- Dallas Venture Capital leads the round, with participation from NVIDIA, Accel, Shastra VC and Micromax, as Simplismart expands its AI inference platform
- The company reports 14x growth in annualised revenue run rate over six months
Bengaluru and San Francisco, 27th October 2026: Simplismart, an AI inference infrastructure company, today announced that it has raised $14 million in Series B funding, led by Dallas Venture Capital, with participation from NVIDIA, Accel, Shastra VC and Micromax. The funding will be used to further develop Simplismart’s AI inference platform, helping enterprises run AI applications more efficiently and enabling cloud providers to turn their GPU infrastructure into managed AI services they can offer customers. It will also support the expansion of cloud-provider and sovereign AI partnerships, alongside growth across North America and other global markets.
Simplismart builds the software layer between AI models and GPU infrastructure. Its modular inference engine allows businesses to configure how AI models run based on their requirements for speed, cost and performance, while its orchestration platform manages deployment and scaling across cloud and on-premises infrastructure. Over the past six months, the company’s annualised revenue run rate has grown 14x, with companies including HeyGen, Higgsfield, InVideo, Shiprocket, Apna and Ema running production workloads across video, voice, multimodal and enterprise AI applications.
“The industry has spent years pushing the limits of what AI models can do. With frontier-class open models now readily available, access alone is no longer enough to set a business apart. The advantage comes from how you run them: what each interaction costs, how quickly the application responds, and whether it stays reliable as demand grows,” said Amritanshu Jain, co-founder and CEO of Simplismart. “We have spent four years building the software layer between the model and the GPU, helping customers optimise how their workloads are deployed and operated. This funding will allow us to take that platform to more enterprises and cloud providers globally.”
Customer communications platform Exotel is one example. After finding the economics of closed-model APIs unworkable for its voice agents, Exotel post-trained an open-source model on its own data and moved inference to Simplismart. By fractionalising GPUs, so that each workload uses only the GPU capacity it needs, Simplismart reduced Exotel’s cost per minute by up to 57% compared with closed-model APIs. It also cut the cost below that of self-hosting the same open model, and brought response times down from 1.1 seconds to under 300 milliseconds.
For its cloud-provider offering, Simplismart works with NVIDIA to make AI infrastructure easier to deploy and operate at scale. Unlike inference providers that buy or reserve GPU capacity and resell it, Simplismart does not own compute. Instead, it provides the software layer that turns third-party GPU fleets into production AI infrastructure, which makes it a data centre enabler. Cloud providers can use the platform to offer managed AI services under their own brands without building their own software stack from scratch. This offering is already being used by Yotta, where its platform underpins the serverless inference layer of Shakti Studio. This enables customers to run AI models without having to manage the underlying inference infrastructure themselves.
“Inferencing is emerging as one of the biggest game changers in AI economics,” said Dayakar Puskoor, Managing Partner at Dallas Venture Capital. “Simplismart is rapidly emerging as a global player helping enterprises run AI workloads more efficiently, while enabling cloud providers to turn GPU capacity into scalable services. We are backing this fast-growing company because of its proven technical prowess in addressing a critical need in the AI infrastructure ecosystem globally, as well as the strong execution capabilities of its founding team.”
Running AI models in production involves much more than choosing a model and a GPU. Decisions around inference frameworks, quantisation and caching directly affect how quickly an application responds and what it costs to run,” said Devansh Ghatak, Co-Founder and CTO, Simplismart. “Our modular inference engine allows us to tailor these configurations to the requirements of individual workloads, while our orchestration platform manages deployment across different infrastructure environments. The new funding will also support our work towards a dynamic GPU scheduling layer that can match AI workloads with available computing capacity across regions and time zones, helping cloud providers make better use of their GPU infrastructure.”
About Simplismart
Simplismart is a full-stack AI infrastructure platform for fine-tuning, deploying, optimizing, benchmarking, observing, and scaling AI models in production. Its proprietary inference engine and orchestration layer let enterprises and AI-native companies run the full post-training lifecycle through a single control plane across AWS, Azure, GCP, sovereign and neo clouds, and on-premises infrastructure, while continuously optimizing cost, latency, performance, and GPU utilization. Simplismart powers production workloads across generative AI, voice, multimodal systems, and agentic applications worldwide.
Read the full article here