ASML CEO: Musk's Terafab Is 'Very Serious,' But Supply Is the Real Bottleneck

That relationship deepened in early June 2026 when Musk virtually attended ASML’s closed-door annual technology conference for a fireside chat with Fouquet. ASML characterized Terafab as a "serious endeavor" and positioned itself as a collaborator . Musk used the forum to outline his vision for a vertically integrated chip plant that would rely almost exclusively on ASML's tools to hit its ambitious targets .

Fouquet has since indicated that the first chips from Terafab could be produced within months of equipment installation, with Intel reportedly joining as a manufacturing partner . However, the path from equipment order to chip output is exactly where the supply warning bites.

Inside the Terafab Megaproject

Announced on March 21, 2026, Terafab is a joint venture among Tesla, SpaceX, and xAI designed to build the largest semiconductor manufacturing facility in history . The numbers are staggering:

• Target output: More than 1 terawatt (1 trillion watts) of AI compute capacity per year—nearly double the entire current US semiconductor output .
• Total investment: Initial phase estimated at $55 billion, with total costs potentially reaching $119–122 billion if all phases are completed .
• Location: Grimes County, Texas, near Austin .
• Process node: Targeting 2nm chips for AI, robotics, Tesla vehicles, and SpaceX applications .
• Footprint: Approximately 100 million square feet, roughly 10 times the size of Tesla's Giga Texas .
• Intel partnership: Intel formally joined the effort in April 2026, providing manufacturing expertise that Tesla and SpaceX lack internally .

Musk has stated that roughly 80% of Terafab's output will be destined for space-based applications, powering AI data centers in orbit via SpaceX's Starlink network, while the remaining 20% will serve terrestrial needs including Tesla's autonomous vehicles and Optimus humanoid robots .

The Broader AI Supply Crunch

Fouquet's Terafab-specific warning sits inside a much larger alarm he has been sounding for months. In a rare interview with Reuters on May 20, 2026, he said the global semiconductor market will remain "supply-limited for quite a while" because "demand on AI is coming so strongly" . He projected the chip market could reach $1.5 trillion by 2030 and warned of sporadic bottlenecks across the entire supply chain .

The constraint is structural, not cyclical. ASML is the only company in the world that makes extreme ultraviolet (EUV) lithography machines—the multi-million-dollar tools required to fabricate chips at 5nm, 3nm, and 2nm process nodes . Without these machines, cutting-edge chip production is impossible. And ASML can only build so many of them each year.

Compounding the pressure: Terafab is not the only mega-project competing for ASML's limited output. Tata Electronics is pursuing an $11 billion semiconductor fab in India that also requires High-NA EUV tools . Meanwhile, existing giants like TSMC, Samsung, and Intel continue to expand their own capacity. The result is a queue of well-funded customers stretching ASML's production capacity into the 2030s .

"Limited production of these machines could cause long delays and higher prices through 2030," one analysis noted, citing Fouquet's remarks .

ASML Stock: A Gatekeeper With a Stretched Valuation

The Terafab narrative has been a powerful catalyst for ASML shares, which closed around $1,804 on June 16, 2026 . The stock has surged roughly 35–75% year-to-date in 2026, depending on the measurement window, driven by AI infrastructure spending and a record order backlog of approximately $45 billion .

Wall Street analysts broadly rate ASML as a Buy, but the consensus suggests the easy money may already be priced in:

• Median price target: Around $1,997, implying roughly 10–11% upside from mid-June levels .
• Bear-case target: Approximately $1,581, suggesting a modest downside risk .
• Trailing P/E ratio: Approximately 62x, which is elevated even by semiconductor industry standards .
• Consensus rating: Strong Buy, but some analysts have recently trimmed estimates. Erste Group Bank cut its FY2026 EPS forecast for ASML to $36.88 from $37.35 in mid-June .
• Key risks: Export controls on China, where ASML's revenue is expected to decline, and the execution risk of scaling production fast enough to meet Terafab-level demand .

Some commentators have been blunt: "Elon Musk Needs ASML for Terafab. You Don't Need ASML Stock in Your Portfolio" . The argument is not that ASML is a bad company—it is arguably the most indispensable hardware supplier in the AI ecosystem—but that the current valuation already reflects years of hypergrowth, leaving little margin for error.

What It All Means

The Terafab-ASML dynamic reveals three core truths about the current state of the semiconductor industry.

ASML is the ultimate gatekeeper. Terafab cannot produce a single advanced chip without ASML's High-NA EUV lithography systems. This gives ASML extraordinary pricing power, order visibility, and strategic leverage that will likely persist through at least 2030 . Every dollar Musk commits to Terafab ultimately flows through ASML's order book.

Supply, not demand, is the binding constraint. The AI boom has created essentially unlimited demand for advanced chips. The bottleneck is not whether someone wants to buy—it is whether the equipment to make them can be built fast enough. Fouquet's repeated warnings suggest that even ASML, with its monopolistic position, is racing to keep up .

Investor sentiment is split. The long-term thesis is compelling: a monopolistic supplier at the heart of a multi-trillion-dollar AI buildout. But the near-term picture is murkier. A 62x P/E leaves ASML vulnerable to any supply-chain stumble, export-control escalation, or sector rotation. The Terafab opportunity reinforces the bull case; the stretched valuation keeps the bears cautious .

Fouquet's message, distilled: Musk is serious, the project is real, and ASML will benefit enormously—provided it can build machines fast enough. In the AI chip race, the real bottleneck isn't ambition. It's lithography.