Intel’s AI pivot gains ground as foundry and x86 bets lift Q3 earnings beyond expectations

Why Intel’s Q3 2025 results show early signs that its AI and foundry strategy may finally be working

Intel Corporation (NASDAQ: INTC) has delivered a robust financial performance in the third quarter of 2025, defying expectations and signaling a potential turning point in its multi-year strategic transformation. With revenue reaching $13.7 billion, the company exceeded the upper end of its guidance and reported non-GAAP earnings per share of $0.23. This marked the fourth consecutive quarter where Intel Corporation outperformed its guidance, a consistency that is beginning to restore investor confidence in the company’s long-term strategic direction.

Chief Executive Officer Lip-Bu Tan characterized the quarter as a validation of Intel Corporation’s renewed execution discipline and its evolving role within the broader artificial intelligence economy. According to Tan, the company is steadily rebuilding around core strengths, including its x86 compute franchise, United States-based logic manufacturing, and design services. Notably, Intel Corporation is framing itself not merely as a legacy CPU company but as a hybrid compute enabler capable of powering the next era of agentic and inference-heavy artificial intelligence systems.

Chief Financial Officer David Zinsner noted that the results were also fueled by improved macro demand, a favorable mix, and better-than-expected gross margin performance. While challenges persist in capacity constraints and product ramp costs, the narrative around Intel Corporation’s recovery is becoming less hypothetical and more grounded in quarterly delivery.

How Intel Corporation’s core businesses outperformed and what that says about end-market demand in AI, PCs, and data centers

Third-quarter revenue of $13.7 billion represented a three percent year-over-year increase and a six percent sequential rise. The Client Computing Group, which posted $8.5 billion in revenue, grew five percent compared to the same quarter last year and eight percent quarter over quarter. This upside was supported by strong notebook demand, particularly within consumer and enterprise segments, where Intel Corporation ramped its Lunar Lake and Arrow Lake processors. The ongoing Windows 11 upgrade cycle and renewed interest in artificial intelligence personal computers also contributed meaningfully to the topline.

Data Center and Artificial Intelligence revenue came in at $4.1 billion, slightly down on a year-over-year basis but five percent higher than the second quarter. Enterprise demand was strong, particularly for Intel Corporation’s Xeon 6 processors, known under the Granite Rapids codename. These processors have demonstrated improved power efficiency and total cost of ownership benefits, and demand has been especially strong among hyperscalers, where artificial intelligence workloads are reshaping server architecture priorities.

Operating income for Intel Corporation’s core products was $3.7 billion, or 29 percent of revenue, up nearly $1 billion sequentially. This margin expansion was driven by stronger product mix, lower operating expenses, and the absence of certain period-specific costs that weighed on the previous quarter.

Chief Financial Officer David Zinsner attributed some of the execution success to real-time demand shaping with customers, including pricing adjustments and supply reallocation to prioritize higher-margin segments where Intel Corporation had capacity and customers had demand.

Why x86 architecture is central to Intel Corporation’s artificial intelligence strategy in 2025 and beyond

Chief Executive Officer Lip-Bu Tan made it clear during the earnings call that Intel Corporation’s recovery is rooted not only in structural changes but in a philosophical shift toward engineering-led execution. Central to that vision is a renewed commitment to the x86 instruction set architecture, which Tan argued remains uniquely positioned for the hybrid compute environments of the future.

Tan believes that inference, edge processing, and multi-modal agentic systems are where the next frontier of artificial intelligence lies, and in these domains, x86 remains highly relevant. Unlike graphics processing units that dominate training workloads, Intel Corporation’s CPUs offer the flexibility, installed base, and software ecosystem needed to scale inference workloads in enterprise and embedded contexts.

The company’s collaboration with NVIDIA Corporation is a cornerstone of this strategy. The two firms are jointly developing multi-generational data center and consumer computing products, leveraging Intel Corporation’s x86 CPU strength and NVIDIA Corporation’s artificial intelligence accelerators through NVLink integration. According to Tan and Zinsner, the joint roadmap is the result of nearly a year of deep engineering collaboration and is designed to position x86 as a central compute layer in modern artificial intelligence platforms.

To support this pivot, Intel Corporation has established the Central Engineering Group, a new internal structure tasked with unifying foundational design and architecture work. This group will also spearhead Intel Corporation’s custom silicon efforts, building application-specific integrated circuits for external clients with workloads ranging from general-purpose compute to fixed-function inference accelerators.

How Intel Foundry is reducing losses and advancing 18A milestones despite soft revenue trends

Intel Corporation’s Foundry business, which includes both internal and external customer volumes, reported $4.2 billion in revenue during the third quarter, a four percent sequential decline. However, the segment showed clear signs of operational recovery, reducing its operating loss by $847 million compared to the previous quarter. This improvement was driven by the absence of an $800 million impairment recorded in Q2 and by better-than-expected execution on Intel 10 and Intel 7 volumes.

Importantly, the company reached key milestones on its Intel 18A node and released hardened process design kits for its 18A-P variant. Fab 52, the company’s new high-volume fabrication facility in Arizona, is now fully operational and producing 18A wafers, which are critical to Intel Corporation’s upcoming client and server roadmaps. These wafers are also foundational to United States Government projects under the Secure Enclave program, a key component of the CHIPS Act partnership.

