Intel’s Foundational Pivot: Navigating the Delicate Balance of Manufacturing Recovery and Merchant Growth

Intel’s second-quarter financial results have provided the most compelling evidence to date that the company’s ambitious multi-year manufacturing overhaul is gaining genuine traction. With Intel Foundry revenue climbing to $5.8 billion—a significant leap from $4.4 billion in the same period last year—the company is successfully demonstrating that its internal engine is firing more efficiently. However, as the semiconductor giant works to transform its fabs into a world-class merchant-foundry business, the industry remains focused on a critical question: Can Intel transition from a captive manufacturer of its own chips to a trusted partner for the global technology ecosystem?

Main Facts: A Turnaround in Progress

The financial data released in Intel’s Q2 report reveals a stabilizing manufacturing arm. Intel Foundry’s operating loss has narrowed substantially, falling from $3.2 billion to $2.1 billion, with operating margins improving from negative 71.7% to negative 36.2%. Leadership within the company attributes this progress to a "triple threat" of operational improvements: enhanced wafer yields, reduced production cycle times, and a significant increase in overall fab scale.

While these internal metrics are undeniably positive, they underscore a lingering reality: the vast majority of this revenue is derived from manufacturing Intel’s own product lines. According to CFO David Zinsner, only $293 million of the segment’s quarterly revenue came from external customers. This disparity highlights the fundamental challenge Intel faces—proving that its fabs are capable of supporting external third-party companies at scale, rather than just serving as an in-house production facility.

Chronology: The Road to the Fortinet Partnership

The journey toward a diversified foundry model has been marked by a series of technical milestones and strategic realignments.

• Early 2024: Intel began aggressive internal restructuring to separate its product design and manufacturing divisions, laying the groundwork for the Intel Foundry reporting segment.
• Mid-2024: Intel accelerated the development of its Intel 18A process node, focusing on defect density and power-performance-area (PPA) improvements to attract external design wins.
• Late 2024/Early 2025: Intel deepened its engagement with various partners in the networking and cybersecurity sectors, specifically targeting firms that require custom silicon but lack the massive R&D infrastructure of tech giants like Apple or Nvidia.
• Q2 2026: The formal announcement of the partnership with Fortinet. This agreement serves as a critical stress test for Intel’s ability to handle external clients. Intel confirmed to EE Times that this is a substantive commercial engagement, marking a shift from theoretical technology evaluation to actual production.

Supporting Data: Capacity, Efficiency, and Capital Allocation

Intel’s capital expenditure strategy has become increasingly aggressive. The company expects to spend more than $20 billion in 2026, with projections indicating further increases in 2027. This spending is heavily earmarked for manufacturing tools, front-end capacity, and the advanced packaging technologies—such as EMIB (Embedded Multi-die Interconnect Bridge)—that are becoming essential for modern, chiplet-based designs.

The efficiency gains are visible in the product pipeline. Intel reported that 18A output grew more than 50% sequentially, exceeding internal targets by approximately 25%. Furthermore, the production cost of the company’s Panther Lake architecture has been slashed by nearly 50% during the first half of the year.

However, these figures create a "chicken and egg" scenario. Are these investments being made purely for external foundry growth, or are they necessitated by the supply constraints of Intel’s internal products, such as the Granite Rapids server processors? Management insists that these investments are aligned with customer demand, but analysts remain cautious about the lack of granular transparency between spending on internal versus external manufacturing needs.

Official Responses and Strategic Vision

Intel’s leadership, including CEO Lip-Bu Tan, maintains a focus on technical milestones. During the latest earnings call, Tan highlighted that the 18A-P process has entered risk production, and that engagement for the 14A node—slated for high-volume manufacturing in 2028—is trending positively.

When pressed on the lack of major external, high-volume production announcements, Tan emphasized that the current strategy is built on "purpose-built silicon." By combining Intel’s intellectual property, custom ASIC design services, and advanced packaging, the company is positioning itself to compete with firms like Broadcom and Marvell, rather than fighting a direct, head-on war with TSMC for commodity wafer volume.

"Fortinet is an Intel Foundry customer," an Intel spokesperson clarified in an exclusive exchange. "They are using Intel’s custom ASIC capabilities built for Intel 4 to develop and produce a solution tailored to their requirements." This confirms that Intel is not just selling wafer space; they are selling a comprehensive development and manufacturing lifecycle.

Implications: The Quest for Trust and Optionality

The Fortinet partnership serves as a barometer for whether Intel can successfully pivot its culture. Historically, Intel’s manufacturing organization was optimized exclusively for its own designs. A merchant foundry, by contrast, must be hyper-responsive to the disparate needs of external clients.

Stephen Sopko, an analyst-in-residence at HyperFRAME Research, suggests that the industry should view Intel as a provider of "supply chain optionality" rather than an immediate threat to TSMC. "TSMC is so overcommitted," Sopko noted. "They’re running flat out to meet demand. Somebody else being able to produce chips and package them is a net positive for the entire industry."

The "Trust" Barrier

The ultimate hurdle for Intel is not just technical performance—it is institutional trust. For a company like Intel to succeed as a foundry, it must convince customers that their intellectual property is safe and that their orders will not be relegated to the back of the queue when Intel’s own products are in short supply.

Because Intel remains a vertically integrated company with its own product division, the potential for a conflict of interest is the primary concern for potential clients. Intel Foundry’s separate reporting structure is a start, but deep-seated cultural change is required to convince a competitor to Intel Products to trust their flagship designs to Intel’s fabs.

Future Outlook: Beyond Optionality

The current phase of Intel Foundry is defined by "optionality." Customers are currently evaluating Intel as a second-source supplier, perhaps moving one or two chips to Intel’s nodes while keeping the rest of their portfolio at TSMC. This is a prudent, low-risk approach for the customer, but it is not the end goal for Intel.

To truly secure its future, Intel must reach a point where a major, neutral player chooses Intel as their primary foundry partner, with a secondary provider serving as the backup. Until that happens, Intel’s progress will continue to be measured by how well it fills its own fabs and how effectively it can integrate its custom ASIC capabilities into the broader cybersecurity and cloud infrastructure markets.

As we look toward 2027 and 2028, the success of the 18A and 14A nodes will be the final judge. If Intel can prove that its process nodes offer a tangible performance advantage that cannot be replicated elsewhere, the question of "internal versus external" will likely fade. For now, the semiconductor world remains in a period of cautious observation, waiting to see if Intel can indeed master the complex art of the merchant foundry.

The path is clear: Intel has the fabs, it has the capacity, and it has the desire. The remaining variable is the industry’s collective willingness to embrace Intel as a foundry partner, a choice that will ultimately depend on Intel’s ability to remain neutral, reliable, and technically superior in a crowded global market.