Intel Corporation designs and manufactures semiconductor products used across computing and related markets worldwide, serving consumers, enterprises, governments, and educational institutions. Its defining trait is that it is an integrated design manufacturer, or IDM: it both designs chips and manufactures most of them itself, unlike fabless rivals such as AMD and NVIDIA. Intel frames itself as the only company in the U.S. running leading-edge process research and high-volume logic manufacturing at scale, a position it ties to national economic and security relevance.
That dual identity is the core of Intel’s business model. In 2025, Intel pushed harder to monetize both halves of that identity independently, most notably by opening its manufacturing arm, Intel Foundry, to outside customers rather than serving only Intel’s own product groups. The workforce behind this stood at 85,100 people as of December 27, 2025. The data below shows a company still profitable on the design side while absorbing heavy losses on the manufacturing side.
How Does Intel Make Money
Intel’s core revenue comes from designing, marketing, and selling CPUs and adjacent silicon: AI accelerators, network interface cards, infrastructure processing units, custom ASICs, and discrete GPUs. Sales run through a direct global sales organization serving large OEMs and cloud providers, plus indirect channels of distributors, resellers, and retailers reaching smaller, fragmented buyers.
A second, structurally different revenue stream is third-party foundry services, where Intel Foundry manufactures chips designed by outside companies for wafer fabrication, packaging, and design-enablement fees rather than product margins. Layered on top are Intel’s Mobileye stake (roughly 80% owned, a global leader in driving-assistance and self-driving systems) and IMS Nanofabrication stake (roughly 68% owned, maker of multi-beam mask-writing tools), both contributing revenue outside the core CPU business.
Business Segments: CCG, DCAI, and Intel Foundry
Intel reports three reportable segments plus a non-reportable “all other” category. Intel Products, grouping CCG and DCAI, generated $49.1 billion in FY2025 revenue, down about 1% year-over-year — a soft but stable base that still funds most of the company’s profitability.
CCG (Client Computing Group) covers PCs and edge devices, selling mainly through distributors and OEMs into consumer, education, healthcare, finance, and government markets, plus emerging AI PC and edge-robotics use cases. CCG revenue was $32.2 billion in FY2025, down 3% year-over-year, but operating income was roughly $9.3 billion — still Intel’s most profitable segment by a wide margin and the segment effectively subsidizing the foundry build-out.
DCAI (Data Center and AI) sells x86 CPUs, AI accelerators, NICs, IPUs, and custom ASICs to hyperscalers, OEMs, telecoms, and HPC customers. Xeon processors drove most DCAI product revenue in 2025. DCAI revenue reached $16.9 billion, up 5% year-over-year, with about $3.4 billion in operating income — modest growth against entrenched GPU competition in AI workloads, and a segment where Intel is fighting to stay relevant as AI compute spending shifts toward accelerators.
Intel Foundry develops leading-edge process nodes, manufactures Intel’s own chips, and increasingly sells wafer fabrication, packaging, and design-enablement services externally. Foundry revenue was $17.8 billion, up 3%, but the segment posted an operating loss of roughly $10.3 billion — the clearest single number showing how expensive this strategic pivot remains. The “all other” category (Mobileye, IMS Nanofabrication, and Altera through September 2025) added about $3.6 billion in revenue and $264 million in operating income, a small but profitable sliver next to Foundry’s losses.
Products and Manufacturing Technology
Intel’s 2025 product story centers on the Intel 18A process node, used for Core Ultra Series 3 processors and the first Intel node combining gate-all-around transistors (RibbonFET) with backside power delivery (PowerVia) at volume. Intel is simultaneously developing Intel 14A, its next node, designed from inception to also serve external foundry customers rather than only Intel itself.
Earlier nodes still carry real volume: Intel 7 remains widely used internally, while Intel 4 and Intel 3, Intel’s first EUV-based nodes, support current Xeon and client lines. Packaging technology includes EMIB for 2.5D integration (with a new EMIB-T variant in 2025) and the Foveros family for 3D die-stacking, plus a newly announced Foveros Direct aimed at future hybrid bonding by 2028.
On products, the Core Ultra family introduced AI PCs in its first generation and has since added dedicated NPUs and Foveros packaging, alongside the Panther Lake client platform for AI-enabled edge workloads. On the data center side, Intel is developing discrete GPUs under the Crescent Island (AI inference) and Jaguar Shores codenames, an area where it still trails NVIDIA and AMD commercially.
