Intel's Memory 'Pet Project': CEO Lip-Bu Tan Plots CPU-Memory Stacking Comeback

Intel CEO Lip-Bu Tan has signaled a major strategic shift, calling new memory architectures one of his "pet projects" and publicly hinting that Intel may re-enter the memory business after exiting it roughly two decades ago. Speaking about the revival of the American chip industry, Tan said memory was once dismissed as a commodity business but has now become "strategically interesting" — the market is "ripe for innovation," and he is actively exploring ways to stack memory and CPU dies together to solve performance bottlenecks .

CPU-Memory Stacking Vision

Tan directly stated: "CPU and memory, there's a lot of ways we can do stacking together and also try to find some new architecture for memory. I think in some way, in memory, a lot of innovation are not there. I used to not invest in memory because it's kind of commodity business, right? But now [it has] become different. There's a lot of new technology come out" . This points to a 3D-stacking approach where memory dies are integrated directly atop or alongside CPU chiplets — a technique that could dramatically reduce latency and power consumption versus traditional separate memory modules.

The Seok-Hee Lee Hire

On June 18, 2026, Intel appointed Seok-Hee Lee, former CEO of SK hynix and SK On, as Executive Vice President of Intel Foundry, reporting directly to Tan . Lee leads all advanced packaging, system integration, back-end technology development, and back-end manufacturing . This is not a memory-manufacturing hire per se — it is a packaging hire — but it is directly tied to Tan's memory ambitions. As one analyst noted, Lee's expertise in memory integration and advanced packaging makes him the ideal executive to execute CPU-memory stacking . Tan himself welcomed Lee on X, calling him a "close friend" and a "highly regarded semiconductor process integration expert and successful CEO" .

Memory Shortage Context and Intel's Momentum

The memory push comes amid a deepening global chip shortage, particularly in AI-critical memory like HBM (high-bandwidth memory). The worldwide memory crunch has made alternative architectures — including on-package or stacked memory — far more strategically attractive .

Intel's financial momentum provides the firepower to pursue this bet:

• Q2 2026 revenue reached $16.1 billion, up 25% year-over-year — Intel's fastest quarterly growth since 2011 .
• Data center and AI revenue surged to $6.3 billion, up 59% .
• The company raised its 2026 capital expenditure forecast to over $20 billion .
• On August 11, 2026, Intel upsized a stock offering to $20 billion (210.5 million shares at $95) to fund its AI and manufacturing expansion .

A Potential Return to Memory

Intel famously exited the DRAM business in the mid-1980s after Japanese manufacturers commoditized the market, pivoting instead to microprocessors. Tan's language marks the first time a sitting Intel CEO has spoken so explicitly about re-engaging with memory technology.

That said, Intel is not likely to build commodity DRAM or NAND fabs from scratch. The more probable path is that Intel pursues proprietary, stacked memory architectures using advanced packaging — effectively designing its own memory solutions that integrate tightly with its CPUs and AI accelerators, produced through its foundry or in partnership with memory makers like SK hynix (where Lee's deep relationships could be invaluable) .

The combination of a massive capital raise, Lee's packaging expertise, Tan's stated "pet project" focus, and a booming revenue base suggests Intel is laying the groundwork to reinvent how memory and logic are integrated — a move that would be the most significant strategic pivot since the company abandoned memory decades ago.