Intel Announces New Blueprint For Agentic AI Tool With SambaNova Shortly After Terafab Deal

The design will combine GPUs for prefill, SambaNova reconfigurable dataflow units for high throughput decode, and Intel’s Xeon 6 processors.
In this photo illustration, a close-up of a printed circuit board (PCB) is seen, while the Intel logo is displayed on a screen in the background. (Photo by Nikos Pekiaridis/NurPhoto via Getty Images)
In this photo illustration, a close-up of a printed circuit board (PCB) is seen, while the Intel logo is displayed on a screen in the background. (Photo by Nikos Pekiaridis/NurPhoto via Getty Images)
Profile Image
Aashika Suresh·Stocktwits
Updated Apr 08, 2026   |   2:04 PM EDT
Share
·
Add us onAdd us on Google
  • The SN50 RDU will focus on high-throughput, low-latency inference, while Xeon 6 will provide memory bandwidth, connectivity, and integrated acceleration. 
  • It is expected to be launched in the second half of 2026. 
  • The announcement comes a day after Intel said that it would be partnering with SpaceX, xAI, and Tesla on silicon fabrication technology.

Intel Corp. (INTC) on Wednesday announced the next phase of its collaboration with SambaNova, revealing a blueprint for advanced agentic tools and systems.

The design will combine graphics processing units for prefill, SambaNova reconfigurable dataflow units for high throughput decode, and Intel’s Xeon 6 processors to address performance, efficiency, and software compatibility challenges. It is expected to be launched in the second half of 2026.

The announcement comes a day after Intel said that it would be partnering with SpaceX, xAI, and Tesla Inc. (TSLA) on silicon fabrication technology that would make it the central foundry for Elon Musk’s integrated AI, robotics, and space initiatives.

Shares of INTC were up nearly 11% at the time of writing.

Agentic AI Push

The design is geared toward a broader industry move that will match each stage of AI inference with the most suitable compute, while staying compatible with the x86 software stack used in modern data centers, Intel said.

The SN50 RDU will focus on high-throughput, low-latency inference, while Xeon 6 will provide memory bandwidth, connectivity, and integrated acceleration.

“The data center software ecosystem is built on x86, and it runs on Xeon— providing a mature, proven foundation that developers, enterprises, and cloud providers rely on at scale,” said Kevork Kechichian, Executive Vice President and General Manager of the Data Center Group (DCG) at Intel Corporation.

“Workloads of the future will require a heterogeneous mix of computing, and this collaboration with SambaNova delivers a cost-efficient, high-performance inference architecture designed to meet customer needs at scale—powered by Xeon 6,” Kechichian added.

Meanwhile, SambaNova said Intel Xeon 6 significantly improves compilation and vector database performance, speeding up coding workflows. It also serves as the system’s control plane, managing tasks, executing code, and validating outputs.

Project Terafab

Project Terafab, introduced by Elon Musk, is a joint initiative by Tesla and SpaceX to build an advanced semiconductor facility integrating design, manufacturing, packaging, and testing in one place. It aims to deliver over one terawatt of compute annually and rapidly iterate on new chips.

Under the new deal announced on Tuesday, Intel will contribute key technologies like PowerVia and advanced packaging, leveraging its in-house manufacturing model to enable secure, scalable AI chip production.

INTC shares have risen nearly 50% this year.

For updates and corrections, email newsroom[at]stocktwits[dot]com.

Follow on Google News
Read about our editorial guidelines and ethics policy