InvestAnswers: SpaceX "stupid cheap" under $135 after AI selloff triple whammy

James frames the AI selloff as a "triple whammy," not a black swan

Responding to a Patreon question about black swan risk amid government "72-hour stockpile" warnings in the UK, Japan and the Netherlands, James says this is routine EU preparedness guidance, not a sign of imminent catastrophe. He argues wars already exist (Ukraine, Russia, Iran) and historically have limited market impact. Instead, he describes last week's selloff as three overlapping "gray swans": capex/debt fears around the AI buildout, midterm politicians threatening to curb AI data centers, and a repricing of "bubble" valuations he considers unjustified, citing Micron at a PE of 5 and Nvidia at 13.

He compares the size of the drawdown to COVID, saying the S&P 500 tech sector fell 7% in its worst day since March 2020, and that overall losses were larger than the March 2020 selloff, though markets have since grown. He rejects dot-com bubble comparisons outright:

"No, it's nothing like it at all. Nothing like it."

SpaceX, Starlink and "Starmind": the space data center thesis

James disputes a claim that Texas canceled data centers, saying local permitting for 2027 builds is locked in and components like Vera Rubin chips and memory are sold out for years. He references Elon Musk's interview with Ron Baron discussing "Starmind," space-based AI data centers, with first units planned for orbit in Q3 2027 and "huge ramps," a 10x increase in 2028 and another 10x in 2029-2030, using radiation-hardened chips co-designed with Nvidia on Vera Rubin NVL72 systems.

He frames this as a possible tipping point where more US data center capacity could eventually sit in space than on the ground, calling it speculative: "this is not a forecast, it's my what-if tipping point." He also notes Starship's upcoming flight 15 double-catch attempt as key to full reusability, which he sees as central to the space data center thesis.

Tesla-SpaceX merger math and portfolio concentration

Multiple viewer questions concerned a possible Tesla-SpaceX merger. James models Tesla reaching a $10 trillion market cap and SpaceX $15 trillion standalone by 2030, implying roughly 6.6x and 7-12x returns respectively from current levels. He says a merged entity could be worth more than the sum, potentially $25 trillion combined, implying a per-share price around $990, giving SpaceX holders a 7.3x return and Tesla holders more due to a premium. He warns about share dilution from the deal and advises against pair-trading in taxable accounts to avoid triggering taxes.

On concentration, he tells one listener with 53% exposure to "Musk Industries" that a 30% drop in those holdings would cut their total portfolio by "1650" (16.5%), calling that risky but notes Ron Baron is "all in" on both stocks and has profited. His own guidance: keep Musk-related exposure near "55% max" and consider buying SpaceX under $140 or Tesla under $350.

Marvell leaps, Google's SpaceX stake, and other flyers

On Marvell, James recaps buying around $80, watching it run to near $300, then fall to $141 after what he calls a "triple whammy" including the Lumentum merger announcement. He says 2029 $200 leaps should still work out:

"I would say you're fine... everything will be recovered and those will all be forgotten."

He cites a Wall Street one-year target of $338 and notes Marvell's market cap fell from roughly $200 billion to $120 billion.

He dismisses concern over Google's additional $500 million investment in SpaceX, framing it as Google buying "a way of getting compute" via launch capacity, noting Google separately "lost 500 billion in a day" on other news. On a drone-play question about UMAC with a cited $27-plus target, James says he holds it as a small "flutter position" (about a tenth of 1% of portfolio) and is unsure how it will play out, cautioning against going "all in on flyers."