Manufacturing runs across a global fab and assembly network. Key 2025 wafer fabrication sites included Oregon (leading the Intel 18A ramp), Arizona (Intel 7 and 18A ramp), Ireland (Intel 4 and Intel 3), and Israel (Intel 7), with assembly-and-test operations spread across China, New Mexico, Vietnam, and Malaysia. This geographic spread is central to Intel’s pitch that it offers supply resilience outside Asia-concentrated foundry capacity.
The IDM Strategy and Foundry Ambitions
Intel’s stated intent is to keep manufacturing most of its own products internally, arguing this optimizes performance, speeds time-to-market, and reinforces its role as a U.S.-based leading-edge manufacturer with national-security relevance. At the same time, Intel pragmatically uses external foundries, including rival TSMC, for select tiles or entire products as it moves toward disaggregated, chiplet-based designs.
This ambition carries real conditionality: Intel has stated that if it cannot secure a significant external customer for Intel 14A, it may pause or discontinue that node and future leading-edge nodes, shifting more manufacturing to TSMC over time. That single disclosure is arguably the biggest strategic risk embedded in the entire business model, since it makes Intel’s long-term manufacturing roadmap contingent on winning a customer it does not yet have.
In 2025, Intel also streamlined its physical footprint — consolidating Costa Rica assembly-and-test operations, slowing Ohio fab construction, and discontinuing planned Germany and Poland expansions. This signals a more disciplined, demand-matched capital approach after years of aggressive build-out under the earlier IDM 2.0 expansion plan.
Customers and Distribution Channels
Intel’s customers range from individual consumers to the world’s largest cloud operators. CCG’s direct customers are PC distributors and device OEMs; DCAI’s direct customers are hyperscale cloud providers and server OEMs, whose own downstream buyers include enterprises, ISVs, systems integrators, and governments. Intel does not publicly name customers exceeding 10% of revenue, instead pointing to segment-note disclosures, typical of large B2B suppliers protecting concentration data.
Distribution runs through a direct global sales force for the largest accounts plus indirect distributors, resellers, and retailers for smaller markets. Demand is seasonal, historically weighted toward the second half of the year and peaking in Q4, and Intel notes customers can cancel or delay orders with little notice — a structural volatility factor that shows up clearly in quarter-to-quarter revenue swings.
Strategic Capital Partnerships in 2025
2025 brought an unusually dense run of capital-structure events tied to funding Intel’s manufacturing transformation. In August, the U.S. Department of Commerce received 275 million Intel shares plus a warrant for up to 241 million more at $20.00 per share, while Intel drew the remaining $5.7 billion owed under its CHIPS Act agreement; 159 million more shares went into escrow as part of the arrangement.
That same month, SoftBank Group bought 87 million shares at $23.00 per share for a $2.0 billion cash investment. In September, NVIDIA agreed to buy 215 million shares at $23.28 per share for $5.0 billion, alongside a partnership to co-develop custom x86 and NVIDIA AI products across hyperscale, enterprise, and consumer markets — a deal that effectively makes a top GPU rival a shareholder and product partner at once.
Also in September, Intel completed the divestiture of 51% of Altera for net proceeds of $4.3 billion, recognizing a pre-tax gain of $5.6 billion. Together, these deals brought billions in fresh capital from a government stakeholder and two major tech partners in a single year — strengthening cash reserves while diluting shareholders and tying Intel closer to NVIDIA’s AI ecosystem and U.S. industrial policy.
Internally, Intel restructured around four priorities: transforming its culture, revitalizing the x86 ecosystem, growing external foundry business, and expanding market opportunities. It simplified its organizational structure, cut management layers, and unified engineering functions — IP development, test-chip design, EDA tooling, design platforms — into one group supporting a planned ASICs and design-services business.
Competitive Positioning
Each segment faces a distinct rival set. In CCG, AMD is the primary x86 competitor, alongside ARM-based threats from Apple’s M-series, Qualcomm’s Snapdragon, and MediaTek’s Kompanio chips — competition Intel expects to intensify through 2026 as ARM-based laptops gain share.
