SambaNova unveils SN50 inference chip alongside fresh funding
SambaNova announced the launch of its SN50 inference chip, positioning the new processor as a purpose-built alternative to traditional GPU-based deployments for running large AI models. The company said the SN50 can deliver 5x faster inference performance at roughly 3x lower cost compared with GPUs, a claim aimed at organizations struggling with the economics of scaling generative AI workloads.
The product debut coincides with a major capital raise. SambaNova said it has secured a $350M+ Series E financing round led by Vista, adding to a growing wave of investment in AI infrastructure as enterprises seek more predictable performance and pricing for inference-heavy applications.
Intel partnership targets broader deployment
In parallel, SambaNova also announced a partnership with Intel. While specific commercial terms were not disclosed, the collaboration is expected to focus on accelerating enterprise adoption and improving integration across data center environments. The move signals an effort to expand beyond standalone hardware claims and into ecosystem-level distribution and deployment support.
SoftBank to deploy first in Japan AI data centers
SoftBank is set to become the first customer to deploy the SN50 in Japan-based AI data centers, according to the announcement. The initial rollout could serve as an early proof point for the chip’s performance and cost assertions in production settings, particularly as Japan ramps up domestic AI capacity.
With the SN50 launch, new funding, and the Intel partnership, SambaNova is betting that specialized inference silicon can compete on both speed and total cost of ownership—two metrics increasingly central to the next phase of enterprise AI adoption.
Related
