From Singapore: Intel's costly gamble on semiconductor manufacturing is showing its first real signs of market traction. During an earnings call this week, Chief Financial Officer David Zinsner told analysts that the chipmaker's Foundry division is close to closing contracts worth billions of dollars annually, pivoting away from the core manufacturing focus that has drained billions quarterly from Intel's balance sheet.
The breakthrough involves advanced packaging technology, an area where demand has intensified amid the AI boom as accelerators have grown more complex, with many GPUs composed of multiple compute and memory dies that must be fused together. This represents a significant commercial opportunity for Intel, which has developed proprietary solutions in this space that could compete directly with established players like Taiwan Semiconductor Manufacturing Company.
Intel has invested heavily in its EMIB and Foveros 2.5D and 3D packaging tech to support multi-die processors like its Xeon CPUs and (now discontinued) Ponte Vecchio family of GPU Max accelerators. The technology advantage is substantial. That tech isn't limited to chips Intel makes in its own fabs; the company has already used it to meld silicon manufactured in-house with chips made by TSMC.
Yet the most striking development involves Intel's 18A manufacturing process. CEO Lip-Bu Tan, who joined Intel in early 2025, originally thought the company should focus on 14A as a foundry node and make 18A just an internal node, but now recognizes it as a good node to offer to external customers.Intel has been getting inbound interest in 18A-P as a foundry node, according to Zinsner. The shift suggests Tan is embracing a broader commercial strategy than his initial assessment implied.
The timing matters.When Tan joined Intel in early 2025, both the functional and parametric yields of chips made using 18A were low and unpredictable, leading him to consider shifting foundry efforts to 14A. Manufacturing yields have since improved.While volatility remains with some wafers yielding significantly less and others more, Tan is focused on minimizing wafer-to-wafer variation, with good improvement achieved and steady yield progression expected through 2026.
For shareholders and Intel's broader turnaround, this matters becausethe Foundry division has cost the company billions each quarter. Zinsner offered a cautious timeline, noting that the division remainsat least a year away from breaking even, with the company expecting to exit 2027 at break-even operating margins. He also raised a subtle but important caveat: higher-than-expected external demand could delay profitability by forcing Intel to spend more on manufacturing capacity expansion. For Intel, he joked, that would be a good problem to have.
The advance in packaging and process technology masks a genuine complexity.A number of large fabless chip designers have evaluated Intel's 18A and 18A-P process technologies, though none have made publicly-announced commitments to use the node. Intel's track record on delivery timelines has been questioned in the past, and the company will need to execute flawlessly to convert interest into signed contracts.
From an Australian perspective, this matters for supply chain security and semiconductor sourcing.American chip developers may increasingly outsource production of non-core products to Intel to reduce geopolitical risks associated with Taiwan and potential tariffs on chips made outside of America. A functioning Intel foundry could reshape semiconductor dependencies across the Indo-Pacific, though Intel must first prove it can compete reliably on quality, cost, and delivery.