In DCAI, AMD competes on x86 servers while NVIDIA’s GPUs dominate the highest-demand AI compute segment; several hyperscaler customers are also competitors through custom silicon programs, and Broadcom competes on ASIC design wins. In foundry, TSMC is the dominant leading-edge rival, with Samsung, GlobalFoundries, UMC, and SMIC also competing for external wafer business.
This makes Intel unusual among large semiconductor companies: it competes simultaneously against fabless designers, pure-play foundries, and hyperscalers’ internal silicon programs, all from one vertically integrated structure. Few rivals face pressure on that many fronts at once.
Industry Context and Human Capital
Intel has been part of the computing industry for more than 50 years, and much of its current business model was reshaped by the IDM 2.0 strategy launched a few years ago to convert manufacturing from an internal-only function into a customer-facing foundry business. That shift is still in its early innings: management itself acknowledges having “few external customers to date” for foundry services, even as it commits multi-billion-dollar capex to the effort.
The broader semiconductor industry backdrop matters here too. Demand has increasingly concentrated around AI accelerators and GPU systems, a category where NVIDIA leads and where Intel’s own DCAI accelerator and custom-ASIC lineup is still building credibility. Intel’s model is a bet that x86 CPUs, AI PCs, and a revived foundry business can each carve out defensible share even as GPU-centric compute captures a growing share of industry capital spending.
On human capital, Intel’s workforce of 85,100 people as of the end of FY2025 supports design, manufacturing, and sales functions across a global footprint. Talent retention in advanced process engineering is a specific point of competitive importance, since Intel’s foundry ambitions depend on the same specialized workforce that TSMC and Samsung also compete to hire.
Financial Performance: FY2025
Intel closed fiscal 2025 (ended December 27, 2025) with total net revenue of $52.9 billion, essentially flat against $53.1 billion in FY2024. Q4 revenue was $13.7 billion, down 4% year-over-year. GAAP gross margin improved to 34.8% from 32.7%, and non-GAAP gross margin rose to 36.7% from 36.0% — modest but real manufacturing efficiency gains despite Foundry’s losses.
Intel narrowed its GAAP net loss to $267 million for the year, from an $18.8 billion loss in FY2024, driven substantially by the Altera divestiture gain. Non-GAAP net income was $1.9 billion versus a non-GAAP loss the prior year. GAAP diluted EPS was $(0.06) versus $(4.38); non-GAAP diluted EPS was $0.42 versus $(0.13), showing the underlying operating business is closer to profitability than the GAAP headline suggests.
Full-year operating cash flow was $9.7 billion, and gross capital expenditures totaled about $17.7 billion — still heavy investment even after trimming some expansion plans. Intel ended the year with $37.4 billion in combined cash and short-term investments, up sharply from $22.1 billion a year earlier, aided directly by the government, SoftBank, and NVIDIA equity infusions.
Total assets grew to $211.4 billion from $196.5 billion, and total stockholders’ equity rose to $114.3 billion from $99.3 billion, largely reflecting the new equity issued to the U.S. government, SoftBank, and NVIDIA. Total debt stood at $46.6 billion, split between $2.5 billion short-term and $44.1 billion long-term, a load the company can service more comfortably now given its larger cash cushion.
R&D spending was $13.8 billion, down from $16.5 billion in 2024 and $16.0 billion in 2023, directed toward AI integration, heterogeneous computing, and next-generation nodes. The declining R&D trend alongside rising capex suggests Intel is shifting spend from research toward the physical build-out of manufacturing capacity. For Q1 2026, Intel guided to revenue of $11.7-12.7 billion, GAAP gross margin of 32.3%, and roughly breakeven non-GAAP EPS.
Outlook
Intel’s business model in 2026 is a company mid-transition: a legacy, profitable PC and data-center design business (CCG and DCAI) funding a capital-intensive attempt to turn manufacturing into a foundry business that can win outside customers the way TSMC has. The segment-level numbers make the trade-off explicit — CCG and DCAI together earned roughly $12.7 billion in combined operating income while Intel Foundry lost about $10.3 billion.
The 2025 capital raises from the U.S. government, SoftBank, and NVIDIA, plus the Altera divestiture, show a company actively reshaping its balance sheet and alliances to buy time and capital for that manufacturing bet. Whether Intel Foundry can move from a loss-making internal function to a credible third revenue leg, particularly by landing a marquee external customer for Intel 14A, is the central question this business model now has to answer.
