Simplismart raises $14 million Series B; its run-rate revenue grew 14x in six months

Simplismart, a maker of AI inference software founded four years ago, raised a $14 million Series B led by Dallas Venture Capital, with participation from NVIDIA, Accel, Shastra VC and Micromax, co-founder and Chief Executive Officer Amritanshu Jain said in a letter on the company's blog dated October 27.

What the company does

Inference is the stage at which a trained AI model answers requests. Simplismart said its software "helps application companies run AI within their cost, performance, and deployment requirements, and enables cloud providers to turn GPU capacity into managed AI services under their own brands."

The company said that over the past six months its annualized revenue run rate has grown 14x. It did not publish the revenue figure. It named HeyGen, Higgsfield, InVideo, Shiprocket, Apna and Ema as companies running production workloads on its platform across video, voice, multimodal and enterprise AI.

How it started

Jain wrote that he and co-founder Devansh "started Simplismart because we were tired of fighting production pipelines." The company wrote its Series A letter two years earlier, he said, and its argument then was that "one size does not fit all in AI inference." The company said it has spent four years making individual models run better.

How it grew

Jain gave two customer examples, both the company's own account. For Ema, the target was about 200 milliseconds of latency at about nine cents per million tokens. One configuration cost five cents but took 350 milliseconds, and another reached 100 milliseconds at 12 cents. Simplismart said it delivered a production configuration at seven cents and 180 milliseconds.

For Exotel, which post-trained an open-source model on its own data, the company said its managed inference brought costs to ₹0.90-₹1.30 per minute, compared with ₹2.10 for a closed-model API setup, "approximately 38-57% lower." It reported latency below 300 milliseconds, against 1.1 seconds for the closed-model setup.

On the cloud-provider side, Jain wrote that at Yotta, Simplismart "powers the serverless inference layer of Shakti Studio." He said the company chose to "build the software that turns third-party GPU fleets into production AI services, rather than build a GPU fleet ourselves."

What the founders say

"On-demand compute is, in my view, the biggest unsolved problem in inference," Jain wrote. He described the difficulty as having "the right model ready when demand arrives without paying for peak capacity around the clock."

On NVIDIA's role, he wrote that Simplismart's engineers work directly with NVIDIA Solutions Architects, "with technical contributions coordinated between the teams."

What's next

The company said the funding will let it "go deeper on the inference engine and orchestration platform, while expanding the cloud partnerships that bring our software to more customers." It said North America is a priority as it grows globally.

Jain also described a dynamic scheduling layer the company is building toward, one that matches demand with suitable capacity across regions and time zones while staying within customers' data-residency and compliance boundaries.