Intel

Intel reported impressive second-quarter revenue of $16.1 billion, up 25% year over year and ahead of the high end of guidance of $14.8 billion. Intel expects third-quarter revenue of $16.3 billion, which would be up 19% year over year and ahead of FactSet consensus estimates of $15.1 billion.

Why it matters: Intel reported a blowout quarter, thanks to tremendous demand for server CPUs to support agentic artificial intelligence. The report was in lockstep with rival AMD’s comments that the server CPU market should grow at a 50% CAGR and be nearly four times higher than its forecast provided in November. Intel’s data center and AI, or DCAI, revenue grew 59% year over year, its highest quarterly growth percentage on record. In turn, Intel’s total revenue growth of 25% was its highest level in over 15 years.

The bottom line: We raise our fair value estimate for no-moat Intel to $105 from $90, driven by more optimistic near- and long-term assumptions for the server CPU end market. Although Intel faces a host of server CPU competitors, we think the rising tide of agentic AI will lift all boats. Shares rose about 3% after hours but are still up over 170% year to date, thanks again to high server CPU demand and steady manufacturing progress. We now view Intel’s shares as fairly valued. In addition to strong DCAI growth, Intel’s PC CPU held up well. The PC market should still be down due to higher memory and component prices, but a favorable product mix toward newer AI PC processors could be an offset for Intel.

Coming up: In addition to 19% revenue growth in the September quarter, Intel forecasts an adjusted gross margin of 42%, up modestly sequentially and about 200 basis points year over year. We’re encouraged with Intel’s manufacturing progress. Yields are improving on Intel’s latest 18A process; its sister technology (Intel 18A-P for external foundry customers) should arrive later this year, and Intel 14A is on pace to reach high-volume production in 2028.

Based on higher server CPU demand, Intel raised its 2026 gross capital expenditure forecast to $20 billion, up from previous guidance of roughly $18 billion. Intel anticipates significantly higher capex in 2027 as well, but didn’t provide a financial figure. We view the higher capex forecast as a good sign, both for demand for internal products and for the potential for meaningful external foundry revenue in the years ahead. We still model $16 billion in external foundry revenue for Intel by 2030, in line with management estimates.

In its prepared remarks, Intel disclosed its commitment to reaching high-volume production of Intel 14A in 2028. This is a positive change in language, in our view, as a year ago CEO Lip-Bu Tan made it clear that he would not commit to 14A production without “a meaningful external customer to drive acceptable returns on our deployed capital, and I will only invest when I'm confident those returns exist.”

Looking at the comment a year later, we think it’s now conceivable that Intel could, in fact, invest mostly in 14A, if not entirely, to satisfy internal manufacturing needs only, given the massive opportunity we see ahead in server CPUs for agentic AI. We model DCAI revenue rising from $16 billion to $54 billion in 2028, thanks to this agentic opportunity that was obscure to all just 12 months ago. However, we also suspect that Intel has external foundry customers within its pipeline—they’re merely striving to not announce any external customers unless the parties do so first. We’d bet that Intel 14A will prosper from both opportunities, not just one.

In advanced packaging, Intel’s EMIB-T process is gaining traction among leading AI customers, given tight capacity in pre-existing chip-on-wafer-on-substrate, or CoWoS, packaging techniques for AI GPUs. Intel thinks it is on target to support customers in 2027, and we anticipate that some of Intel’s first, meaningful external foundry revenue will likely come from EMIB in the quarters ahead.

For further insight into the server CPU market, please see our July 1, 2026 report 'From Tokens to TAM: Mapping the $154 Billion CPU Opportunity.'

Our fair value estimate is $105 per share, which implies a 2026, 2027, and 2030 price/adjusted earnings multiple of 71 times, 40 times, and 24 times, respectively.

Intel’s revenue fell 14% in 2023, due to a significant pause in PC spending after a couple of strong years of upgrades during the covid-19 pandemic, but also because of market share losses and more-cautious spending by data center customers. Intel didn’t see a huge rebound in 2024 or 2025 as market share losses continued, and sales fell 2% and 0.5%, respectively.

We anticipate that strong demand for server CPUs as part of the agentic AI buildout and Intel’s foundry buildout will lead to strong growth, which we model at a 21% CAGR over the next five years and 14% over the next 10 years.

In the DCAI segment, we project 49% growth in 2026, as robust demand for x86 server CPUs boosts the businesses of both Intel and AMD. AMD now thinks the server CPU market will reach $220 billion by 2030, nearly four times higher than its projections just a few months ago, thanks to the rise of agentic AI. We model growth of 54%, 39%, and 27% in 2027, 2028, and 2029, respectively, as Intel prospers from agentic AI. We still model modest share loss for Intel to AMD in x86 server CPUs, and suspect that the x86 CPU camp will concede some share to Arm-based CPUs too. Still, this rising tide should lift all boats, including Intel.

Looking at Intel’s client computing group, which includes the PC CPU business, we anticipate 6% revenue growth in 2026, as memory chip shortages may weigh on overall PC demand throughout the industry, but pricing should remain strong amid processor supply shortages. We model a low-single-digit revenue growth CAGR thereafter. We anticipate that Intel will continue to lose share to AMD and emerging ARM-based CPU vendors, but might be able to stop the bleeding somewhat, thanks to its partnership with Nvidia to include GPU tiles into its x86 CPUs.

