Starlink is real and the launch business is real. No argument there.
But the market values SpaceX at $2.1T. Morningstar, who is the most bearish, values it around $780B - and that already includes Starlink and launch. So the extra ~$1.3T the market is paying is for the AI and orbital story. That is the part worth checking.
It comes down to two things.
The first is the orbital data center projection. Goldman, who led the IPO, expects the AI division to grow 100x to $322B by 2030. That number is what holds up the valuation. But SpaceX's own filing says it "relies on unproven technologies and may not become commercially viable." It also needs full Starship reusability, which has not happened yet - the ship has never been caught, only the booster has, and the latest booster was lost in May.
The second is the AI revenue that actually exists today - around $26B from Anthropic ($1.25B/month) and Google ($920M/month) renting Colossus. People see that and assume SpaceX is already winning AI infrastructure. It is the opposite. Anthropic and Google are renting Colossus because xAI couldn't use it properly - around 11% utilization vs 40% for rivals, and xAI moved its own training to Colossus 2. xAI lost $6.4B last year and lost all 10 of its co-founders.
So the main proof of AI execution is competitors paying to rent the GPUs that SpaceX's own AI couldn't use.
Starlink is the floor. But the premium on top - the part that gets you to "bigger than Nvidia" - is an unproven orbital bet plus a data center its own AI couldn't run.
That is a very different company than the one being priced.
I think many people look at Elon Musk and envision SpaceX’s long-term potential.
They believe that getting in at any price is better than waiting and buying later once the stock has shot up even more.
That said, I believe the massive hype surrounding the IPO — which I don’t think has been matched by any other IPO to date (at least until Anthropic’s and OpenAI’s potential offerings this fall) — is causing a lot of people to jump on the hype train mindlessly out of FOMO.
Interestingly, around 60% of IPOs underperform the broader market in their first three years.
So if investors get in now, don’t mind underperforming the market over a three-year horizon, or scared of missing out or suffering a drawdown, then buying at these levels isn’t necessarily bad.
That said, in my opinion, it’s still better to be invested in SpaceX at any price than to stay on the sidelines entirely and risk missing out on its long-term potential.
I think many people look at Elon Musk and envision SpaceX’s long-term potential. They believe that getting in at any price is better than waiting and buying later once the stock has shot up even more.
That said, I believe the massive hype surrounding the IPO — which I don’t think has been matched by any other IPO to date (at least until Anthropic’s and OpenAI’s potential offerings this fall) — is causing a lot of people to jump on the hype train mindlessly out of FOMO.
Interestingly, around 60% of IPOs underperform the broader market in their first three years.
So if investors get in now, don’t mind underperforming the market over a three-year horizon, or scared of missing out or suffering a drawdown, then buying at these levels isn’t necessarily bad.
That said, in my opinion, it’s still better to be invested in SpaceX at any price than to stay on the sidelines entirely and risk missing out on its long-term potential.
It only makes sense for the Fed to shift its focus toward inflation rather than the labour market.
The labour market has proven resilient in the past couple of releases, while inflation keeps drifting upwards - even when stripping out hot oil prices.
Core PCE is running ~3.8% on a 3-month annualised basis and 3.3% year over year, and the reacceleration is now showing up in goods, not just energy.
That's not an oil story anymore.
Wrote this 4 days ago.
Since then: semis swung 10% on the week, and SPCX IPO'd and closed at a $2T valuation on day one.
MSCI and Russell are already adding it - meaning passive funds are forced to buy a stock with barely any float, whether they want it or not.
The concentration keeps building.
The bet gets more leveraged, not less.
Everyone says index funds are the safe, diversified way to invest. And usually they're right.
But at least for now, that may not hold for the S&P 500 — at least not if you're buying it for diversification rather than chasing the returns.
The S&P 500 isn't really 500 companies anymore. It's a concentrated bet on AI with 493 names along for the ride.
Take the AI names out and this year's rally disappears. The index is up 7.34% since the war began in late February. Strip out AI and that number drops to zero. Every bit of the 2026 rally traces back to a handful of stocks.
Friday made it impossible to ignore. The S&P had its biggest single-day drop since October 2025, falling 2.64%. But the index without AI stocks moved 0.02%. The 493 companies that make up 60% of the index did nothing. The entire selloff was AI unwinding.
The number that should scare people isn't the concentration. It's the gap between concentration and earnings.
The top 10 stocks now make up roughly 41% of the entire index — but they only generate about 32% of its earnings. Concentration has risen faster than the profits that justify it.
By the end of 2000, at the peak of the dot-com bubble, the top 10 made up about 23% of the index — and peaked during that year at around 27%. By the end of 2025, the top 10 accounted for nearly 41%. It has more than doubled in just 10 years — the highest concentration in modern market history.
One honest nuance: in 2000 the mega-caps traded at 43x earnings with under 20% of index profits. Today's leaders trade around 31x with 32% of earnings. So there's more fundamental backing this time. It's not pure fantasy.
But concentration risk is concentration risk. When you buy the S&P today, you're not diversified. You're making a leveraged bet on AI continuing to deliver — whether you know it or not.
That's exactly why equal-weight indexes exist — or indexes like the Dow, which carry a heavier weighting of defensives and leave you genuinely more diversified.
The key is to do your own research and invest according to your own risk tolerance. Diversified portfolios take less of a hit when things go bad — but they also enjoy less of the upside when things are good.
The 2000 dot-com crash is the perfect example. The S&P fell 49% over the following two years. The Dow, with less tech concentration, fell 37%.
And here's what actually outperformed when that concentrated trade unwound:
Defensive sectors led. Utilities and consumer staples posted positive returns while the index fell sharply — money flows to stable, dividend-paying businesses when growth stocks unwind.
Energy and commodities rallied as crude recovered and materials demand returned.
And US Treasuries rose around 13% as capital fled equities for the safety of fixed income.
The pattern is always the same. When the crowded trade breaks, money rotates into everything that was ignored during the boom — the boring, the defensive, the unloved.
Less concentration. Less hit on your investments.
That's the trade-off nobody thinks about until it's too late.