Tesla Is The New Data Center King | SpaceX Merger Is Coming

The video explains how SpaceX’s record-breaking IPO—which made Elon Musk the world’s first trillionaire—is the catalyst for a likely SpaceX-Tesla combination, with Starlink’s profits funding massive AI compute ventures while Tesla’s real-world driving data becomes the ultimate moat. It highlights that Starlink is the true cash engine, generating over $11 billion in revenue, while SpaceX’s AI arm loses billions yet is rapidly scaling through xAI’s Colossus supercomputer and rented compute deals with competitors. Tesla already benefits from the AI boom through its Cortex supercomputer and Megapack battery sales, but the key structural problem is that the most lucrative AI-compute revenue flows to SpaceX and xAI, not Tesla. SpaceX president Gwynne Shotwell has publicly acknowledged a convergence between the companies, and analysts estimate an 80% chance of a merger, likely via SpaceX absorbing Tesla in a stock swap worth over $3.4 trillion. The decisive advantages are build speed—xAI built Colossus in 122 days while incumbents take years—and Tesla’s 10 billion miles of FSD driving data, which competitors cannot buy. Ultimately, the pieces are already in place, and the video argues that the convergence is not a speculative promise but a structural consolidation of Musk’s companies under one AI-driven vision.

• ▶ 0:00 SpaceX's record-breaking IPO made Elon Musk the world's first trillionaire, with the company immediately surpassing Tesla's market cap on day one.
• ▶ 1:00 SpaceX raised ~$75 billion by selling 555 million shares at $135, valuing the company around $1.8 trillion; the stock closed up ~19% on day one.
• ▶ 2:26 First public financials revealed Starlink as the real cash engine: $11.4B revenue (61% of total), $4B+ operating income, ~10M subscribers—while rocket launches actually lost money.
• ▶ 4:19 SpaceX's AI division (Space XAI) is a massive, money-losing bet: ~$3B revenue but billions in operating losses, including nearly $2.5B lost in the most recent quarter—funded by Starlink's profits.
• ▶ 4:54 Tesla operates one of the largest private supercomputers in the U.S. at Giga Texas — Cortex — starting with ~50,000 Nvidia H100s, scaling toward 350,000 chips, and drawing up to 500 MW at full scale to train FSD and Optimus; Cortex 2 adds another 5M+ sq ft by 2026.
• ▶ 8:21 Tesla Energy sells Megapack grid-scale batteries that smooth power spikes for AI data centers; in 2025, Tesla sold ~$430M of Megapacks to xAI, and xAI’s Colossus supercomputer runs on 168 Megapacks providing ~150 MW of backup.
• ▶ 9:18 Tesla Energy deployed ~47 GWh of storage last year at higher margins than the car business, with data centers driving growing demand — so even without a merger, Tesla is already profiting from the AI infrastructure boom.
• ▶ 10:48 Compute is the core competitive advantage in AI: more compute means smarter models, and falling behind is nearly impossible to reverse.
• ▶ 11:46 AI data centers are heavy industry, not software—the real bottleneck is building, powering, and cooling sites fast, with costs reaching $38 billion per gigawatt.
• ▶ 13:36 The decisive edge is build speed: xAI’s Colossus went up in 122 days and is scaling to ~2 GW and 555,000 chips, beating incumbents who take 1.5–3 years per site.
• ▶ 16:44 The entire AI industry is hitting a physical ceiling of power, land, and water, so xAI is moving compute to orbit — an advantage only SpaceX can provide.
• ▶ 18:16 xAI turned a hardware-flawed Colossus site into a rental business, signing massive deals with competitors: Anthropic (~$1.25B/month) and Google (~$920M/month).
• ▶ 19:30 Grok has become a top-tier model, trained on ~100x more compute, while these rental deals push xAI to over $2 billion per month in revenue.
• ▶ 21:48 Musk's core commodity is "raw intelligence by the gigawatt" — massive AI compute sold as a utility-like service, projected to yield ~$26 billion per year once fully operational.
• ▶ 21:55 That annual revenue would exceed SpaceX's entire last-year revenue (rockets plus Starlink), with Wall Street reportedly calling the business "Elon Web Services."
• ▶ 22:07 The key problem: the compute revenue flows to SpaceX and xAI, not Tesla, creating a structural gap between the entity earning the AI-compute "rent" and Tesla, which actually needs the compute for its AI ambitions.
• ▶ 22:26 SpaceX president Gwynne Shotwell officially acknowledged Tesla-SpaceX convergence for the first time, speaking publicly on the day SpaceX went public.
• ▶ 22:41 Shotwell explicitly said there is "a convergence of what all of these companies are trying to accomplish in the future" and suggested a Tesla deal might "make Elon's life a little easier."
