Artificial intelligence infrastructure startup Simplismart has secured $14 million (approximately ₹133 crore) in a Series B funding round to scale its inference engine and orchestrate GPU workloads globally.
The investment round was led by Dallas Venture Capital, which contributed nearly $5 million, according to details shared by cofounder and chief executive officer Amritanshu Jain in an interview with Inc42. The round also drew strategic participation from global chip giant Nvidia, early-stage venture firm Accel, Shastra VC, Micromax, and select angel investors.
Tackling the GPU Inference Bottleneck
While substantial venture capital has traditionally targeted AI foundation models and training clusters, production deployments face intense cost pressures during inference—the operational phase where trained models process user prompts and generate live predictions. As enterprises scale real-time AI applications, running models across fragmented GPU infrastructure often introduces high latency and unsustainable cloud spending.
Founded in 2022 by Amritanshu Jain and Devansh Ghatak, Simplismart operates as an optimization and software middle layer between raw graphics processing units and AI models. The platform allows enterprise engineering teams to configure model execution based on specific parameters such as latency tolerance, throughput, and target compute cost.
"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," Jain noted. The fresh capital injection will accelerate product rollouts across cloud providers and sovereign AI initiatives while expanding operations directly into North American markets.
Deployment Models and Dynamic Scheduling
To address fluctuating hardware availability, Simplismart is engineering a dynamic GPU scheduling system designed to balance computational workloads across distributed geographical regions and operating time zones, according to cofounder and chief technology officer Devansh Ghatak.
Simplismart provides three commercial deployment structures to align with varying enterprise security and consumption needs:
| Deployment Model | Infrastructure Setup | Billing & Pricing Structure |
|---|---|---|
| Shared Endpoints | Multi-tenant managed cloud infrastructure | Pure consumption pricing billed per token with no minimum commitments |
| Dedicated Endpoints | Dedicated GPU hardware bundled with optimization layer | Usage charges linked directly to GPU operational runtime |
| Bring-Your-Own-Compute (BYOC) | Deployed inside customer's private cloud or on-premises servers | Software licensing based on orchestrated workload volume |
Expanding Sovereign AI and Cloud Partnerships
Simplismart plans to use the new funds to deepen integration with sovereign AI operators and regional cloud vendors seeking alternatives to dominant hyperscalers. By enabling businesses to deploy open-source models onto diverse hardware environments without complex custom rewrites, the startup aims to capture market share among mid-market and enterprise operators seeking lower compute overheads.
The participation of Nvidia indicates growing industry emphasis on middleware that maximizes hardware efficiency, ensuring that scarce GPU clusters deliver predictable throughput at commercial scale.
