Intel unveils major overhauls for AI, foundry, and x86 — but deep losses raise concerns about the future

Intel reported flat second-quarter 2025 revenue of $12.9 billion, but its losses widened to $2.9 billion as gross margins fell amid rising operating expenses. The decline stemmed largely from restructuring charges, including severance payments, asset impairments, and accelerated depreciation. CEO Lip-Bu Tan outlined sweeping operational changes, emphasizing a focus on streamlining the company, reducing redundancies, and improving agility. He revealed plans to cut the workforce to 75,000 by year-end, with over 30,000 layoffs already completed in Q2. Tan stressed the need to retain top talent internally while attracting external expertise, stating, “We need to right-size and scale back the company while ensuring we are retaining our best internal talents and hiring the best external talents.” A key shift in Intel’s strategy involves a renewed emphasis on its x86 architecture. Tan announced the return of Hyper-Threading to client processors, a feature absent since 2017, to boost performance. He also mandated that all major chip designs require his personal approval before finalization, aiming to ensure architectural discipline and simplify product lines. The Panther Lake processor, based on the Intel 18A node, is set for a late 2025 launch, with additional SKUs expected in 2026. While Panther Lake is projected to strengthen Intel’s notebook market position, the company acknowledged struggles in high-end desktops. To address this, Intel will reintroduce simultaneous multithreading (SMT) to P-cores, a move expected to improve competitiveness. For server CPUs, Granite Rapids is progressing well but faces challenges in outperforming AMD’s EPYC line, particularly in performance-per-watt metrics. Intel’s foundry strategy is also undergoing a major overhaul. Tan admitted past investments in manufacturing capacity were “unwise and excessive,” leading to underutilized facilities. Moving forward, the company will align capital spending with customer demand and milestone-based triggers, avoiding overexpansion. Projects in Germany and Poland have been halted, assembly and test operations in Costa Rica are being consolidated, and construction in Ohio will slow unless demand increases. Intel remains committed to its 18A node for internal products, with plans to later offer it to external customers. However, the upcoming 14A node, its first foundry-first design, hinges on securing major external clients. Without such demand, Intel may abandon or delay 14A development, signaling a potential shift away from leading-edge process technology. In AI, Intel is pivoting toward inference and software integration, moving away from its previous fragmented approach focused on training workloads. The new strategy leverages existing x86 CPUs and Xe GPUs to create a cohesive hardware-software platform, with increased investment in AI software to enhance competitiveness. Tan emphasized targeting areas like agentic AI, rather than directly competing with Nvidia’s H100 or B200 accelerators. The company will now prioritize understanding AI workloads before designing hardware, aiming for a full-stack solution. Financially, Intel’s Data Center and AI (DCAI) segment saw a 4% revenue increase to $3.9 billion, driven by stronger server CPU sales. However, average selling prices (ASPs) dropped 8% due to pricing pressures and product mix shifts. The Client Computing Group (CCG) revenue fell to $7.9 billion, with operating income declining due to lower volumes and restructuring costs. Meanwhile, Intel Foundry reported $4.42 billion in revenue but remained unprofitable, posting a $3.17 billion operating loss. Most foundry revenue came from internal use, with external sales at just $22 million. Despite cost savings from restructuring, the segment is expected to break even by 2027. For Q3 2025, Intel projects revenue between $12.6 billion and $13.6 billion, a $200 million decrease from the midpoint. However, gross margins are expected to rise to 36% as restructuring charges primarily impacted Q2. Tan acknowledged the challenges ahead, stating, “We are laser-focused on strengthening our core product portfolio and AI roadmap to better serve customers.” While the changes reflect a strategic reset, the financial strain and competition from rivals like AMD and Nvidia suggest Intel’s path to recovery remains uncertain. The company’s ability to secure external foundry demand and execute its AI initiatives will be critical to its long-term success.