• ▶ 22:58 She gave a measured statement—not announcing a deal—while emphasizing SpaceX's current focus remains on its core operations like rockets, ISS supply, and internet access.
• ▶ 23:13 The SpaceX IPO is the critical catalyst that makes a SpaceX-Tesla combination realistic, giving SpaceX public stock as acquisition currency.
• ▶ 23:29 The likely direction is SpaceX absorbing Tesla via a stock swap, not the other way around.
• ▶ 23:44 Wall Street analyst Dan Ives puts the odds of the combination at about 80%, with a combined valuation exceeding $3.4 trillion.
• ▶ 24:08 Musk has a long history of missing deadlines, but most of what is discussed is not a forward-looking promise—it is already built, with data centers rented out, a chip factory under construction, and xAI already operating inside SpaceX.
• ▶ 24:20 Viewers don’t need to believe a single forward-looking statement to recognize that the foundational pieces are already in place.
• ▶ 24:25 This is a recurring consolidation pattern: SolarCity was folded into Tesla in 2016, xAI was started in 2023 and folded into SpaceX by February, and SpaceX is now public—pulling Musk’s companies closer together under his personal control, with Tesla as the central piece.
• ▶ 24:46 A Tesla-SpaceX merger hasn't happened yet; structuring and regulatory clearance will take time, so it likely plays out over the next year or two rather than immediately, though pieces are lining up.
• ▶ 25:07 Combining the two would put SpaceX's rapid data center build speed and compute capacity on Tesla's balance sheet, alongside Tesla's own Cortex supercomputers and Megapacks.
• ▶ 25:35 The biggest asset would be data: Tesla's FSD fleet has passed 10 billion miles driven, adding roughly 29 million miles per day, with every car recording rare driving moments to train the AI.
• ▶ 25:52 Tesla’s true moat is the data itself—competitors can rent chips and data centers, but they cannot buy 10 billion miles of real-world driving data.
• ▶ 25:59 The advantage is a self-perpetuating flywheel: more cars → more driving miles → better AI → more sales, and the loop speeds up as it grows.
• ▶ 26:23 Once AI can drive with no one inside, the same loop expands beyond cars to produce robo-taxis and robots, each generating new data that feeds the flywheel.
• ▶ 26:34 The robo taxi is Tesla's first payoff product, acting as the commercial bridge between its AI/autonomy work and the core vehicle business.
• ▶ 26:44 Robo taxi rides are priced at ~$1.00 per mile versus Uber's $1.50–$2.00, with costs expected to keep falling because removing the driver eliminates nearly the entire operating expense of ride-hailing.
• ▶ 26:57 Tesla plans to produce more Cyber Cabs than all its other vehicles combined, using a purpose-built two-seat car designed to drive itself and aggressively minimize cost per mile.
• ▶ 27:15 Musk believes Optimus will eventually be worth more than everything else Tesla does combined, framing it as the next trillion-dollar bet aligned with SpaceX and xAI's AI/compute strategy.
• ▶ 27:34 The core investment thesis: the global market for physical human labor is worth tens of trillions per year, so a robot that can reliably replace even a slice of that work would claim a massive market.
• ▶ 28:00 The real bottleneck for millions of working robots isn't hardware/factories — it's compute and training time, because unpredictable human environments (like kitchens) are far harder for AI to master than regulated roads.
• ▶ 28:30 Nvidia transformed from a gaming graphics card maker into the core supplier for AI compute, now central to the massive infrastructure behind a Tesla-SpaceX convergence.
• ▶ 28:51 Nvidia's AI chip business generates over $30 billion per quarter, with top chips priced at $30k–$40k each and bought in hundreds of thousands, giving it continuous pricing power.
• ▶ 29:08 Every major AI player—including xAI, Anthropic, and even Google with its own custom chips—depends on Nvidia, so the more the industry spends on AI, the more Nvidia earns.
• ▶ 29:22 Tesla is strategically reducing dependence on Nvidia by aiming to own the entire AI compute stack rather than just purchasing capacity.
• ▶ 29:36 Tesla is designing one unified AI5 chip family used across its cars, Optimus robot, and data-center training, having shut down its old separate training computer.
• ▶ 29:44 Three Musk-affiliated companies are collaborating on a Texas chip factory called Terafab to fabricate AI5 silicon in-house at massive scale, controlling design, fabrication, and the full compute stack.
• ▶ 29:56 Apple's chip strength didn't translate to AI leadership because it chose not to build frontier AI, leaving it "way behind" and renting Siri's AI from Google for ~$1B/year.
• ▶ 30:18 Tesla is doing the opposite—building all AI (chips, models, inference, training) in-house—but a merger with SpaceX/xAI would plug Tesla into far more compute, enabling FSD training in days instead of weeks.
• ▶ 31:11 Musk's vision is one integrated machine, and Gwynne Shotwell's IPO-day comments confirm it: joining the pieces means owning every part of the supply chain—full vertical integration.
