The Greatest Merger In History: What the Market Is Missing

For years people argued about whether Elon would ever merge Tesla and SpaceX. That argument is basically over. Two thirds of the stack already sit together. Tesla is the last piece still outside. The obvious question is why the most valuable company is the one left out. The answer starts with what he did on June 17, 2026.

On that day he exercised options from his old 2018 CEO award: about 304 million shares at $23 apiece. He converted them into restricted stock with immediate voting rights. That single move lifted his Tesla voting power from roughly 13% to about 20%. He now sits on around 700 million shares, roughly a fifth of the company. To pay for the exercise he sold 17 million shares for about $7 billion. Options are not free. You owe the strike. Then you owe the tax. The sale covered both. The shares themselves stay locked until January 2028.

Look at the calendar. SpaceX started trading as a public company on June 12. Five days later he did this. Motley Fool ran a headline asking whether a SpaceX merger could be next. That timing is not an accident.

The June exercise is only one step in a longer march. In November 2025 Tesla shareholders approved a new pay package with over 75% support. Twelve tranches. Around 424 million shares. Worth up to roughly a trillion dollars over ten years if the milestones hit. Those milestones are extreme: an $8.5 trillion market-cap goal, 20 million vehicles, a million Optimus robots in the world. As the tranches vest, his ownership keeps climbing toward about 25% voting control. That number is not random. In January 2024 he said out loud he wanted roughly 25% — enough to be influential, not so much that he cannot be overturned. The June move looks like the first brick of a Tesla-SpaceX combination.

Voting matters because of how both sides are stacked. On the SpaceX side Musk controls something like 82% to 85% of the vote through supervoting shares. In any Tesla-SpaceX deal he sits on both sides of the table: lock-it-down control on SpaceX, approaching 25% on Tesla. He is functionally negotiating with himself. More votes help pass a deal. They do not make the deal lawsuit-proof. When a controlling insider sits on both sides, the legal tests get stricter, not looser. Winning the shareholder vote and surviving court are two different problems. That is exactly why xAI and SpaceX went first and Tesla got left for last.

We have seen this movie. In 2016 Tesla bought SolarCity for about $2.6 billion. Elon was on both sides then too. That deal went to a full trial in Delaware and survived. A judge ruled it entirely fair in 2022. Affirmed in 2023. It only survived because there was a genuinely independent board, a minority shareholder vote that excluded Musk's shares, and a price that could be defended. That is a hard gauntlet.

The xAI-SpaceX merger walked around that gauntlet by staying private and incorporating in Nevada, where the legal standard is friendlier to insiders. In Nevada you essentially have to prove known misconduct. In Delaware you have to prove the deal was entirely fair — a much higher bar for the company. Tesla is public, heavily owned by outside shareholders, and sued constantly. Those shareholders are already in court over Tesla's $2 billion investment into xAI from January 2026. So the order came down to friction. The pieces that could merge quietly went first. The public, lawsuit-filled, minority-owned company got saved for last.

A cleaner investor table was only half of what that merger handed him. The other half was a balance sheet. In June 2026 SpaceX priced its first investment-grade bond. Demand was staggering. They targeted $20 billion, upsized to $25 billion, and the order book came in around $90 billion — oversubscribed about three and a half times. Investors wanted to lend almost four times what it asked for. Five tranches. Coupons from about 5.35% up to 6.65%. Maturities stretching to 2056. Roughly $1.5 billion a year in interest. Lead banks: Bank of America, Citi, Goldman, JPMorgan, Morgan Stanley. Ratings came back investment grade across the board.

What is the money for? It refinances expensive debt left over from the 2022 Twitter acquisition and costly xAI loans from 2025, plus general corporate use. SpaceX also disclosed a cash pile of about $100 billion. Follow the structure: the Musk ecosystem's debt now routes through SpaceX, and the public bond market refinances it at cheap investment-grade rates. SpaceX became the financial foundation. Any Tesla combination will lean on that foundation.

