Elon Musk has spent two decades aggressively building components in-house across Tesla, SpaceX, and xAI, yet he still purchases critical technology from a small handful of suppliers. These companies have survived what analysts call the "Elon Test": they tried to build what these suppliers make, and failed. Understanding which companies pass this test reveals the true limits of vertical integration in advanced manufacturing and AI chip production.

What Is the "Elon Test" and Why Does It Matter?

Musk's empire operates on a simple principle: if a component can be built cheaper, faster, or more controllably in-house, it gets brought in-house. Over twenty years, this instinct has swallowed batteries, motors, chips, software, seats, castings, rocket engines, avionics, launch towers, and entire data centers. The suppliers still standing today aren't standing because Musk is loyal to them. They're standing because his engineers tried to replicate what they do, and couldn't.

The companies that survive tend to operate in domains governed by unforgiving physics, decades of accumulated process knowledge, brutal capital intensity, and ecosystem lock-in. These are competitive moats that cannot be dug across in a few quarters, no matter how many engineers you deploy. Identifying which suppliers have passed the Elon Test reveals where the true structural advantages lie in advanced manufacturing and AI infrastructure.

Which Companies Has Musk Tried and Failed to Replace?

The clearest example is semiconductor fabrication. Musk designs chips, but he does not fabricate them. Every custom Tesla chip, including the Dojo D1 training processor, was manufactured by Taiwan Semiconductor Manufacturing Company (TSMC). The D1 is a 50-billion-transistor die built on TSMC's 7-nanometer process, delivering 362 trillion floating-point operations per second per tile. Even Terafab, Musk's roughly $25 billion venture to fabricate chips in-house, proves the point: Intel announced in April 2026 that it had joined the effort, targeting one terawatt of AI compute per year at Giga Texas. Intel's own CEO framed the partnership as needing Musk's help, not the other way around.

Behind TSMC sits an even purer monopoly: ASML Holding N.V., the only company on Earth that manufactures the extreme-ultraviolet lithography machines required for advanced chip nodes. Musk cannot build a leading-edge chip without TSMC, and TSMC cannot build one without ASML. These represent two irreplaceable choke points with no substitutes.

Nvidia presents a different kind of moat. Musk is one of the most aggressive custom-silicon builders in technology, designing Dojo, FSD inference chips, AI5, and AI6 processors. Yet when he needed to train Grok, he purchased roughly 200,000 Nvidia GPUs for Colossus, xAI's 100,000-GPU cluster that was built in 122 days and later doubled to 200,000 GPUs. When SpaceX unveiled its orbital data-center satellites, Musk named Nvidia as the supplier for the first wave. The moat isn't the silicon itself; it's CUDA, the software layer that the AI research world has built on for nearly two decades. Switching means rewriting years of code, so Musk, of all people, keeps reaching for his checkbook instead.

How to Identify Suppliers With Genuine Competitive Advantages

• Fabrication Expertise: Companies that control extreme-ultraviolet lithography or advanced semiconductor manufacturing processes have moats that cannot be replicated quickly, regardless of capital investment or engineering talent.
• Software Ecosystem Lock-in: Suppliers whose technology is embedded in years of research code and developer workflows create switching costs so high that even aggressive in-sourcers like Musk choose to buy rather than rebuild.
• Specialized Manufacturing Knowledge: Suppliers that have accumulated decades of process knowledge in niche areas, such as silicon-carbide power chips or precision casting, retain advantages that resist rapid replication.
• Capital Intensity and Physics Constraints: Suppliers operating in domains governed by extreme physics or requiring billions in upfront capital investment create natural barriers that time and engineering alone cannot overcome.

Tesla's battery strategy illustrates the narrowing nature of some moats. Tesla manufactures its own 4680 battery cells and has made them its lowest-cost cell per kilowatt-hour by late 2025, undercutting external suppliers. Yet Tesla's best-selling Model 3 and Model Y still run on cells from outside suppliers, and Tesla's finance chief has called batteries the company's "biggest constraint globally." In July 2025, one of those suppliers announced a multibillion-dollar deal to supply lithium iron phosphate (LFP) battery cells starting in 2027, with U.S. regulatory filings in March 2026 confirming Tesla as the buyer, even as Tesla races to replace exactly that kind of contract with its own cells. This represents the weakest moat on the list: these suppliers survive because Tesla needs volume today, not because Tesla cannot eventually build the product itself.

Specialist survivors include power-chip manufacturers and precision-casting companies. Tesla's Model 3 was the first mass-market electric vehicle to use a full silicon-carbide traction inverter, and silicon-carbide remains brutally hard to manufacture, historically costing ten times more than plain silicon. Even after Tesla said in 2023 it would cut silicon-carbide content by 75 percent, the company said it would keep buying die from multiple outside manufacturers. Musk's most theatrical manufacturing flex involves casting a car's entire rear underbody in one shot. The giant press that does this is built by a single Italian manufacturer, majority-owned by a Hong Kong-listed parent, and is the only die-caster, by Musk's own account, willing to take the risk of building a machine to his specifications.

What Are the Limits of This Framework?

Surviving the Elon Test today does not guarantee survival forever. Tesla's in-house battery cells represent the weakest link precisely because the moat is built on "Musk needs volume right now" rather than on physics or ecosystem lock-in. As Tesla's production scales, this moat erodes. The companies that look unkillable today are exactly the names a founder like Musk tends to fixate on next.

Additionally, these are real businesses with real customers outside Musk's empire. TSMC, ASML, and Nvidia would matter to the world even if Musk vanished tomorrow. Their exposure to Musk's ventures is a layer on top of already-large franchises, not the whole story. Investors should treat position sizing accordingly, recognizing that these suppliers derive value from far broader customer bases than just Musk's companies.

The pattern reveals a fundamental truth about advanced manufacturing and AI infrastructure: the companies that survive aggressive in-sourcing are those operating at the intersection of extreme physics, accumulated expertise, and ecosystem effects. These are the genuine moats in technology, and they remain remarkably durable even against the most determined vertical integrators in modern capitalism.