Elon’s Missing Pieces

“Diligence is the mother of good luck.” – Benjamin Franklin

Elon Musk doesn’t simply run separate companies anymore. He runs an interlocking machine.

You see, xAI folded into SpaceX earlier this year. Then, SpaceX went public in one of the largest offerings in market history. And now, many are dropping hints that Tesla could be next – that the automaker and the rocket company might eventually become a single entity.

I want to be careful here… because there isn’t an official merger agreement in hand. This is my read of the evidence. But the evidence, in my opinion, points to a coming merger.

In fact, SpaceX President Gwynne Shotwell has said publicly that she sees “no question” there are synergies between the two companies. And Wall Street analyst Dan Ives puts the odds of a merge above 80% and has called it Musk’s “holy grail” move.

And the investors who see this coming could profit big-time.

But if you’re trying to play this by buying SpaceX or Tesla directly, you’re already too late for the kind of gain that changes a portfolio. SpaceX and Tesla are each worth well over $1 trillion. A company that size can double and still not move your net worth the way a $10,000 stake in the right small company can.

But when it looks like giants might make major moves, I look behind the scenes – to the companies filling gaps the larger firms can’t fill themselves. A trillion-dollar company can’t build everything in-house. So it leans on outside suppliers for the pieces that take years of focused, patient work to get right, and those suppliers are where I go looking.

They’re the ones already feeling the weight of an empire that intends to build the largest chip factory, the largest supercomputer, and the largest satellite network on Earth – all at once. That buildout is happening whether or not the two companies ever become one.

But a merger would do something else, and it’s the part most investors will miss until it’s too late to act on. The moment that announcement crosses the wire, every analyst on Wall Street will start mapping the new company’s supply chain – and names that have gone unnoticed for years could get discovered all at once.

That’s exactly what I went looking for here: the suppliers filling a critical gap in Musk’s empire that he can’t operate without.

And I found three companies that tie his businesses together. Let me walk you through them.

Missing Piece No. 1: The Battery – Amprius Technologies

Start with the biggest physical constraint on Musk’s AI ambitions: power in space. SpaceX has filed with the Federal Communications Commission to eventually launch as many as a million AI-capable satellites – full data-center racks in orbit rather than the simple internet relays that make up today’s Starlink network. Each one would need dramatically more power than a Starlink unit, and every extra pound launched into orbit costs real money. SpaceX’s own batteries, built for lighter Starlink duty, aren’t the right tool for this heavier job.

That’s the opening for Amprius Technologies (NYSE: AMPX), a Fremont, California-based battery maker I’ve been tracking for its silicon-anode technology. Instead of the graphite anode found in a conventional lithium-ion cell, Amprius uses silicon, which packs in roughly twice the energy at the same weight.

Amprius’ cells already powered the Zephyr aircraft, a solar-powered plane that flew continuously in the stratosphere for a record 67 days. The company also has presented its battery performance data at NASA’s Aerospace Battery Workshop, and an independent lab has validated its cells against the U.S. military’s nail-penetration safety standard.

Amprius’ silicon-anode cells are also becoming standard equipment for U.S. national security. The Defense Innovation Unit has now put $14.8 million behind the company’s drone batteries, and Amprius counts L3Harris Technologies among its customers.

The business is scaling fast. Second quarter revenue more than doubled year over year to $34 million, and gross margin expanded to 27% from 20% the quarter before. Management also raised full-year revenue guidance twice this year, most recently to at least $140 million.

Plus, in its second quarter call, the company announced a multi-year supply deal with Barcelona-based Stark Future worth more than $100 million. It also has a $24 million order from a European drone maker. Amprius ended the quarter with $74.5 million in cash and no debt.

The company also just extended its Fremont manufacturing lease through 2034.

Now, in May, a short seller published a report questioning some of Amprius’ manufacturing partnerships. Management hasn’t restated any financials, and the business has beaten its own raised guidance twice since the report came out, but it’s a live controversy worth knowing about.

