Recently, FriendliAI, a startup focused on AI inference, announced a successful $20 million seed extension funding round. This funding will be primarily used to accelerate the development of FriendliAI's platform, especially its unique inference optimization technology, aimed at improving the efficiency of AI model deployment and reducing related costs.
FriendliAI's platform focuses on the inference stage of AI models, which is a crucial step from training to practical application. By optimizing the inference process, the platform can significantly reduce the computing resources required for AI model operations, thereby lowering corporate costs and energy consumption. This is highly attractive for companies looking to deploy AI solutions in an economically efficient manner, as it allows for more cost-effective operations without compromising performance.
Image source note: The image is AI-generated, and the image licensing service provider is Midjourney.
In the current cloud computing environment, computing resources are billed based on usage, so fast inference speed is crucial. FriendliAI aims to help companies perform AI model inference faster, more economically, and more conveniently. Companies do not need to spend large amounts of money and time establishing, optimizing, and maintaining complex AI inference infrastructure; they can simply use FriendliAI's optimized GPU platform to easily deploy and scale their AI inference business.
FriendliAI's technology has received widespread recognition, with its platform capable of saving up to 90% in GPU costs for users and leading the market in large model inference performance. Although FriendliAI has not disclosed specific revenue data, with the rapid popularization of generative AI applications, the company's usage and revenue have seen significant growth since 2025, with expected revenue reaching 6 to 7 times that of 2024 in 2025.
Currently, FriendliAI has approximately 25 to 30 large clients, with partners spanning various AI-driven enterprises from startups to large corporations. Recently, the company has partnered with LG Electronics and become the exclusive API provider for the EXAONE foundational model of the LGAI Research Center.
Key Points:
🌟 FriendliAI has secured $20 million in funding, aiming to accelerate the development of inference optimization technology.
⚙️ The platform significantly reduces the cost and energy consumption of enterprise AI model operations.
📈 Expected revenue in 2025 will be 6 to 7 times that of 2024, with a growing customer base.
