The potential deal could strengthen Intel’s foothold in the fast-growing AI hardware market
Intel is reportedly in preliminary discussions to acquire SambaNova Systems, a California-based AI chip designer, as part of its broader push to regain momentum in the rapidly expanding artificial intelligence hardware market, according to reports citing people familiar with the matter.
The talks come as Intel seeks to bolster its AI capabilities amid fierce competition from Nvidia and AMD, whose GPUs dominate the sector. Intel’s previous attempts to position its chips as lower-cost alternatives have seen limited success. SambaNova, founded in 2017, has hired bankers to explore acquisition interest, with sources suggesting that a potential deal could value the company below the $5 billion mark set during its 2021 funding round.
SambaNova develops custom AI processors and systems designed for high-performance machine learning and large-scale inference. It has drawn backing from major investors, including Intel Capital and Walden International, the latter led by Intel CEO Lip-Bu Tan, who also serves as SambaNova’s executive chairman. Other investors include SoftBank’s Vision Fund, which shares investment ties with Intel.
Renewed focus on AI data center chips
In parallel with the acquisition talks, Intel is preparing to re-enter the AI GPU market with its next-generation Crescent Island data center chip, expected in 2026. The chip, featuring 160GB of memory, is designed to handle the increasing demand for real-time AI inference.
Intel’s Chief Technology Officer Sachin Katti recently emphasized the need for “heterogeneous systems” tailored to evolving AI workloads. The potential SambaNova acquisition would further reinforce Intel’s long-term strategy to integrate advanced AI capabilities across its product portfolio and reclaim its competitive edge in the semiconductor industry.
See What’s Next in Tech With the Fast Forward Newsletter
Tweets From @varindiamag
Nothing to see here - yet
When they Tweet, their Tweets will show up here.



