Intel shares slide after Musk confirms TSMC in Terafab talks

Main points from this report:

• Musk recognised TSMC was holding discussions about Terafab, calling them purely preliminary. He indicated TSMC's chips would probably complement, rather than supplant Intel's.
• Terafab currently lists Intel as its sole declared production partner, providing Intel with its highest-profile external contract chipmaking win.
• If TSMC operates its own factory at the site, Intel's unique billing would become a shared arrangement.
• The foundry unit generates modest external revenue, so the company's valuation depends heavily on future client wins.
• Key milestones to monitor are a binding TSMC contract and the debut of Intel's 14A design kit this month.

A brief remark from Elon Musk was enough to shake a chip company valued at over $600 billion. Here is why Intel is so exposed to the Austin foundry plan.

Intel's stock declined on Monday when Elon Musk acknowledged that TSMC had entered talks concerning Terafab, the Texas-based chipmaking initiative supported by his companies. Musk characterized the dialogue as preliminary, noting it might lead to an agreement, and forecast that any TSMC chips would complement, not replace, Intel's. Intel shares lost more than 4% in pre-market activity and roughly 2% during the morning session. TSMC's American depositary receipts gained about 1%, and its stock in Taiwan ended the day roughly 3% higher.

It might seem excessive that a social media post about a company valued at over $600 billion could trigger such a move. The reaction becomes clearer when the significance of Terafab to Intel, the economics of contract chip fabrication, and the amount of anticipated growth embedded in Intel's stock are considered.

The Terafab project and Intel's central place in it

In March, the approximately $25 billion Terafab scheme was unveiled in Austin, described as the biggest chip fabrication plant ever conceived. Its initial target is roughly 100,000 silicon wafers monthly, with a long-term goal of about 1 million, making chips for robotics, transport, and aerospace uses. Roughly $16.8 billion has allegedly been allocated to the initial stage.

Intel came on board in April as the exclusive identified production partner, assigning its 14A node, its most cutting-edge expected manufacturing technique, to the site. This elevated Terafab beyond a single client agreement for Intel. It offered the clearest demonstration that external firms were ready to place trust in Intel's fabrication units.

The foundry business model and the weight of winning clients

A foundry fabricates semiconductors that other entities have designed. TSMC is the leading instance: it makes chips for numerous global tech giants yet does not market processors under its own brand.

Intel has conventionally engineered and manufactured its own CPUs. As part of its revival plan, it is pursuing contracts to make chips for other firms. This is a costly objective. State-of-the-art plants demand tens of billions in investment, and every new manufacturing node involves huge expenditure prior to generating returns.

This explains the weight of customers. A foundry requires sufficient demand to operate its factories at capacity, recoup its outlay, and finance the following node. A substantial, dedicated client that occupies a facility from the outset is called an anchor customer. Securing one indicates to other chip architects that the foundry's technology merits confidence.

The figures reveal how nascent Intel's foundry effort is. In the second quarter, external clients brought in only roughly $290 million in revenue for Intel's foundry segment, which posted an overall loss of about $2.1 billion.

What a TSMC presence would alter

The Culpium newsletter, which broke the news of TSMC's interest on Friday, outlined a probable framework where TSMC owns and runs the plant, with Terafab serving as the anchor client. That arrangement would parallel TSMC's partnerships in Japan and Germany.

The arrangement's importance goes beyond just splitting the manufacturing work. If TSMC operates a unit at Terafab, Intel could find itself demoted from the single production ally to one among two, vying for capacity at the exact initiative that was supposed to highlight its contract manufacturing aspirations. Analysts have cautioned that if Intel steps back entirely or becomes merely one of several technology vendors, investors might reassess the leeway they have afforded Intel. Some have proposed that TSMC might operate a factory solely serving Tesla, SpaceX, and xAI.

The root of the stock's vulnerability

Intel's stock price has approximately trebled over the last twelve months, and the company reportedly trades at about 71 times projected earnings. Such a high multiple indicates that investors are largely paying for anticipated future profits rather than current earnings. A considerable portion of that expectation is tied to the foundry operation securing substantial external clients.

When worth depends on hypothetical victories, news casting doubt on any one of them can shift the share price even absent a concluded deal. The market is essentially adjusting the odds of an outcome rather than responding to a certain loss.

The Terafab update coincided with a day when analysts projected Intel had ceded more ground in the PC processor market during the third quarter, while competitor AMD advanced, imposing further strain.

Why the reaction might be too hasty

Musk framed the conversations as purely preliminary, with no deal finalized. His remark that TSMC's chips would probably complement Intel's indicates that Terafab's intended capacity might handle both. A venture of this scope might gain from dual suppliers, lowering the hazard of depending on a single producer.

Signals to track

The main signal is whether TSMC formalizes an agreement, and on what conditions. A transaction in which TSMC possesses and manages its own factory would generate more concern for Intel than one where it delivers a smaller volume of chips. Intel's scheduled release this month of the 14A process design kit, the utilities clients require to develop chips for the node, is another test of external firms' willingness to commit to Intel's plants.