Intel @ BOFA Global Tech Conference 2026. Executive Summary & Full Transcript
Event: Fireside chat, Bank of America Global Technology Conference, June 2, 2026
Speakers: David Zinsner (EVP & CFO), John Pitzer (CVP, Investor Relations & Treasurer)
Host: Vivek Arya (BofA)
Date: June 2, 2026
Executive Summary
Intel’s CFO used the BofA fireside chat to frame the company’s roughly 14-month turnaround under CEO Lip-Bu Tan as a culture and execution story first, and a strategy story second. Zinsner argued that Intel’s two core assets, decades of x86 CPU design and front-end/back-end manufacturing, were never the problem; execution was, and the root cause was cultural.
Tan’s early moves were structural: collapsing management layers from twelve to six, cutting the VP count from over 400 to around 200, and reducing headcount from above 100,000 to under 80,000, paired with a “bring me the problem or get fired” transparency mandate. Zinsner credited this with the improved Q1 results and the current-quarter guide. A parallel push on financial discipline strengthened the balance sheet, cut operating expenses, and recruited outside talent into CEO-staff roles.
On products, management leaned into the view that AI is turning the data-center CPU market from a GDP-plus grower into an explosive one, as the workload mix shifts from training toward inference, agentic, and multi-agent computing. Zinsner declined to endorse a specific TAM figure but insisted the market is “big enough” that execution on supply is the real constraint.
He flagged genuine ASP strength: rising core counts lift data-center ASPs mechanically, and for the first time in years like-for-like ASP-per-core pricing has stabilized and in places is rising. Intel is using long-term agreements to lock in price and volume rather than riding a supply shortage, which Zinsner contrasted explicitly with the memory market’s boom-bust dynamics from his prior career at Micron.
On competition, management acknowledged x86 rival strength and conceded Intel is weaker in multi-threaded data-center performance: Diamond Rapids will lack SMT, with simultaneous multithreading returning in the following product. Client is positioned as the bright spot, with 18A described as the fastest-ramping client product in five years, now running in two fabs (Oregon and Arizona). Pitzer added that NVIDIA’s client push validates rather than threatens the x86 ecosystem, and that client CPUs will be a major beneficiary of physical AI and edge AI.
On foundry, Zinsner said 18A yields are now improving at an industry-standard monthly cadence after stabilizing performance first, helped by opening data to vendors. He suggested some yield milestones may be pulled in by a quarter or more, with Foundry break-even targeted as Intel exits 2027, and claimed 14A is already ahead of where 18A was at the equivalent maturity. The financial north star is a multi-year “rule of 45” aspiration, underpinned by Pitzer’s argument that Intel can shift from a historical low-single-digit grower to structural double-digit growth.
On capital intensity, Zinsner argued Intel’s combined manufacturing-plus-product margin stack lets it fund heavy capex better than pure-play peers. He intimated that Intel’s capital intensity may have to rise again in the medium term if demand is sufficiently strong. He noted that Intel was in a good position in terms of having a number of “empty shell” buildings waiting in the wings should they be needed.
Full transcript below the fold.
