What Are Wafer Starts?

Wafer starts are the fundamental metric of semiconductor manufacturing. A wafer is a thin slice of silicon, typically 300mm in diameter, on which chips are fabricated. A “wafer start” is the beginning of the manufacturing process for a batch of wafers. The number of wafer starts per month determines how many chips a facility can produce, and it is the primary measure of a chip factory’s capacity.

When Luke Lango says the Terafab will “double American chip production,” he is talking about dramatically increasing the number of wafer starts in the United States. Currently, 97 percent of high-end chips are made in Taiwan, primarily by Taiwan Semiconductor (TSMC). The Terafab, a 100-million-square-foot facility in Texas, is designed to shift a significant portion of that production to the U.S.

As we explain in our full review of the INI XPanse presentation, the Terafab is the integration point for the XPanse thesis. Here we focus on what wafer starts mean for investors and why the Terafab’s production capacity matters.

The Strategic Importance of Wafer Starts

The concentration of wafer starts in Taiwan is what Lango calls an “economic apocalypse” waiting to happen. Treasury Secretary Scott Bessent calls it “the single biggest threat to the world economy.” If Taiwan were blockaded, the global supply of high-end chips would be severely disrupted, affecting everything from smartphones to AI data centers to military hardware.

The U.S. government has recognized this vulnerability through the CHIPS Act, which provides funding to increase domestic chip manufacturing. The Terafab, announced on May 6 as a $25 billion initiative, represents what Lango calls a facility that “will singlehandedly by itself double American chip production.” Intel has partnered to “fast-track the project.”

The strategic importance is clear: if you control wafer starts, you control the chip supply chain. And if you control the chip supply chain, you have a significant advantage in the AI economy, where every data center, every robot, and every AI model depends on chips. For more on the national security angle, see our article on sources for Tesla and SpaceX.

The Terafab’s Production Scale

While Lango does not specify exact wafer start numbers for the Terafab, the scale he describes is significant. A 100-million-square-foot facility is enormous by any standard. For comparison, TSMC’s most advanced fabs are approximately 4-5 million square feet. The Terafab would be roughly 20-25 times larger than a single TSMC fab.

Lango says the Terafab “will singlehandedly by itself double American chip production.” If U.S. chip production is currently a fraction of global capacity, doubling it represents a meaningful shift in the global semiconductor landscape. The combination of Intel’s manufacturing expertise and Tesla’s capital and ambition makes this a credible, if ambitious, target.

The Terafab will produce the AI5 chip, which Jeff Brown (in the MAGI presentation at Brownstone Research) says is “40 times more powerful” than current chips. The connection between wafer starts and the AI economy is direct: more wafer starts mean more chips, and more chips mean more AI infrastructure capacity. For more on the MAGI thesis, see our MAGI review.

The Three-Pillar Integration

What makes the Terafab’s wafer starts different from a traditional chip factory is the integration with the other pillars of the XPanse thesis:

AI-driven manufacturing. The Terafab is “trained by xAI’s Grok.” AI can optimize chip design and manufacturing processes, potentially increasing yields (the percentage of functioning chips per wafer) and reducing defects. Higher yields mean more effective wafer starts, as a larger percentage of each wafer produces sellable chips. For more on this angle, see our article on AI model training.

Orbital data center power. The Terafab is “powered by SpaceX’s AI1 satellites.” If orbital data centers can provide the enormous compute power needed for AI-driven manufacturing optimization, the Terafab’s efficiency increases. The data center crisis on Earth (land, power, water constraints) makes orbital alternatives attractive. For more, see our article on AI compute capacity.

Robot labor. The Terafab is “ultimately staffed by Tesla’s Optimus robots.” A 100-million-square-foot facility running 24/7 requires enormous labor. If Optimus robots can perform manufacturing tasks, the Terafab can operate at full capacity without the labor constraints that limit traditional factories. For more, see our article on Optimus AI.

The Investment Angles

The wafer starts angle connects to several of Lango’s investment picks:

The Terafab chip supplier (paid pick #2) is what Lango calls “the ONLY U.S.-based pure play in this niche industry” that is “well-positioned to supply this linchpin tech for Terafab’s ramp-up.” This company supplies critical technology for the facility that will increase U.S. wafer starts. For more, see our article on Tesla chips.

The rare earth kingpin (paid pick #3) provides the raw materials needed for semiconductor manufacturing. Rare earth elements are used in chip fabrication equipment, magnets for manufacturing motors, and various electronic components. The mine 70 miles from the Terafab site creates a vertically integrated supply chain. For more, see our article on sources for Tesla and SpaceX.

For the overall investment framework, see our articles on growth potential and Innovation Investor.

The Speed Precedent

Musk’s ability to build at scale rapidly is what makes the Terafab’s wafer start timeline credible. Lango provides several examples:

  • The Tesla Shanghai gigafactory, which now produces 950,000 vehicles per year, was built in 168 days
  • The Colossus AI supercomputer was built in 19 days, a process Jensen Huang says would normally take four years
  • At age 12, Musk completed a six-month coding course in three days

If Musk applies this speed to the Terafab, with Intel’s manufacturing expertise, the timeline for increasing U.S. wafer starts becomes more credible. The Shanghai gigafactory going from permits to 950,000 annual vehicles in 168 days is a real data point. For more on Musk’s execution speed, see our article on Elon Musk’s next project.

Considerations

The wafer starts angle is grounded in real strategic importance. The Taiwan chip dependency is a genuine national security concern. The Terafab announcement is real. Intel’s partnership is documented. And the combination of AI-driven manufacturing, orbital data center power, and robot labor represents a genuinely novel approach to chip production.

What to consider: Doubling American chip production is an ambitious target. The specific wafer start numbers are not detailed in the presentation. The chip manufacturing industry is highly competitive, with established players like TSMC, Samsung, and Intel. And the integration of AI, orbital data centers, and robots into a chip factory is an unproven combination at this scale.

What makes the wafer starts angle worth considering is that it addresses a real strategic vulnerability with a real solution, backed by real partnerships and real government funding. The companies supplying the critical technology for the Terafab’s ramp-up are positioned to benefit from both the facility’s construction and the broader policy push for domestic chip manufacturing. For investors interested in the semiconductor supply chain, understanding wafer starts provides a concrete metric for evaluating the Terafab’s potential impact.

Ready to learn more? Click here to access Luke Lango’s full research.

This is not financial advice. Always do your own research before investing.