Overall, Amprius has already proven its technology works in the field, is growing revenue, and now counts the Pentagon’s own innovation arm as a customer chasing supply security as much as performance. That’s a rare combination to find in a company still worth less than $2 billion.

Recommendation: Buy Amprius Technologies (NYSE: AMPX) up to $15. While this should be considered a long-term position, some Members may prefer to use a 25% trailing stop to protect their principal and profits.

Missing Piece No. 2: The Testing Gatekeeper – Aehr Test Systems

Terafab is the chip-manufacturing megaproject Musk announced in March 2026. It’s a joint venture between Tesla and SpaceX, built in Grimes County, Texas. It’s designed to produce chips for Tesla’s self-driving systems and Optimus robots, SpaceX’s satellites, and xAI’s AI models – all under one roof.

The first phase alone carries a $16.8 billion price tag, and county filings show the full multi-phase buildout could run as high as $119 billion. At more than 100 million square feet, Musk has said it will be the largest building on Earth, though production isn’t expected to start until late 2027.

And every chip that comes off Musk’s planned Terafab factory in Texas has to survive one thing before it ever reaches a car, a robot, or a satellite: burn-in testing. That means running the chip at full power and high heat to identify any potential weakness.

A chip that fails in your laptop is an inconvenience. A chip that fails in a self-driving car, or a data center a quarter-million miles from a repair technician, is not.

Aehr Test Systems (Nasdaq: AEHR), headquartered in Fremont just a few miles from Tesla’s own factory, is one of the only companies in the world that can burn-in test an entire silicon wafer at once.

Rather than testing them one at a time after they’re cut apart, Aehr can test thousands of chips in a single pass.

Two years ago, Aehr was a niche supplier that depended almost entirely on one customer base: chipmakers using silicon-carbide for electric vehicles. Today, AI processors make up the bulk of its revenue instead.

It quickly transitioned from a single-industry supplier to a critical player in the AI buildout. And the numbers indicate the shift is paying off.

Bookings hit $60.7 million in Aehr’s fiscal fourth quarter, up significantly from $11.1 million a year earlier. Backlog stands at $100.6 million on an effective basis, and management projects fiscal 2027 revenue of $130 million to $150 million.

Plus, Aehr’s lead AI customer is moving all of its burn-in testing to Aehr’s wafer-level systems, and the company recently completed benchmark testing with a second major AI chip supplier that’s now moving toward production orders of its own.

Aehr’s business model is built to compound. Customers buy the testing systems themselves as one-time capital purchases – and typically buys more of them over time as its production ramps up.

And AI processors are also just one part of a broader growth story. Aehr has completed a test of its equipment with a leading flash-memory chipmaker. It sees high-bandwidth memory (HBM) – a faster, more expensive type of memory chip packed tightly next to the processor in AI servers – as a major opportunity ahead.

One analyst estimates a single HBM customer win could be worth hundreds of millions of dollars in equipment sales alone, before counting the repeat adapter revenue that would follow. The company is also moving into testing the optical components that shuttle data at high speed between the thousands of processors inside a modern AI data center.

None of this is a signed, material contract yet, but it points to a company whose potential customer base keeps widening rather than one riding on a single relationship.

There’s also no signed deal between Aehr and Terafab yet. But Aehr doesn’t need one to win. Every company building AI data centers needs its chips tested, whether it’s Musk’s or a rival’s, and Aehr holds patents on wafer-level burn-in technology it’s actively defending.

Now, the stock has climbed significantly over the past year, and a large share of the growth ahead depends on just a couple of large AI customers ramping on schedule. Any slip in their ordering pace could hit the numbers.

Aehr has already reinvented itself once, from a niche electric vehicle parts supplier into the company AI chipmakers rely on. With record bookings, a positive revenue outlook, and new markets in memory and optical components still ahead of it, this is a small company with several separate paths to keep growing.

