For Intel, from 🤫 hussh
American-made silicon for the sovereign AI factory.
Intel is the only American company building leading-edge logic at home. Its Core Ultra NPUs already run private, air-gapped AI agents on a person's own machine with OpenVINO, and 18A Xeon is made in the United States. 🤫 Private Agent One belongs on both: on every AI PC, and in a grid of garage and warehouse AI factories built on American silicon. We'd love to build it with you.
What Intel and 🤫 make possible.
Core Ultra NPUs already run private, air-gapped AI agents on the person's own machine.
🤫 Agent One is that on-device agent: the same stack, for every person, consent-first.
A flagship AI PC agent that shows why on-device AI sells Core Ultra.
18A is "the most advanced semiconductor process developed and manufactured in the United States," and Foundry is built around committed demand.
Our American garage and warehouse AI-factory grid is a distributed buyer of US-made silicon.
Real demand for US-made 18A Xeon, node by node, with a customer behind every one.
You are the American, sovereign silicon story, with the US government as a shareholder.
We serve government, military and veteran customers with a consent-first, sovereign posture.
An American-made, sovereign AI factory, for the people who need it most.
x86, on-device NPUs and American manufacturing.
Intel is in the middle of a turnaround under Lip-Bu Tan. Intel Products, client and data center, funds the build-out of Intel Foundry, which is targeting break-even around 2027.
Roughly flat year over year, with Products funding the Foundry build-out.
Data Center and AI grew 22% year over year, the brightest spot in the business.
"The most advanced semiconductor process developed and manufactured in the United States," live in Panther Lake and Clearwater Forest Xeon.
A passive stake of about 9.9%. American sovereignty, built into the company.
NVIDIA agreed to invest $5B and develop x86 CPUs and RTX-chiplet SoCs with Intel.
Intel's open toolkit for inference on CPU, GPU and NPU. The software that makes on-device agents practical.
The AI PC, the open edge toolkit and American-made silicon. Exactly what a private agent and a sovereign grid need.
The x86 base
Still the default across enterprise PCs and servers. Client and Data Center generated about $49B in 2025.
Core Ultra NPUs
Integrated NPUs for on-device AI. Panther Lake, Core Ultra Series 3, is the first client platform on US-made 18A.
A US foundry
The only US-headquartered company building leading-edge logic at scale at home. 18A is made in America.
US government backing
A government stake of about 9.9% and a national industrial-policy mandate.
OpenVINO
The open-source toolkit for optimizing and running inference at the edge, across CPU, GPU and NPU.
The NVIDIA alliance
A $5B NVIDIA investment and x86 plus RTX-chiplet products developed together.
Where Intel is going.
Every one of these priorities has room for a private agent the person owns.
Process leadership, 18A to 14A
Lip-Bu Tan: Intel is "going big time into 14A," with every fab investment tied to real customers.
Foundry to break-even
Driving Foundry toward break-even around 2027 by landing committed external customers.
The AI PC
Core Ultra NPUs and the NVIDIA RTX-chiplet SoCs: on-device AI as the client growth story.
Cost discipline
Spend directed only to programs with clear strategic or financial payback.
Manufacturing as national security
Tan frames American chipmaking as central to economic and national security, backed by the government stake.
Panther Lake, Clearwater Forest and the AI PC.
What Intel shipped and announced most recently, and where Agent One plugs in.
Panther Lake (Core Ultra Series 3)
The first client platform on Intel 18A, launched at CES 2026 and broadly available in Q1 2026: an 18A CPU tile, Arc graphics and an NPU for on-device AI.
Xeon 6+ Clearwater Forest
Intel's first 18A data-center CPU, launched around June 2026, with up to 288 E-cores. US-made silicon for the compute grid.
OpenVINO and Core Ultra
The open stack already running private, air-gapped agents on the person's own machine, across CPU, GPU and NPU.
Crescent Island
A data-center inference GPU entering customer testing in the second half of 2026, as Intel focuses on inference GPUs and rack scale with Jaguar Shores in 2027.
Buy together. Build together. Sell together. Take risk together.
The proof is already public: EdgeRunner runs air-gapped, on-device AI agents on Intel Core Ultra with OpenVINO. A private agent stack runs on Intel silicon today. Here is what we will build on it.
Build: Agent One on the AI PC
Core Ultra NPUs and OpenVINO run inference on the person's own machine.
We will ship 🤫 Private Agent One as a first-class AI PC experience on Core Ultra with OpenVINO: consent-first, no phoning home, on the person's own machine.
Intel's NPUs carry a flagship private agent, and the person owns their AI and compute.
Sell: American compute, at both ends
Intel is growing enterprise and government reach and demand for US-made silicon.
Together we market Core Ultra AI PCs as the endpoint the person owns, and US-made 18A Xeon as the compute behind the garage and warehouse grid.
The 🤫 Yellow Pages and directory run on American-made edge compute, and Intel gains a new channel.
Buy: US-made silicon for the grid
Intel Foundry is built around committed volume tied to real customers.
We source US-made 18A Xeon (Clearwater Forest) and Core Ultra for Puppy One and edge nodes: sovereign, and lowest cost per watt where power is cheap.
A distributed-grid buyer for Intel, with a real customer behind every node.
Take risk: a sovereign edge grid
Intel and the US government are betting on American leading-edge manufacturing.
We'd love to design on 14A early and be a reference customer for a sovereign, consent-first edge AI grid.
An American sovereign AI factory, shared and built in the open.
The people building Intel.
Lip-Bu Tan
CEO (since March 2025)
Leads the turnaround: process leadership, Foundry customers, cost discipline and AI, with American manufacturing as the mission.
David Zinsner
EVP and CFO
Leads cost and capex discipline and the path to Foundry break-even.
Michelle Johnston Holthaus
Leads Intel Products
Leads the client, server and edge portfolio that funds the build-out.
Naga Chandrasekaran
EVP, Intel Foundry Manufacturing
Leads fab execution and the 18A and 14A ramp, the US-made supply behind a sovereign edge grid.
American-made, sovereign compute, for those who serve.
Intel is the American-made, sovereign leading-edge silicon story. Pair it with private on-device agents and consent-first design, and you have a template for the nation. 🤫 Agent One is free for life to every federal and state government employee and every past or present service member, a natural fit for a sovereign, American compute stack.