Ron Baron Bets on Tesla at 5x and Data Centers in Space. What Would Prove Him Right?
The 83-year-old investor says he is fine with or without a Tesla-SpaceX merger. What he is really backing is AI-era power, autonomy and orbital infrastructure.
The 20-second read
1. A $30 billion gain, and now hundreds of billions more?
Ron Baron, founder of Baron Capital, made a striking forecast on CNBC on Nov. 6. He expects hundreds of billions of dollars in additional gains from Elon Musk-related investments over the next 10 years.
The confidence has a track record behind it. Baron Capital has held Tesla and SpaceX for a long time. By Baron's account, the two holdings have produced about $30 billion in cumulative gains for clients. The firm has previously disclosed positions of about $25 billion in SpaceX and about $5 billion in Tesla. Those are valuations at the time, not the original capital invested.
Asked about a merger of the two companies, Baron gave a simple answer. A merger would be fine, and so would no merger.
He said he has passed Musk his private views on the pros and cons, has not made them public, and will back the decision of an independent committee and Musk's management judgment.
What matters here is less whether a deal happens than where Baron locates the source of returns.
A merger is a separate question. Gains and losses for each shareholder would depend on the exchange ratio and premium offered to existing Tesla holders, how conflicts of interest are handled, and regulatory review. Technical synergies between the two companies do not by themselves make a merger good for shareholders.
2. In US AI, power may be a bigger constraint than chips
The figure Baron stressed during the interview was China's pace of power build-out.
He said China builds about 300 GW of generating capacity a year, while the US builds about 20 GW.
The direction is right, but the actual statistics are more complicated than his remarks.
| 2025 | What it counts | |
|---|---|---|
| China | About 452 GW | New renewable generating capacity |
| US | About 53 GW | All new utility-scale generating capacity |
Sources: China's National Energy Administration and the US Energy Information Administration. The scopes differ, so the two figures should not be read as a like-for-like comparison.
According to China's National Energy Administration, new renewable capacity in 2025 was about 452 GW. New utility-scale capacity in the US that year was about 53 GW, per the EIA. US plans for 2026 call for 86 GW of additions.
Capacity is not the same as actual generation. Solar output depends on weather and daylight hours, and data centers need stable, round-the-clock power.
Still, it is clear that China is expanding its power infrastructure much faster.
The US leads in advanced AI chips and software. But if it cannot secure power for data centers in time, the pace at which it can add computing capacity could be constrained.
3. SpaceX's Starmind sends the data center off the planet
That is why Baron is focused on SpaceX's Starmind.
Starmind is a concept to place AI computing hardware in orbit, powered by solar energy collected in space. SpaceX's published AI1 satellite design lists peak compute capacity of up to 250 kW and average compute capacity of 175 kW. Results are relayed over laser links to the Starlink network.
In space, there is no need to acquire large tracts of land or connect to a ground grid. Heat is shed through radiators in a vacuum, and a well-chosen orbit can extend the hours of available sunlight.
None of that, however, proves it is cheaper than a ground-based data center.
Space offers no air to carry heat away, so large radiator structures are needed. Radiation-induced chip errors, the difficulty of repairing hardware, satellite lifetimes, launch costs and data transmission delays all have to be solved.
SpaceX's biggest strengths are its large-scale launch capability through Starship and its existing Starlink production and communications infrastructure. The key question is whether it can use them to cut costs enough to compete with ground-based data centers.
Baron projected that space data centers would operate at meaningful scale from 2028 and grow sharply in the one to two years after that. That is an optimistic forecast about the speed of commercialization, not proven business results.
4. Tesla at 5x in 10 years: how far has the robotaxi come?
Baron said he expects another fivefold return on Tesla over the next 10 years.
He explained that during Tesla's big past run he did not try to time short-term highs and lows, and that holding for the long term produced his results.
This time there is also different evidence.
According to Tesla's website, the robotaxi service operates in limited areas of Austin, Dallas, Fort Worth, Houston, Miami, Orlando and Tampa.
Tesla's published safety report shows about 3.1 million miles of autonomous driving through Oct. 18, 2026. Safety monitors and operating methods may differ by market, a distinction worth keeping in mind.
The expansion of service is a positive. But valuing the robotaxi business takes different numbers.
How many hours a day does each vehicle run? How much fare revenue does it actually earn? How much is left after vehicle, insurance, maintenance and remote-support costs?
Miles driven are a measure of technical validation. Contribution profit per vehicle is a measure of business viability. They are not the same.
If Tesla's market value is assumed to be about $1.4 trillion, a fivefold share price would, other things equal, imply a valuation of about $7 trillion. Growth on that scale would likely require substantial cash flow not only from robotaxis but also from energy storage, autonomous driving software and the Optimus robot business.
The question the interview raises
The question is simple.
It is not whether Tesla and SpaceX merge. It is which company, 10 years from now, can supply the compute and power AI needs most economically, and return the results to shareholders.
Baron believes the answer lies with Musk's companies. For investors, the task is not to follow that belief but to watch which numbers, over time, prove or refute it.
- No. Depending on the exchange ratio, new share issuance, asset valuation, and the transfer of business risk, existing shareholders' value could be diluted.
- Whether business synergies turn into higher shareholder value has to be examined separately.
- That is unlikely. Ground infrastructure may be better for latency-sensitive services and large-scale data storage.
- Space computing is likely to grow first as a complement that takes on certain AI workloads.
- Expanding to more cities shows progress toward commercialization, but it does not prove profitability.
- You need to look at revenue per vehicle, service costs, and the payback period on vehicle investment together.
- Holding for the long term is different from holding unconditionally. If the business thesis is damaged, your judgment should change too.
- Keep checking progress in the technology and the business, and avoid concentrating your portfolio too heavily in one executive or company.
What to watch
Insight Times Editorial Desk
