Intel: From Falling Knife to Portfolio Multiplier

Intel: From falling knife to portfolio multiplier

Intel can report zero earnings per share and still multiply the money of those who bought at the bottom. The story began with an explicit recommendation: enter Intel with a forecast of a minimum 30% rise in one or two years. The most repeated response at first was the falling knife. The figures circulating now in that same conversation tell a quite different story.

The initial thesis: cheap assets and a plan for NPUs

The starting point was simple. Prices at rock bottom, a failed processor generation, and loss of market share to competitors, especially in servers. From this came the conclusion: a company with Intel's assets is undervalued. The future argument relied on NPUs, processors that run artificial intelligence without depending on a graphics card and, according to this reading, will eventually be integrated into all kinds of devices, even appliances.

On this basis was added a factor not found in any balance sheet: direct backing from the United States. The idea that Intel is too strategic to be allowed to fail reappears in various parts of the discussion, even among those who criticize its management.

The bearish argument: lower profits and the GPU train

The most repeated objection is accounting-related. Fewer profits are generated, and the market has shifted toward GPUs, territory where Nvidia rules without dispute. The comparison with IBM appears several times, with the warning that Intel could end up living off patents and little else. With a high P/E ratio of 20 for a company that generates no profits, the conclusion of this stream is that there is no bargain.

There is more. The factories announced in Germany with public funding were, according to this version, left on the edge. And the management diagnosis is harsh: an ossified structure that needs a thorough cleanup. The analogy with Nokia—the giant that failed to choose a platform—is launched as a warning of what could happen.

Why does the United States support Intel?

Because it is considered a strategic piece rather than a business. The dominant argument here is that the State is interested in having its own semiconductor manufacturing on American soil and reducing dependence on Taiwan and TSMC. Hence the subsidies and grants to build plants, and hence the reading that Intel will ultimately become the country's white-label factory.

This public cushion is what sustains the case even when the accounts are poor. There is talk of a possible acquisition by Qualcomm, of alliances with other tech giants, and of a possible spinoff, scenarios that, if confirmed, would completely change the valuation. The plants, moreover, take years to bear fruit: the thesis demands patience.

The mirror of Meta and entry prices

Market sympathy comes and goes with the price. It is recalled that Meta fell below $100 and that no one wanted to touch it then, and that the narrative changed completely afterward. With Intel, according to this reading, the same will happen: when it rebounds, favorable opinions will start to rain down. It is the classic herd effect applied to semiconductors.

The prices declared by participants draw a wide range: entries at $17, very close to the minimum, and others at $21, $22, $30, $45, and even $49. One of the most repeated narratives is that of the one who bought in haste at $21, saw the price at $19, sold at $21.05, and missed the move. Declared gains range from +12% to +150%, with a +50% that fell short and a day of +8% in the heat of a possible spinoff. There are also those who recall having put in only a few coins and who got off too early.

AI as an excuse: NPU and Gaudi 3 accelerators

The attempt to jump on the artificial intelligence train was made before and went badly, according to the most technical analysis: there was a graphics project that did not work and nothing peine because the desktop and server processor business remained comfortable. Now the situation is reversed. In servers, they sell a worse product than the competition for more money, and in desktops, lost ground is notable.

The different trump card is the Gaudi 3 accelerators, chosen by IBM for AI workloads. There remains one unanswered question: how Intel will be affected by the bursting of the AI bubble when it has little presence in the sector. And another, more uncomfortable, the usual one: whether it is worth what it manufactures or what the United States needs it to manufacture.