Then SpaceX turned inherited compute into a landlord business. In June 2026 it signed Reflection AI to rent compute at $150 million a month — about $6.3 billion through the end of 2029. Reflection is an open-source AI startup founded by ex-DeepMind people, Nvidia-backed, pitched as an American answer to DeepSeek. That sits on top of Anthropic and Google deals worth tens of billions. Colossus was xAI's training supercomputer, the cluster that trained Grok. After the merger SpaceX owns Colossus and flipped it into a commercial multi-tenant cloud — what the industry calls a neo-cloud. Google is in at roughly $920 million a month. Anthropic signed for around $45 billion over multiple years. Committed compute revenue from outside customers now tops $80 billion through 2029. That backlog helps justify SpaceX's roughly $1.8 trillion valuation and helps service the new bond debt. It also makes SpaceX a credible AI infrastructure partner for what Tesla needs next: autonomy and Optimus.

The third piece is silicon. Terafab is the planned shared chip megafab announced in March 2026 near Austin, reported as a joint venture across Tesla, SpaceX, and xAI. That it has to be a three-way JV already tells you the corporate walls are getting in the way of building the thing. Cost estimates climbed from about $25 billion in March to as much as $120 billion all-in by May 2026. It is supposed to make custom chips for FSD, Cybertruck, and Optimus, plus radiation-hardened chips for data-center satellites. Tesla has been here before. Dojo was killed in August 2025. Musk said all paths converge on AI6 and called Dojo 2 an evolutionary dead end. The lead and about 20 staff walked. Months later the program was revived. Tesla targets small-batch AI5 by late 2026 and volume in 2027. Analysts are cooler. Morgan Stanley projects meaningful output closer to 2028 to 2030.

Meanwhile the pieces that already exist keep stacking. AI5 tapeout landed in April 2026. Tesla claims it runs inference roughly ten times cheaper than Nvidia hardware. Megapod is a trademark Tesla filed in June 2026 for a turnkey container-style AI data-center box, reportedly powered by Megapack batteries. Grok already ships inside millions of Tesla cars. It went live in the fleet in July 2025 and is the conversational brain inside Optimus Gen 3. Starmind, just trademarked, is SpaceX's orbital AI layer, with its first satellite unveiled in June 2026. Four companies on paper. Chips, models, and data centers already plugged into each other.

Step back. There are two kinds of companies in this story. Tesla acts on intelligence. It takes a trained model and spends compute to turn it into physical motion: FSD into steering, Optimus into hands, a robotaxi into a left turn through traffic. Every car and robot is an edge node turning a finished model into movement. Call that inference consumption. SpaceX and xAI make the intelligence. Colossus trains Grok. Starmind runs inference in orbit. Starlink beams it down. Starship drops the cost of putting compute anywhere. Call that inference production. One side makes the intelligence. The other spends it.

Right now those two costs sit in separate companies. Every time a unit of intelligence crosses from producer to consumer it gets taxed: IP licensing, related-party reviews, duplicate overhead, lawyers on both sides of every deal. Put them in one company and that friction goes to basically zero. One trained model gets amortized across the largest surface on Earth — robotaxis, millions of Optimus robots, satellites, and outside customers renting compute. The biggest cost in AI is training the frontier model. Spread that across more units than any company alive. One custom silicon line from AI5 into AI6 into cars, robots, and data centers. One energy stack with Megapack and solar powering the data centers making the intelligence and the fleet burning it. The Apple playbook taken to its extreme: chip, software, device. Except the device is a car, a humanoid, and a satellite, and the chip flies in orbit.

Inevitable can be true and still be a bad deal for one side of the table. Put these together and you get a roughly $4 trillion entity. Tesla's GAAP profit is positive but small relative to that whole. So we are in a tense moment between Tesla and SpaceX investors — many of whom own both — asking what a fair relative valuation is. The higher Tesla's relative value, the better for Tesla holders. The lower, the better for SpaceX holders. Given Elon's voting power and the trust he gets for execution, the price will likely be whatever he thinks is right. Then the questions become how that lands with Tesla investors and how it lands in court after the lawsuits.

This deal is getting done. Too many pieces between these companies benefit from becoming one entity. When it happens, it will be one of the most historic events in business history.

Check the video here.