AI Inference Optimization Startup Secures $14M, Backed by Nvidia

An Indian startup specializing in optimizing AI inference has closed a $14 million Series B funding round. The investment was led by Dallas Venture Capital, with significant participation from industry giant Nvidia, alongside Accel and Shastra VC. The round consisted entirely of primary shares, valuing the company at approximately $100 million post-money.

From Tech Giants to a New Venture

Founded in 2022 by former Oracle engineer Amritanshu Jain and former Google engineer Devansh Ghatak, Simplismart operates with a team of 58. The founders identified a critical market need: as AI models grow in size and complexity, the cost and efficiency of running them in production environments have become major hurdles for businesses.

The company's software acts as an intelligent layer between AI models and the underlying computational hardware. Its primary focus is optimizing the inference phase—when trained models perform tasks like generating text or analyzing images.

The Core Technology: Balancing Speed, Throughput, and Cost

The platform's key capability is dynamically adjusting an AI model's response latency, request throughput, and corresponding computational resource consumption. This allows enterprise clients to tailor their inference services based on business priorities, whether that's maximum speed or cost control.

  • Flexible Deployment: Can be deployed on its own platform or within clients' existing cloud and on-premises environments, easing integration.
  • Model Fine-Tuning: Provides tools to optimize models for specific inference tasks.
  • Automated Operations: Features auto-scaling and performance monitoring, dynamically adjusting resources based on load to ensure stability and prevent waste.

Market Validation: Tangible Cost Savings Spearhead Adoption

The startup points to concrete results to demonstrate its value. It claims its platform helped AI video creation platform InVideo reduce its GPU service costs by a significant 56%. This compelling case study has attracted attention from other AI application developers.

Its current client roster includes AI companies like HeyGen, Higgsfield, and Ema. Notably, Indian cloud provider Yotta leverages the technology to power the serverless AI inference services on its Shakti Cloud platform, underscoring the technology's robustness.

Rapid Growth Fueled by U.S. Market Demand

Co-founder Amritanshu Jain revealed that the company's annual recurring revenue has skyrocketed 14-fold over the past six months, now reaching "eight-figure" dollar amounts. More than half of this revenue originates from the United States, indicating successful penetration into the world's most competitive tech market and validation from global clients.

Nvidia's participation in this round is a strong industry signal. It shows that while advancing computational hardware (like GPUs), industry leaders are also prioritizing the efficiency of existing compute power. The "AI inference optimization" space that Simplismart represents is becoming an essential layer in the AI infrastructure stack.