IT Home On October 28, as the AI boom continues to occupy computing resources, Nvidia participated in the funding of Indian startup Simplismart for $14 million (Note from IT Home: approximately 93.573 million RMB at current exchange rate). This investment targets software that can improve the efficiency of AI models, and such software can help companies reduce GPU usage in certain situations.
This round of financing is the B-round, and it is entirely equity financing. In addition to NVIDIA, Accel, Shastra VC, and Micromax also participated in the investment, with Dallas Venture Capital leading the investment. After the transaction is completed, this startup, which was founded just four years ago, is valued at approximately 100 million US dollars (equivalent to about 668 million RMB at current exchange rates), including the funds from this round of financing.
Today, companies are investing a large amount of money in purchasing GPUs to support the operation of AI applications. However, achieving stability and cost-effectiveness when deploying these models on a large scale still presents significant challenges. Different workloads have varying requirements for running speed, cost, and performance. To meet business needs, companies often need to make complex parameter adjustments when deploying and operating AI models.
Simplismart states that its software can automatically handle a large amount of the work, adjusting its operation on existing computing infrastructure according to the actual needs of the model, without requiring customers to manage these complex configurations themselves.
Amritanshu Jain, co-founder and CEO of Simplismart (the person on the left in the image above), said in an interview: “In fact, no one really needs a solution that is the fastest to reason about or the lowest in cost.” He explained that what businesses really need is a solution suitable for their specific use cases, allowing AI models to operate within reasonable speed and cost limits.
Simplismart’s software sits between AI models and underlying computing infrastructure. By adjusting the deployment and management methods of the models, it meets the specific needs of customers. The company states that its software can optimize indicators such as model response speed, request processing volume, and computational resource consumption, helping enterprises make better use of existing GPU resources.
The startup headquartered in Bangalore, India, offers various ways for businesses to deploy and run AI models: they can operate through Simplismart’s own platform, or directly use the cloud infrastructure or on-premise deployment environments already available to customers.
In addition, the company also offers model fine-tuning tools that can automatically scale computing resources based on demand changes and monitor the performance of the model. Cloud service providers can also use Simplismart’s software to build and operate their own AI inference services on GPU infrastructure.
Simplismart states that its software has helped the AI video generation platform InVideo reduce GPU service costs by 56%. The company also collaborates with AI video companies such as HeyGen and Higgsfield, as well as the enterprise-level AI startup Ema. Additionally, the Indian cloud service provider Yotta uses Simplismart’s technology on its Shakti Studio platform to provide serverless AI inference services.
Jain told TechCrunch that in the past six months, Simplismart’s annual revenue operating rate has increased 14 times, reaching an octal figure. However, this metric is based on recent customer usage, not on contracted recurring revenue, and does not include income paid to infrastructure providers. The company did not disclose the specific amount.
Jain said that although Simplismart has more customers in India, more than half of its revenue comes from the US market.
While investing in Simplismart, NVIDIA is also working on technical integration cooperation with the company, including integration around NVIDIA’s Dynamo reasoning software. Jaan told TechCrunch that the collaboration involves Simplismart engineers and NVIDIA solution architects. NVIDIA also helps Simplismart connect with potential customers and partners.
It is worth noting that this investment seems to present a contradiction: as a leading supplier in the AI chip field, NVIDIA is supporting a software that can help companies reduce their GPU usage.
Regarding this, Jain believes that improving AI inference efficiency could ultimately benefit NVIDIA. The reason is that more efficient inference capabilities can enhance the appeal of its hardware to customers, and drive more AI applications to be implemented, thereby expanding overall market demand.
The AI reasoning software market in which Simplismart operates is becoming increasingly crowded. Its competitors include not only the solutions provided by NVIDIA itself, but also large cloud service providers such as Amazon Web Services, Google Cloud, and Microsoft Azure, as well as ultra-large-scale cloud computing companies.
As the demand for AI applications continues to grow, companies are also seeking ways to improve the utilization of expensive GPU resources. As a result, more and more startups are developing related tools to make AI models run faster and at lower costs.
Jaan said, “We often have an advantage when customers have stricter requirements for Service Level Agreements (SLA) and Total Cost of Ownership (TCO).” This means that customers expect the system to meet strict performance standards while also needing to control overall costs.
He pointed out that, unlike some competitors who offer pre-set high-speed reasoning or low-cost reasoning solutions, Simplismart aims to find a more suitable balance between speed, cost, and performance based on the specific needs of each customer.
Simplismart was established in 2022 by former Oracle engineers Jan and Devansh Ghatak (the person on the right side above in the text). They met during their studies at BITS Pilani in India, and later founded the company based on their experience in building and deploying machine learning systems.
Jain said that currently, Simplismart has 58 employees.
Regarding the use of the funds obtained from this round of financing, Simplismart plans to further expand the US market, increase its team size, and continue investing in the research and development of AI reasoning software. Jaan told TechCrunch that the US remains a key market for the company, and Simplismart hopes to further expand its business locally based on its existing customer base.
Before the funding news was released, Indian English media outlet The Economic Times reported that Simplismart was negotiating to raise between $40 million and $50 million (at current exchange rates, approximately 267 million to 334 million RMB), with a target valuation of between $300 million and $400 million (at current exchange rates, approximately 2.005 billion to 2.674 billion RMB).
Regarding this report, Jaan refused to comment, stating that the company is not actively seeking a new round of financing at present. He also emphasized that the current Series B financing has been completed, and no additional funding will be provided.