SambaNova may now be another Silicon Valley unicorn.

SambaNova Systems, whose founders are the crème-de-la-crème of Silicon Valley, raised $250 million in a Series C funding. The startup is at the cutting edge of AI and large-scale computing. It is dedicated to developing “the industry’s most advanced systems platform to run artificial intelligence (AI) and data-intensive applications.” (VentureBeat)

Funding round C

The funding round was led by BlackRock (NYSE: BLK). Existing investors, including GV, Intel Capital, Walden International, WRVI Capital, and Redline Capital, also participated.

“Raising $250M in this funding round with support from new and existing investors puts us in a unique category of capitalization,” said Rodrigo Liang, co-founder, and CEO, SambaNova Systems, in a statement. “This enables us to further extend our market leadership in enterprise computing.”

SambaNova has raised a total of $456 million to date, including the current round.

Highly esteemed founders

Rodrigo Liang, Kunle Olukotun and Chris Re. founded SambaNova in 2017.

Liang previously led an engineering group that designed chips for Oracle’s enterprise servers.

Kunle Olukotun and Chris Re. are Stanford University professors. Known as the “father of the multi-core processor,” Olukotun leads the Stanford Hydra Chip Multiprocessor research project.

Re. is a MacArthur Genius Award winner and an affiliate of the Stanford AI Lab.

AI is running into hardware limitations

As data increases in size and AI applications get more and more complex, currently available hardware is proving inadequate. Moreover, it is very expensive to create silicon wafer chips dedicated to AI applications. However, software-defined AI hardware is an emerging solution.

“SambaNova Systems offers an integrated hardware and software solution with an architecture optimized for dataflow from algorithms to silicon,” the company said. “SambaNova Systems’ reconfigurable dataflow architecture enables applications to drive optimized hardware configurations.”

Therefore, fixed hardware will no longer impose limitations on data and software.

Related Story:   Artificial Intelligence: New Magnetic Memory Device Could Punch Through The Physical Barriers To Faster AI