By Yang Jie and Sherry Qin

Intel laid out a strategy to capitalize on the latest shifts in artificial intelligence on Tuesday, reiterating its focus on the full-stack ecosystem as it looks to pull off a pivot to AI.

Emphasizing value over single chips, Intel Chief Executive Lip-Bu Tan outlined four pillars to meet modern software and hardware demands: PCs, edge and agentic AI, foundational data centers, and intelligence centers.

Each of these ecosystems represents a generational opportunity, Tan said at the Computex tech conference in Taipei.

Among the examples the chip maker showcased was a partnership with Nvidia-backed AI search startup Perplexity to deploy a data-processing system to manage confidential corporate information across local devices and cloud networks.

At the event, the two companies demonstrated what they dubbed "hybrid agentic AI inference", a system that splits workloads between local devices and the cloud.

The collaboration aligns with Intel's broader strategy to push more advanced AI capabilities directly onto everyday hardware.

AI is changing how people interact with their devices, making local, on-device processing a critical pillar of Intel's future development, according to the Intel CEO.

San Francisco, Calif.-based Perplexity rolled out its AI operating system in February, which CEO Aravind Srinivas has said can juggle different AI models, tools and files simultaneously.

Using Intel's latest processors, Perplexity can run smaller AI models locally on a user's computer --a hybrid approach combining local hardware efficiency with cloud power.

That prevents the risk of data leaks when uploading sensitive documents to the cloud, Tan and Srinivas said, as the Intel-powered system automatically protects classified information.

"[The] computer decides what should leave the device and what shouldn't, and each of these things is done with local AI," said Srinivas.

The system can deploy additional AI agents to gather external research and collaborate with the local model, all without ever uploading private files to the cloud, he added.

"This is the architecture we both believed in," Tan said. "The future is more compute in the data center and more compute on the local machine."

Write to Yang Jie at jie.yang@wsj.com and Sherry Qin at sherry.qin@wsj.com

(END) Dow Jones Newswires

June 02, 2026 05:29 ET (09:29 GMT)

Copyright (c) 2026 Dow Jones & Company, Inc.