INTEL CORP

INTEL CORP is a US-listed company materially exposed to Sovereign Cloud & Classified Compute.

Intel is an integrated device manufacturer (IDM): it designs and manufactures semiconductor products. That is the single most important structural fact about Intel. Most competitors are either fabless designers — AMD, NVIDIA, Qualcomm — or pure-play foundries — TSMC, GlobalFoundries. Intel operates both sides, which gives it potential supply-chain control but also exposes it to massive capital intensity and process-node execution risk.

Intel’s FY2024 10-K reports total net revenue of $53,101 million, down from $54,228 million in 2023 and $63,054 million in 2022 [S1, Item 8, Consolidated Statements of Operations].

Intel’s segment reporting in FY2024 groups revenue into Client Computing Group (CCG), Data Center and AI (DCAI), Network and Edge (NEX), Intel Foundry, Altera, and All Other. The following table is built from Intel’s FY2024 10-K segment data [S1, Item 8, Segment Results].

Percentages may not sum exactly due to rounding.

The key takeaway from the mix is concentration. Client Computing plus Data Center and AI represent roughly 81% of Intel’s total revenue. Client is the largest segment, but Data Center and AI is the strategic battleground because it carries higher margins and is the natural entry point for AI compute. Intel Foundry — the revival leg — is still only about 3.5% of revenue at roughly $1.9 billion. That is a tiny base for a strategy that requires tens of billions of capex.

Intel’s FY2024 consolidated income statement provides the following exact or directly derived figures [S1, Item 8, Consolidated Statements of Operations]:

The gross margin of 32.7% is historically low for Intel and reflects both competitive price pressure and significant manufacturing/foundry charges. R&D at 30.9% of revenue is extremely high — Intel is effectively spending nearly one-third of every revenue dollar on R&D. That is structurally different from a fabless company like AMD, whose R&D burden is lower as a share of revenue because it does not carry process-development costs.

A useful unit-economics calculation is R&D intensity:

\[

\text{R&D intensity} = \frac{\text{R&D}}{\text{Net revenue}} = \frac{16{,}428}{53{,}101} = 30.9\%

\]

That is not necessarily “bad” in the long run if it funds a return to process leadership, but it means Intel has very little flexibility to cut costs without cutting its strategic engine.

Intel monetizes primarily through:

• Processor unit sales — Core and Xeon processors sold to OEMs, ODMs, distributors, and cloud hyperscalers.
• Platform and adjacent silicon — chipsets, networking silicon, Ethernet controllers/IPUs, FPGAs, and accelerator cards.
• Foundry services — wafer fabrication sold to external customers under Intel Foundry.
• Full platform economics — a single Xeon server socket also pulls chipset, Ethernet, and potentially accelerator revenue.

Intel does not have a meaningful recurring software/services revenue stream. Its “oneAPI” software ecosystem is open and used to drive hardware sales rather than direct license revenue.

Essentially, four groups pay Intel:

1. PC OEMs — such as Dell, HP, and Lenovo — buy Core processors for laptops/desktops.
2. Server OEMs/ODMs and hyperscalers — directly or indirectly buy Xeon processors for cloud and enterprise servers.
3. Network and edge equipment vendors — buy Ethernet, IPU, and edge processors.
4. External foundry customers — buy wafer fabrication; this is the newest and least-established revenue stream.

Intel’s FY2024 10-K concentration disclosure names Dell as a major customer, accounting for about 19% of 2024 net revenue [S1, Item 1 / Risk Factors / Concentration of Credit Risk; exact section needs primary source validation]. No other customer is disclosed as exceeding 10% in the same filing. This makes Dell a load-bearing customer relationship. [SINGLE-SOURCE]

Intel’s model remains enormously capital-intensive. Intel is spending billions per year on fab construction, EUV lithography, packaging, and process development. Intel’s plan to restore process leadership and attract external foundry customers requires continued high capex even while profits are negative. If Intel were fabless, it would be a very different financial risk profile; as an IDM, its failure mode is capex-funded cash burn without a commensurate revenue ramp.

From Intel’s FY2024 10-K, the customer classes can be grouped as follows [S1, Item 1, Item 1A, Segment Results]:

The most important named customer is Dell, at about 19% of FY2024 revenue [S1]. That is a meaningful concentration risk and also a signal that Intel’s PC/server OEM channel is still crucial.

Intel is not heavily dependent on any single customer outside Dell, but the combined dependence on a handful of large PC and server OEMs means that design wins at Dell, HP, and Lenovo decide much of Intel’s revenue. If AMD wins a major consumer or commercial PC design cycle at Dell or HP, Intel feels it immediately in the CCG segment.

• INTC — Intel SGX is the TEE foundation cited in Microsoft’s Azure confidential computing work ([Intel SGX overview](https://www.intel.com/content/www/us/e

• - INTC — Intel SGX is the TEE foundation cited in Microsoft’s Azure confidential computing work ([Intel SGX overview](https://www.intel.com/content/www/us/e source

Formula:

\[

\text{Intel 2024 growth} = \frac{53{,}101 - 54{,}228}{54{,}228} = -0.0208 = -2.1\%

\]

\[

\text{Industry gap} = 19.1\% - (-2.1\%) = 21.2 \text{ percentage points}

\]

Inputs:

• Intel FY2024 revenue: $53,101M [S1]
• Intel FY2023 revenue: $54,228M [S1]
• Global semiconductor industry growth: +19.1% [S3]

Sensitivity: If global semiconductor sales growth had been 15% instead of 19.1%, the gap would still be 17.1 points. If Intel’s 2024 revenue had been $56,000M instead of $53,101M, its growth would have been +3.3%, but the industry gap would still be about 15.8 points. The conclusion — massive market underperformance — is not sensitive to minor input errors.

This calculation is illustrative only and uses assumed inputs because Intel does not provide a TAM in its 10-K.

Formula:

\[

\text{Incremental DCAI revenue} = \text{AI accelerator TAM} \times \text{Intel target share} \times \text{platform capture rate}

\]

Inputs:

• 2027 AI accelerator TAM: $300 billion (assumed; needs primary source validation)
• Intel target share: 5% (assumed)
• Platform capture rate: 0.60 (assumed, because accelerator sales often pull Xeon CPU and platform revenue)

\[

\text{Incremental DCAI revenue} = 300B \times 0.05 \times 0.60 = \$9.0B

\]

Sensitivity:

What this shows: a modest AI accelerator win rate of 5% at a $300B TAM would add $9B of revenue — roughly 70% of Intel’s entire current DCAI revenue. That is why this segment is the core line in the investment narrative. It also works in reverse: if Intel captures close to zero, the DCAI segment remains a melting ice cube.

Intel’s revenue peaked in FY2022 at $63,054M [S1]. The two-year CAGR is:

\[

\left( \frac{53{,}101}{63{,}054} \right)^{(1/2)} - 1 = -8.2\%

\]

Two years of negative compounding at that scale explains much of the market’s skepticism. This is not a one-quarter miss; it is a multi-year product-cycle and competitive failure.