South Korean startup FriendliAI has raised $20 million in a seed extension round to support its efforts to accelerate AI inferencing.

FriendliAI offers an inference platform designed to accelerate AI model inference - the process where a trained machine learning model generates predictions and outputs from new input data. Faster inference reduces computational demands on GPUs and other resources, while enabling rapid issue detection and automated responses.

Capstone Partners led its funding round, joined by Sierra Ventures, Alumni Ventures, Korea Development Bank (KDB), and KB Securities.

The firm was founded in 2021 by CEO Byung-Gon Chun at Seoul National University (SNU). The company moved its headquarters to Redwood City, California, in 2023; the same year it released its core offering of Friendli Inference.

Behind the solution

The inference solution is driven by iteration batching, which groups multiple prediction requests together at speed to make processing more efficient.

Its computation cache, meanwhile, can store and reuse partial results from previous calculations, preventing need for the system to recompute with each request.

The platform also supports native quantization, which allows models to use lower-precision formats like 8-bit floating point (FP8), 8-bit integer (INT8), and adaptive weight quantization. Such formats reduce memory and computational demands while maintaining prediction accuracy.

Friendli Inference is also claimed to have access to over 400,000 open-source models from Hugging Face.

Customers using the platform to transition AI prototypes into production include telecom operators SK Telecom and KT, alongside South Korean AI startups Scatter Lab and Upstage.

“The AI inference market is exploding as more organizations move from AI experimentation to production deployment,” said Byung-Gon Chun, CEO and co-founder of FriendliAI. “This funding underscores the strength of our approach and positions us to capture the massive opportunity ahead.”

“FriendliAI represents exactly the kind of infrastructure innovation needed to support the next wave of AI adoption,” said Ben Yu, partner at Sierra Ventures. “Their continued growth demonstrates strong market demand for efficient, scalable AI inference solutions.”

Increasing inference interest

The interest in inference has risen with the demand for enterprise AI. Lumen last year expected the next wave of AI market pressure to come from the AI model inference phase, with demand led by financial services, health care, and retailers.

Lumen CTO Dave Ward told SDxCentral this year that the AI inference push will crossover into hybrid architectures with data held on-premises and using inference models associated with trained models in the cloud.

On the hardware side, the inference push saw UK startup Fractile benefit from a $15m funding round last year, with the company promising faster AI model inference on its chips.