Simplismart raises $14m Series B led by Dallas Venture Capital

Simplismart raises $14m Series B led by Dallas Venture Capital

27 October 2026: Simplismart, which makes software for running AI models in production, says it has raised a $14 million Series B led by Dallas Venture Capital, with NVIDIA, Accel, Shastra VC and Micromax taking part. The company says its annualised revenue run rate has grown 14 times over the past six months.

Not yet independently verified. This rests on the chief executive's own letter. The revenue growth and customer cost figures are Simplismart's, the earlier $7 million Series A comes from Entrackr, and we have not confirmed the head-office city on the company's own pages. We will update this when it can be confirmed, and remove this note.

Amritanshu Jain, co-founder and chief executive, announced the round in a letter on the company's site. He started the company with co-founder Devansh four years ago, he writes, because they were tired of fighting production pipelines.

Software for the inference layer, not a GPU fleet

Simplismart's pitch is that it builds the software, not the GPU fleet. It sells to two kinds of customer. Application companies use it to run models within their own limits on cost, speed and where the workload runs. Cloud providers use it to turn GPU capacity into managed AI services sold under their own brands; at the Indian cloud company Yotta, Simplismart powers the serverless inference layer of Shakti Studio.

Customers named in the letter include HeyGen, Higgsfield, InVideo, Shiprocket, Apna and Ema, across video, voice, multimodal and enterprise AI. Simplismart's engineers work directly with NVIDIA's solutions architects, and NVIDIA is now also an investor.

Inference as a set of trade-offs

The letter argues that inference is "a spectrum, not a leaderboard". A voice agent that answers correctly but slowly loses the user; an overnight document job can wait if waiting is cheaper. Underneath every endpoint are choices about hardware, precision, batching, caching, kernels and scheduling, and improving one measure often worsens another.

Its example is Ema, whose target was about 200 milliseconds of latency at about nine cents per million tokens. One configuration came in at five cents but took 350 milliseconds; another reached 100 milliseconds at 12 cents. Simplismart says it delivered seven cents and 180 milliseconds, meeting both targets. For the Indian telephony company Exotel it reports managed inference at INR 0.90 to 1.30 per minute against INR 2.10 for a closed-model API set-up, with latency under 300 milliseconds instead of 1.1 seconds.

Where the money goes

Jain says the funding will go deeper into the inference engine and the orchestration platform and expand the cloud partnerships that bring the software to more customers, with North America as a priority. Entrackr, which reported the round the same day, says the company raised a $7 million Series A in October 2024.

For buyers, Simplismart competes less with model vendors than with in-house platform teams and with the managed endpoints of the big clouds. Our catalogue lists 7 products in machine learning software and 136 models on our AI model board.

Company profile on TrustList: Simplismart

Related on TrustList:

• RapidCanvas raises $29m to put field AI engineers in mid-size firms
• Higgsfield adds TCV and NVIDIA to its Series B for AI video
• Which funding round is your company ready for?

Sources

• Simplismart: NVIDIA backs our $14M Series B (letter by Amritanshu Jain), 27 October 2026
• Entrackr: Simplismart raises $14 Mn from Dallas Venture Capital, NVIDIA and others, 27 October 2026