The speaker discusses the potential merger between Tesla and SpaceX, noting that while the team has suggested it could be announced this year (late September), there isn't much time left, but it will "definitely" be proposed to investors at some point.
The biggest obstacle to the merger is whether enough current Tesla shareholders deeply understand SpaceX's potential. The speaker's theory is that the more Tesla investors attempt to value SpaceX, the more they will be in strong support of a merger. This involves understanding the opportunities in SpaceX's AI compute business, AI models business, Starlink (including increased bandwidth per launch on Starship and Direct to Cell), the launch business, and StarMind.
Elon and Tesla, the speaker suggests, will likely only propose a merger if they believe there's a high probability of shareholder approval, requiring confidence that Tesla shareholders recognize it as a "shockingly good deal," even if SpaceX represents the majority of the combined entity. The speaker refers to their "20 year Tesla and SpaceX valuation model" which includes a dedicated sheet for this merger scenario, detailing company valuations and fair ratios.
Elon's long-term goal is Mars, with "stops along the way" including global broadband connectivity, the "Dyson swarm," and the "Neocloud business" (referring to xAI). Elon believes the AI leader will be the company with the most computing capacity at the cheapest cost, with compute being the limiting factor.
The speaker provides hypothetical figures for SpaceX's AI compute business:
* In Q4 this year (heard about early 2027), SpaceX could bring online 1+ gigawatts of AI compute, monetizing most of it and generating approximately $50 billion annually in recurring revenue. This would be "half a Tesla worth of annual revenue in a single quarter."
* This estimate is based on Musk's public posts about 660,000 GB300s.
* If fully monetized, this could yield $50+ billion/year recurring revenue in a single quarter.
* If this pace continues for the entire year in 2027, it could add another $200+ billion, potentially exiting 2027 with more than $250 billion in annual revenue from the AI compute business alone (compared to Tesla's roughly $100 billion today).
* The AI compute business could "meaningfully outscale Tesla's autonomous vehicle business."
The speaker emphasizes that SpaceX AI doesn't need to reach a certain valuation, but rather needs to prove its potential for massive revenue and profit growth in the coming quarters and years to investors. The speaker asserts that anyone who models out SpaceX will inevitably become a supporter of the merger.
The speaker's largest holdings, and "only holdings," are Tesla and SpaceX, and they eagerly anticipate the merger. While gaining enough Tesla shareholder support is currently a "hard challenge," it's considered a matter of time. The speaker believes it will likely require "at least a few additional quarters of SpaceX absolutely hammering it," particularly with explosive growth in AI compute infrastructure and AI models, before enough Tesla investors recognize the enormous opportunity. Elon Musk has enough voting power in SpaceX, so support from SpaceX is certain; the hurdle lies solely with Tesla shareholders.
The podcast also includes a sponsorship message for AG1, encouraging listeners to visit drinkAG1.com/smr for a free welcome kit.