Intel is building its foundry business from the ground up. We model $2.1 billion in revenue from Intel Foundry in 2027, primarily because wafers for the formerly internal Altera business unit are now classified as external wafers following Intel's spinoff of Altera. Still, we think that new external customers will gravitate toward Intel, and we model the firm capturing $16 billion in external foundry revenue by 2030 and $50 billion by 2035. We acknowledge that the timing and magnitude of such revenue are highly uncertain. It’s conceivable that Intel’s foundry gains might come from companies that compete with Intel in products.

Intel’s manufacturing missteps and a soft PC market caused GAAP gross margin to crater from 62% in 2018 to 35% in 2025 (and non-GAAP gross margin to fall from 63% to 37%). We think this is the bottom for Intel’s business as we model a 40.5% GAAP gross margin in 2026, expanding to 54% in 2030 and 57.5% by 2035. On a non-GAAP basis, we model a 42% gross margin in 2026, expanding to 55% in 2030 and 58% in 2035.

GAAP operating margin has also plummeted from 33% in 2018 to a 4% operating loss in 2025. We foresee Intel achieving a 4% GAAP operating margin in 2026 and improving to 34% in 2030. On an adjusted operating margin basis, we expect a 15% margin in 2026 and a recovery to 37% in 2030.

We assign Intel a no-moat rating. The company’s returns on invested capital have fallen in recent years, and we do not foresee excess returns on capital in the years ahead. The deterioration stems from the firm’s manufacturing struggles and hefty investment in new manufacturing processes. If Intel were to perfectly execute on its aggressive technological road map, it may warrant an economic moat rating in the future. However, this outcome is likely years down the road with no guarantee of success.

Intel’s chip design team has decades of experience in processor designs, mostly around central processing units, which could conceivably warrant a narrow moat rating as a stand-alone, asset-light business. However, we view the chip manufacturing business as being at a cost and technological disadvantage today, as the firm has stumbled and lost its lead to TSMC, the world’s largest foundry. In our view, Intel’s manufacturing headwinds more than offset any competitive advantages that its chip design expertise may warrant.

Intel is perhaps best known for its CPUs for PC desktops and notebooks within its client computing group, or CCG, business segment. We think that Intel has a competitive advantage over startups and most other digital chipmakers in PC CPU chip designs, not only because of chip design expertise honed over decades, but also because Intel is one of two prominent firms (along with AMD) to hold an x86 instruction set architecture license.

For many years, effectively all PC software was designed for the x86 architecture. Today, however, Apple has already done the heavy lifting of converting its Mac software away from x86 and toward Arm’s instruction set. Microsoft is now supporting Windows on Arm-based devices and may increase these investments over time. Thus, we don’t think the x86 architecture is as moaty as it once was, although we still foresee x86-based processors from Intel and AMD as making up a significant portion of the PC market for the next several years.

Intel also had a historical cost advantage in the PC processor market because of its manufacturing leadership along Moore’s Law, which was coined by one of Intel's chief founders, Gordon Moore, and suggests that the number of transistors per unit area doubles approximately every two years. For decades, the company held a manufacturing advantage that drove PC hardware makers to buy Intel chips, which gave it a bigger R&D budget to invest in the next wave of chip innovations, thus coming out with even better processors and keeping the positive flywheel going.

Unfortunately, this flywheel unwound the other way, as Intel stumbled in internal manufacturing and could no longer keep up with Moore’s Law. Intel conceded PC and server CPU share to AMD and no longer has a large enough business to justify internal-only usage of leading-edge chip manufacturing. Thus, Intel’s viability depends on its ability to attract external foundry customers.

In manufacturing, Intel is making admirable progress on Intel 18A and 14A, coming closer to parity with the chipmaking leader, TSMC. Intel also has interesting advanced packaging assets with EMIB that might attract AI customers. Perhaps more important, the US government owns a stake in Intel and has a national security incentive to ensure that a US company develops leading-edge chip manufacturing on US property. Thus, the US might nudge its leading chip designer firms, such as Nvidia, AMD, Apple, and maybe others, to outsource at least a portion of their chip production to Intel over time.

Nonetheless, the massive costs associated with chip production still lead us to project that Intel won’t achieve excess returns on capital over the next few years. Even with pieces of the business that are hard to replicate, we don’t yet foresee enough excess returns to assign Intel a narrow or wide moat rating.

Intel is one of the largest semiconductor companies in the world and still holds leading share in the PC and server processor markets.

The Intel Products business segment has an innovative, profitable design team, and shifting more production to TSMC might enable it to fend off recent market share losses.

Intel Foundry is one of only three companies that can come close to building leading-edge processors. Additional financial support might enable the business to emerge as a viable manufacturer down the road.

There are no guarantees that Intel can execute on its manufacturing aspirations, such as the release of Intel 14A in 2028 or 2029.

Even if Intel can improve its manufacturing road map, AMD is now a far more credible chip designer in the x86 space for PC and server CPUs, while ARM-based CPUs are a much greater threat than in years past.

Nvidia’s GPUs have captured most of the AI accelerator market, and we don’t foresee Intel making a dent here anytime soon.