SambaNova Systems Offers A New AI System Claimed Three Times Faster
JAKARTA - SambaNova Systems, a Silicon Valley startup that builds artificial intelligence and has been supported by SoftBank's investment, announced on Wednesday, September 14 that it has started offering a new AI system called DataScale SN30. This chip is claimed to be three times the speed of the previous product.
The startup is one of several AI chip companies competing with Nvidia Corp, which currently dominates the world AI market. Part of SambaNova's strategy is to offer a robust chip with software platform that makes it easy for companies without AI expertise to start using AI in their business. SambaNova also offers subscriptions for AI hardware and software.
“For companies, your value is in your data. What we're trying to do is take this very complex aspect of the issue and we're bringing it down for you," said Rodrigo Liang, co-founder and CEO of SambaNova.
Gartner AI infrastructure analyst Chirag Neare said Nvidia was starting to face competition from startups like SambaNova as well as from AI offerings by Alphabet Inc's Google and Amazon.com Inc's cloud businesses.
SEE ALSO:
"If you're trying to increase the productivity of your data scientist, if you're trying to increase the throughput of your deep learning ecosystem, then ecosystems like SambaNova suddenly start to become more serious competitors," he said.
The subscription model has the potential to put a heavy financial burden on SambaNova as it will have a lot of hardware. But the company has raised more than $1 billion with its most recent valuation last year being $5 billion.
IDC's senior analyst Matt Eastwood for cloud and enterprise infrastructure said computer hardware companies were interested in the model to compete with the cloud.
“For a supplier like SambaNova, it just means that instead of having some flexibility in their quarter-to-quarter revenue stream as transactions come in, they can build a smoother, more predictable revenue stream that will appeal to investors,” he said.