AMD fell as much as 7% after hours, while SpaceX dropped about 6%. The earnings reports were not bad. The problem was that neither stock had been priced for a normal beat.
AMD came into earnings with expectations set too high. SpaceX is simply spending money too fast.
AMD: The Surprise Had Already Been Priced In
AMD reported second-quarter revenue of $11.54 billion, up 50% year over year, with adjusted EPS of $1.66. The company also guided for roughly $13 billion in third-quarter revenue.
Data center revenue reached $6.72 billion, up 107%, as both EPYC CPUs and Instinct accelerators continued to ramp. Put these numbers on almost any other semiconductor company, and investors would have very little to complain about.
AMD still got hit after hours.
The first reason was simple: the stock had already gained about 7% before the report.
Investors were not betting on AMD to deliver a small beat. They were betting that MI450 and Helios would pull forward a much larger earnings breakout.
Instead, revenue beat expectations by roughly 2%, EPS by about 2.5%, and third-quarter guidance by around 4%.
Those are solid numbers, but not enough to justify the run-up ahead of earnings.
The second problem was gross margin.
AMD’s adjusted gross margin remained around 56%, while free cash flow fell from $2.57 billion in the previous quarter to $1.56 billion. AI chip sales are growing rapidly, but margins are not expanding with them.
That naturally leaves investors wondering whether AMD is winning business through stronger products or lower pricing.
This selloff looks more like a classic “buy the rumor, sell the news” reaction than a sudden collapse in the AI story.
The market did not stop believing in AMD. It simply realized that the biggest contributions from MI450 and Helios may not arrive until the fourth quarter or even 2027. Some of the expectations priced in ahead of earnings had to come back out.
This is also a warning for optical networking and memory stocks.
AMD confirmed that AI infrastructure demand remains strong, but the market no longer rewards companies simply for having orders. Those orders must quickly become revenue, and that revenue must eventually become profit.
Otherwise, even a good earnings report can still lead to a selloff.
SpaceX: The Faster Revenue Grows, the More Investors Worry About Spending
SpaceX reported approximately $7.8 billion in second-quarter revenue, up 92% year over year and well above expectations. Its losses also narrowed sharply.
Starlink remained the strongest part of the business.
Connectivity revenue reached about $4.29 billion, with operating profit of nearly $1.66 billion and a margin close to 39%. Paid subscribers climbed to roughly 12 million.
Starlink is no longer just an expensive project chasing users. It has become SpaceX’s most reliable source of cash.
The problem is that even this cash cow may not be enough to support the company’s current expansion plans.
SpaceX spent nearly $18.4 billion on capital expenditures during the quarter—more than twice its total revenue. Around $15.8 billion went into AI infrastructure.
The market had expected AI capital spending of about $13 billion. SpaceX spent nearly $3 billion more than that.
That was the real reason behind the after-hours selloff.
Investors did not just see SpaceX building a new AI business. They saw a company constructing data centers at a pace far beyond its current revenue. The cash generated by Starlink is immediately being consumed by AI infrastructure and Starship development.
The AI business has reached positive adjusted EBITDA, but it is still losing money on an operating basis. Starship also continues to require heavy investment.
In other words, Starlink remains the only part of SpaceX currently producing stable profits, while the company’s other major businesses still need more capital.
If AI contracts continue to grow and Starship stays on schedule, today’s spending could eventually create enormous value. But at $18.4 billion of quarterly capital expenditures, every delay becomes extremely expensive.
SpaceX had also rallied ahead of earnings as short sellers covered their positions. Once the results came out and investors saw how far capital spending had exceeded expectations, short-term traders started locking in profits.
There is another problem: around 912 million shares are expected to enter their first lockup-expiration window on August 6.
The market is not worried about weak earnings. It is worried about absorbing a potentially huge increase in share supply.
Even a strong quarter may not be enough to offset the possibility of nearly one billion shares becoming eligible for sale.
Why Could Other Space Stocks Get Hit as Well?
SpaceX has become the valuation benchmark for the entire commercial space sector.
It grew revenue by 92% and generated a Starlink operating margin of nearly 39%, yet the stock still sold off because spending was too high.
That naturally leads investors to ask a harder question: if even SpaceX still needs enormous amounts of capital, when will less mature space companies generate real cash flow?
If RKLB, ASTS, LUNR and RDW come under pressure, it does not necessarily mean they suddenly lost contracts. It may simply mean the market is recalculating how much money the entire industry will need before reaching profitability.
RKLB still has one of the cleaner setups. SpaceX is reserving more Falcon 9 capacity for Starlink, so customers genuinely need a second reliable commercial launch provider.
But until Neutron succeeds, RKLB is still trading on expectations. The higher the stock rises, the less patience investors will have for delays or technical failures.
ASTS faces a more direct problem. Starlink proves that satellite communications demand is real, but it also proves that SpaceX is becoming an increasingly powerful competitor.
ASTS will ultimately need to prove itself through satellite deployment and carrier partnerships.
LUNR and RDW depend more heavily on NASA contracts, defense spending and individual project execution. They should not be bought simply because SpaceX reports strong earnings, but they should not be written off just because SpaceX falls after hours either.
I have not turned bearish on the space sector, but this is not a good level to chase the entire group.
For SpaceX, the first test is whether the market can absorb the selling pressure after the August 6 lockup expiration. For RKLB, it is still Neutron. For ASTS, it is satellite deployment.
Companies with real contracts, differentiated technology and enough capital to survive until cash flow turns positive can still work. Stocks rising purely on SpaceX-related excitement deserve more caution.
This selloff is not the market rejecting AI or commercial space.
It is the market asking a much more practical question:
After spending all this money, when will investors finally get paid?
Comments