• ▶ 31:31 None of the convergence outcomes are guaranteed, and there are real reasons the optimistic scenario may not materialize.
• ▶ 31:43 Morningstar values SpaceX at roughly $63 per share—less than half the IPO price—and warns it's priced for flawless execution; a finance professor adds that overvalued assets like SpaceX, OpenAI, and Anthropic could hurt ordinary people through pension funds.
• ▶ 32:13 Price targets range from $75 to $200, illustrating an exceptionally deep disagreement between skeptics and believers.
• ▶ 32:20 The core financial imbalance: AI operations are losing billions and bleeding cash, while Starlink profits are used to subsidize them.
• ▶ 32:30 The whole system depends on investor patience with losses and customers continuing to rent compute from xAI, with a threat of them switching to other data centers.
• ▶ 32:43 Vertical integration becomes a double-edged sword—if AI spending fails to generate returns, the company is left carrying a massive, unmonetizable bet.
• ▶ 32:51 The biggest structural risk is that the AI buildout is a speculative bubble, with the industry spending hundreds of billions annually on returns that may never materialize.
• ▶ 33:12 Much of the spending is circular: the same companies invest in and rent from each other, creating an illusion of demand that could collapse if growth stalls.
• ▶ 33:26 Musk’s massive supercomputing bet only works if AI keeps consuming a larger share of the economy every year — but nobody actually knows if that will happen.
• ▶ 33:32 The discussion centers on the systemic risk of unchecked concentration of power, with one individual controlling foundational technologies and infrastructure.
• ▶ 33:39 Musk’s control would span the primary U.S. rocket company, satellite constellations, leading AI, chip factories, humanoid robots, self-driving cars, and the data centers training it all.
• ▶ 33:51 This scale of overlapping, infrastructure-level control invites a “hard antitrust look,” and a Tesla–SpaceX merger would face real regulatory scrutiny before closing.
• ▶ 33:57 Musk is already worth over $1.4 trillion, positioning him as one of the wealthiest individuals in history before any additional gains.
• ▶ 33:57 A Tesla–SpaceX convergence would likely push Musk’s net worth even higher, creating massive personal financial upside beyond company-level benefits.
• ▶ 34:13 The political wind is at Musk’s back, with a nationalist rationale favoring a single American company championing AI, space, autonomous vehicles, and robotics.
• ▶ 34:16 The central governance problem is that Elon Musk would be negotiating against himself, effectively setting the terms of any Tesla-SpaceX deal from both sides.
• ▶ 34:20 The 2016 SolarCity merger serves as a precedent: it triggered years of litigation over accusations that Musk was selling one of his own companies to another, and a SpaceX-Tesla combination would reignite that fight at a far larger scale [34:31–34:35].
• [34:39–34:46] Tesla and SpaceX are already deeply interconnected through massive vertical integration under Musk, making the conflict-of-interest question especially pronounced and difficult to disentangle in a formal merger.
• ▶ 34:46 SpaceX's IPO gives Elon Musk the one essential tool he previously lacked to make a Tesla-SpaceX convergence "official"—public stock.
• ▶ 34:55 A potential merger is strategically aligned with when Tesla needs more compute for future products.
• ▶ 35:01 The transformation could turn Tesla from a car company into one of the backbones of artificial intelligence for the future.
• ▶ 35:08 Host commits to honest reporting, covering both positive and negative aspects of the technologies discussed.
• ▶ 35:15 Asks viewers to comment on whether these developments are "great for the future" or "bad for the future."
• ▶ 35:24 Promotes a separate full video on Tesla's Optimus project, linked on screen and in the description.

• ▶ 0:00 SpaceX IPO and the AI Thesis (0:00 - 4:54) - SpaceX’s record IPO, Musk’s trillionaire status, first public financials, and the Space XAI segment plus xAI acquisition.
• ▶ 4:54 Tesla’s AI Compute and Energy Role (4:54 - 10:50) - Tesla trains AI on Cortex, sells AI power, and Tesla Energy Megapacks grow as the merger-line catch emerges.
• ▶ 10:50 The AI Infrastructure Race (10:50 - 16:44) - Compute scaling makes data centers heavy industry; Rivian’s compute gap and Colossus’s build speed highlight the spending and talent race.
• ▶ 16:44 xAI Becomes an Infrastructure Powerhouse (16:44 - 21:53) - The power/land/water ceiling, Colossus’s hardware flaw, Anthropic and Google rental deals, Grok’s rise, and over $2B/month in compute revenue.
• ▶ 21:53 SpaceX-Tesla Convergence and Musk’s Consolidation Pattern (21:53 - 35:35) - Elon Web Services, Tesla’s compute gap, Gwen Shotwell’s remarks, IPO-enabled combination, and Musk’s consolidation pattern.