At first, people said Intel was the national stock of the United States, but I was dismissive - why did stock trading turn into MEME? Later, it skyrocketed from 20 to 80, and people were still saying they underestimated it. There's nothing we can do, it's not scary to step into the air, what's scary is not knowing why. Along the way, I have been researching TMD: Intel is actually two companies: One is Intel Products - a mature cash flow machine that sells x86 CPUs (servers+PCs+AI PCs). In 2025, the total revenue is about 46 billion US dollars, with a gross profit margin of 40-45%, which belongs to the typical "long slope, not thin snow" business in the sense of Duan Yongping. The other is Intel Foundry, an advanced process foundry that gambles money on American semiconductor sovereignty. In 2025, external revenue will only be a few hundred million US dollars, with an annual operating loss of 8.5-13.4 billion US dollars. However, we have obtained a 10% equity stake from the US government, 2 billion US dollars from SoftBank, 5 billion US dollars from Nvidia, and the first order from Elon Musk Terafab 14A - this is a typical candidate for a 0-1 monopoly. The valuation logic, growth logic, and risk structure of these two companies are completely different. Put them together with a P/E score, and the result will inevitably be a paste. To see Intel clearly, one must look at it separately. 1、 Watch the trend: Where is Intel in three structural trends Trend 1: AI infrastructure is a structural theme for 10th grade This one is basically undisputed. Gartner、 McKinsey and Goldman Sachs are both saying similar things - by 2028, global inference computing power consumption will reach more than three times that of training, and in China it may be 4:1. The AI inference market is expected to reach approximately $106 billion by 2025 and $255 billion by 2030, with a compound annual growth rate of 19%. More importantly, the economic structure and training of reasoning are different. Training is a one-time investment, GPU is the protagonist; Inference is the continuous consumption of billions of requests per day, and the role of CPU as the scheduling hub is irreplaceable. When Agentic AI goes from concept to enterprise deployment, the main battlefield for serial logic processing, context switching, tool invocation, API orchestration - these workloads - is the CPU. This trend is beneficial for Intel, but Intel is not the biggest beneficiary. The biggest beneficiaries are Nvidia (training), TSMC (manufacturing), and SK Hynix (HBM). Intel is a participant, not the main force. What it can eat is the "control plane" part of the CPU in the AI system - this is an important position, but not the top position. Trend 2: US semiconductor sovereignty is a 10-year irreversible geopolitical strategy This is the truly unique trend for Intel. In August 2025, the Trump government exchanged the outstanding funds of the CHIPS Act for 10% of Intel's equity. This is not a one-time event, but a change in the national industrial policy paradigm. The previous CHIPS was "government funding, companies doing it themselves"; The subsequent model is "government shareholding, government endorsement, and government backstop". The possibility of extending this model to other strategic industries is being discussed, such as nuclear power, critical materials, defense chips, and so on. Intel's unique position in this trend is the only option for advanced process outsourcing in the United States. Even if TSMC's factory in Arizona is put into operation for 2nm in 2028, the US Department of Defense's Secure Enclave project will always only be given to Intel. The Pentagon will not hand over its most confidential chips to a company headquartered in Taipei, regardless of its production capacity. This is a 'secret' in the Thiel sense - a truth that few agree with but may be extremely important: if the situation worsens, America's AI hegemony will be like a tower on sand without Intel. This secret itself may not necessarily need to be fulfilled, but it defines Intel's "bottom line" - the US government will not allow Intel to disappear. The manifestation of this trend in Intel is an exclusive monopoly. Trend 3: Supply chain diversification pressure from ultra large scale customers AWS、 Microsoft, Google, Meta, OpenAI - none of these clients are willing to burden TSMC with all their AI infrastructure. It's not that they don't trust TSMC, it's that no board of directors can accept a 'single point of failure' supply chain. This creates a structural window - there is only Intel in the United States that can provide truly advanced process outsourcing. Microsoft Maia2, AWS custom AI chips, and Google's collaboration on IPUs are all established under this logic. But the benefits of this trend to Intel are conditional. The condition is that Intel can catch the order - with good yield and large production capacity. The harsh reality revealed by institutional level analysis is that the monthly production capacity of Intel 18A is about 10000-15000 pieces, while TSMC N2 is about 1.3-1.4 million pieces, a difference of 10 times. Even if all super large scale customers want to allocate 20% of their orders to Intel, it cannot afford it. The upper limit of the benefits of this trend is locked in by Intel's own capacity ramp up speed. The conclusion of the superposition of three potentials The track Intel is on is real -10 years of progress in AI infrastructure, 10 years of irreversible US semiconductor sovereignty, and 10 years of sustained supply chain diversification. But Intel's position in the three trends is: AI Infrastructure: Participants (led by Nvidia) US Semiconductor Sovereignty: The Sole Main Force (No Alternative) Supply chain diversification: candidate beneficiaries (locked in by production capacity) This is a target that is' powerful but not the biggest beneficiary '. This position has its advantages - the lower limit is supported by national strategy and is not easy to disappear. But it also has its limitations - the upper limit is suppressed by the respective dominance of Nvidia and TSMC, making it difficult to truly move towards explosive growth. 2、 Looking at Business Models: Two Companies, Two Perspectives (Please move to the link for better formatting) Intel INTC Research Report: National Stocks in the United States? How to translate https://mp.weixin. (qq.com)/s/5T_AGkb8D8mU-psyL2dxYA