Models & Releases

SambaNova raises $676M at $5B valuation from SoftBank, Temasek

SoftBank-led round brings total funding to $1B for AI hardware startup

Deep Dive

SambaNova Systems, an AI startup offering integrated hardware and software for running AI applications, closed a $676 million funding round led by SoftBank Vision Fund 2. The round values the company at $5 billion and includes first-time investors Temasek and GIC, as well as funds managed by BlackRock, Intel Capital, GV, and Walden International. SambaNova has now raised a total of $1 billion since its founding in 2017.

Founded by Stanford University professors, including a MacArthur Genius Award winner, the 400-person company only began publicly detailing its product in December 2020. CEO Rodrigo Liang described the company's novel approach: providing core AI technology as a subscription service across industries, allowing businesses to process data without hiring as many data scientists. The hardware-software system is versatile, enabling use cases like detecting cancer cells in high-resolution medical images with greater accuracy and translating complex materials into multiple languages.

SambaNova's customers include the U.S. Energy Department's Argonne National Laboratory and Lawrence Livermore National Laboratory. The startup competes directly with Nvidia, which recently unveiled its first server microprocessor for complex computing work. SoftBank's Deep Nishar, joining SambaNova's board, noted the company now has commercial traction after being too early-stage for the Vision Fund three years ago.

Key Points
  • SoftBank Vision Fund 2 led a $676M Series D, joined by Temasek, GIC, BlackRock, Intel Capital, GV, and Walden International.
  • SambaNova's subscription-based AI hardware+software platform is used by U.S. national labs for cancer detection and language translation.
  • Total funding reaches $1B; company valued at $5B with 400 employees and competitors like Nvidia.

Why It Matters

SambaNova's subscription model lowers the barrier for enterprises to deploy high-end AI without hiring huge data science teams.