Intel Corporation also continues to develop its Intel 14A node and is expanding its advanced packaging capabilities, including Embedded Multi-Die Interconnect Bridge and its evolution, EMIB-T. Management sees these technologies as critical differentiators, particularly as artificial intelligence accelerators demand new packaging paradigms to address power and thermal density.

Chief Executive Officer Lip-Bu Tan reiterated that the foundry’s growth will be guided by discipline. Capacity expansion will only occur when demand is secured, and operational scalability will be prioritized over speculative buildouts. The overarching goal is to make Intel Corporation a trusted, customer-first foundry that can consistently deliver wafers with power, performance, yield, and schedule reliability.

What the Q4 2025 guidance tells us about Intel Corporation’s near-term margin pressure and capacity planning

Intel Corporation issued guidance for fourth-quarter revenue in the range of $12.8 billion to $13.8 billion. At the midpoint of $13.3 billion, the company expects non-GAAP earnings per share of $0.08 and a gross margin of 36.5 percent, down from 40 percent in the third quarter. This compression reflects the higher costs associated with the initial ramp of Core Ultra 3, as well as the removal of Altera’s financial contribution following its deconsolidation.

Within its product portfolio, Intel Corporation anticipates that Data Center and Artificial Intelligence will grow strongly in the fourth quarter, while Client Computing Group revenue will decline slightly as wafer capacity is redirected toward higher-value server shipments. Intel Foundry is expected to post sequential revenue growth, driven by increased internal demand for 18A wafers and a full quarter of external foundry revenue that is no longer blended with Altera’s contribution.

Inventory levels remain healthy, but capacity constraints on Intel 10 and Intel 7 nodes are limiting Intel Corporation’s ability to fully capitalize on near-term demand. Zinsner noted that while supply tightness is expected to persist into 2026, the company is working closely with customers to manage allocation, pricing, and inventory balancing.

Capital expenditures remain on track at approximately $18 billion for full-year 2025, and Intel Corporation expects to deploy over $27 billion in capital investments this year, up significantly from $17 billion in 2024. This increase is primarily tied to continued investments in United States-based manufacturing, advanced packaging, and the build-out of 18A and 14A nodes.

Why investor sentiment around Intel Corporation is shifting and what risks remain

Institutional investors are beginning to respond more positively to Intel Corporation’s recent performance. The company’s ability to close the Altera divestiture, monetize its Mobileye stake, and secure over $20 billion in funding from partners like NVIDIA Corporation, SoftBank Group, and the United States Government has helped de-risk the balance sheet. Cash and short-term investments now total $30.9 billion, providing ample flexibility to manage debt maturities and future capital needs.

Analysts have noted that Intel Corporation’s positioning in artificial intelligence inference workloads, along with its design services pivot, provides a differentiated approach that contrasts with the dominant training-centric narratives in the market. However, risks remain. Execution on process technology remains a critical watchpoint, and any slippage in roadmap milestones, particularly on 18A and 14A, could impact both customer confidence and revenue visibility.

Furthermore, while Intel Corporation’s focus on x86 revitalization is strategically sound, the broader industry’s transition toward heterogeneous compute and domain-specific accelerators will continue to put pressure on Intel to deliver innovative, timely, and high-performing products that can compete in diverse workloads.

Still, with steady margin recovery, four consecutive quarters of improved execution, and deepening institutional partnerships, Intel Corporation appears to be turning the corner on its turnaround narrative.

Key takeaways from Intel Corporation’s Q3 2025 earnings and strategic direction

• Intel Corporation reported third‑quarter 2025 revenue of $13.7 billion, up 6 percent sequentially and above the high end of its guidance range.

• Non‑GAAP earnings per share stood at $0.23, reflecting higher revenue, improved product mix, and disciplined cost management.

• The Client Computing Group generated $8.5 billion, boosted by strong notebook demand and ramp‑up of Lunar Lake and Arrow Lake processors.

• Data Center and AI revenue reached $4.1 billion, supported by robust enterprise demand for Xeon 6 (Granite Rapids) chips.

• Intel Foundry revenue of $4.2 billion declined 4 percent quarter‑over‑quarter but delivered a $847 million reduction in losses and achieved key 18A process milestones.

• The company secured $20 billion in fresh liquidity from the U.S. Government, SoftBank Group, and NVIDIA Corporation, strengthening its balance sheet to $30.9 billion in cash.

• Intel Corporation is positioning x86 architecture at the heart of AI inference and hybrid computing, aligning with its new Central Engineering Group and expanded ASIC design services.

• Guidance for Q4 2025 projects revenue of $12.8–13.8 billion, gross margin of 36.5 percent, and non‑GAAP EPS of $0.08, with margin pressure from Core Ultra 3 ramp costs.

• Capital expenditures remain near $18 billion, with more than $27 billion expected to be deployed during 2025 to advance U.S. manufacturing and foundry capacity.

• Analysts view Intel Corporation’s progress as a measured but credible turnaround, dependent on consistent roadmap execution across 18A and 14A nodes and sustained AI‑driven foundry demand.