RadixArk Emerges from Project SGLang with $400 Million Valuation Amid Inference Boom
RadixArk, a new player in the artificial intelligence landscape, has recently completed its transition from the open-source project SGLang, achieving an impressive valuation of $400 million. This startup was unveiled in August and is backed by notable investors, including Lip-Bu Tan, CEO of Intel.
Leading RadixArk is Ying Sheng, an accomplished former engineer at xAI, who announced her role as co-founder and CEO on LinkedIn last month. Sheng, previously a research scientist at Databricks, has been instrumental in SGLang’s development, which is utilized by companies like xAI and Cursor to expedite AI model training processes.
Key Highlights:
- Investment and Support: RadixArk has garnered angel investments and is poised for further growth in the rapidly evolving AI sector.
- Focus on Inference Optimization: The startup, like its predecessor SGLang, specializes in enhancing inference processing, enabling AI models to operate more efficiently on existing hardware. This optimization is crucial, as inference and model training significantly contribute to AI operational costs.
- Market Trends: The surge in demand for inference infrastructure is evident, with various startups securing substantial funding. For instance, Baseten recently raised $300 million at a $5 billion valuation, while Fireworks AI secured $250 million last October.
RadixArk is committed to advancing SGLang as a leading open-source AI model engine, while also developing Miles, a framework tailored for reinforcement learning, aimed at enhancing the intelligence of AI models over time. As the AI industry continues to thrive, the focus on efficient inference processing remains critical for businesses looking to optimize costs and performance.