Recommendation: Buy Aehr Test Systems (Nasdaq: AEHR) up to $120. While this should be considered a long-term position, some Members may prefer to use a 25% trailing stop to protect their principal and profits.

Missing Piece No. 3: The Power Chip – Navitas Semiconductor

Every AI chip needs a second, less glamorous chip sitting right beside it. Its job is to take raw electricity and deliver it in exact, controlled doses, thousands of times a second. Too much power, and the processor fries; too little, and it shuts down.

Nowhere does that problem matter more right now than at xAI’s Colossus supercomputer in Memphis, which already draws roughly 380 megawatts of power to train Grok and which Musk’s own team has said will scale toward 2 gigawatts and beyond a million graphics processing units.

And every one of those watts has to pass through a power-delivery chip first. For 50 years, that chip was built from a material that wastes energy under loads like this, and the industry has been racing toward new materials that don’t.

Navitas Semiconductor (Nasdaq: NVTS) builds exactly those chips, using gallium nitride and silicon carbide. And it holds more than 300 patents on the technology. In May 2026, Nvidia and Navitas unveiled a landmark collaboration: an 800-volt power delivery board built on Navitas’ chips, designed specifically for next-generation AI data center racks. Nvidia showcased the technology at the Computex conference in Taipei this year.

Navitas operates as a fabless chip designer. It owns the intellectual property and hands actual manufacturing off. That structure keeps Navitas capital-light, since it doesn’t need to fund its own multibillion-dollar factory to scale production. Its competitive position rests on more than 300 patents covering how its gallium-nitride and silicon-carbide chips are designed and packaged, protecting a technology position that’s difficult for a rival to simply replicate.

The business itself is in the middle of a real transformation. It’s shifting away from low-margin mobile and consumer chips… and toward the same high-power AI, data center, and grid markets Musk’s empire depends on.

Management calls it “Navitas 2.0.”

High-power markets – AI data centers, grid infrastructure, industrial electrification – grew more than 50% year over year in the second quarter. Management expects AI infrastructure alone to make up more than a third of total sales by the fourth quarter of this year.

Guidance for the third quarter calls for $13.5 million in revenue, 28% sequential growth, and a return to year-over-year gains.

Profitability is still a work in progress for Navitas, but what matters more at this stage is the balance sheet. It raised $373 million in fresh equity this year, pushing cash to $557 million. With zero debt, the company has years of runway to fund the AI pivot without needing to raise money again on unfavorable terms.

Now, Navitas is currently defending a patent-infringement lawsuit filed by a rival, even as it pursues its own infringement claim against another company. Navitas disputes the claims and says it will defend its products, but litigation like this takes time and adds uncertainty.

There is tremendous opportunity with this company, however. Navitas doesn’t need a signed contract with Tesla or SpaceX specifically to benefit here. It only needs to keep winning business in the same AI data center architecture Musk’s own compute buildout already depends on.

Recommendation: Buy Navitas Semiconductor (Nasdaq: NVTS) up to $20. While this should be considered a long-term position, some Members may prefer to use a 25% trailing stop to protect their principal and profits.

Ready for the Spotlight

A merger between Tesla and SpaceX doesn’t mean Musk will be buying more batteries or testing more chips.

But it will shine the spotlight on his suppliers in a meaningful way.

For years these three companies have done essential work in near-total obscurity. But a merger announcement of this scale will drag the firms into the center of the most-watched story in the market.

The businesses don’t change. The number of people who know they exist will change enormously.

And that is precisely why you want to own them before a merger announcement hits.

I believe each of these companies could generate massive gains if Musk’s empire keeps expanding the way I expect. There are no guarantees, and never invest more in any one position than you can comfortably afford to lose.

But after 50 years of watching this pattern repeat, I can tell you it almost never changes. The giant gets the headlines. The suppliers get the returns.