s revenue originates from the U.S., despite the startup having a larger customer base in India.
In addition to its investment, Nvidia has been collaborating with Simplismart on technical integrations, particularly concerning its Dynamo inference software. This partnership involves the engineers from Simplismart working alongside Nvidia’s solutions architects, with Nvidia also facilitating introductions to potential customers and partners, according to Jain.
Interestingly, this investment presents a seeming contradiction, as Nvidia, the leading supplier of AI chips, is supporting software that enables companies to manage certain workloads with fewer GPUs. Jain contended that enhancing the efficiency of AI inference could ultimately benefit Nvidia by making its hardware more appealing to customers and fostering broader adoption of AI applications.
Simplismart is catering to clients in an increasingly competitive market for AI inference software, contending with offerings from Nvidia itself, as well as major cloud providers and hyperscalers such as Amazon Web Services, Google Cloud, and Microsoft Azure. A growing number of startups are also developing tools aimed at making AI models faster and more cost-effective to operate, driven by the rising demand for AI applications and the necessity to optimize the use of expensive GPU resources.
“Our competitive edge lies in scenarios where clients must adhere to stringent service-level agreements (SLAs) and total cost of ownership (TCO) requirements,” Jain remarked, highlighting the importance of meeting demanding performance standards while maintaining tight cost constraints. Unlike competitors that offer predetermined options for faster or cheaper inference, Simplismart seeks to achieve a customized balance tailored to each client’s specific needs.
Founded in 2022 by former Oracle engineer Jain and former Google engineer Devansh Ghatak, Simplismart emerged from their experiences in building and deploying machine-learning systems. The two met while studying at BITS Pilani, and the startup currently employs 58 individuals, according to Jain. Simplismart intends to utilize the new funding to enhance its presence in the U.S., expand its team, and further invest in the development of its inference software. Jain noted that the U.S. remains a crucial market for the startup as it aims to build upon its existing customer base there.
The funding announcement follows a report by Indian English daily The Economic Times indicating that Simplismart was in discussions to raise between $40 million and $50 million at a valuation ranging from $300 million to $400 million. When asked about the report, Jain refrained from commenting, stating that the startup was not actively seeking additional funding. He added that the Series B round has been closed and will not be extended.
