Maria's Market Note · SpaceX Earnings

SpaceX Beats Expectations — But Wall Street Questions the Cost

Revenue nearly doubled and every division expanded. The selloff shows that investors are not questioning growth — they are questioning the price of reaching it.

A Strong Report — and a Falling Stock

SpaceX generated $7.81 billion in second-quarter revenue, beating Wall Street's roughly $6.9 billion expectation. Its net loss narrowed from $1 billion to $541 million.

The business mix shows that SpaceX is no longer simply a rocket company. Starlink and connectivity produced $4.29 billion, up 66%. AI, including xAI and Grok, generated $2.56 billion, up 247%. Space operations contributed $962 million, up 29%.

Starlink remains the largest division, but AI already produces more than twice the revenue of the traditional space business. Even so, shares fell after hours. The market's concern was not growth — it was spending.

The Price of Ambition

SpaceX reportedly invested approximately $18.4 billion during the quarter, including about $15.8 billion in AI infrastructure — more than six times the division's quarterly revenue.

Musk plans to build SpaceX's AI infrastructure exclusively with Nvidia chips and reach as much as 10 gigawatts of computing capacity by the end of 2027. At the same time, the company is funding Starship, Starlink satellites, mobile connectivity and data centers.

SpaceX has about $100 billion in cash and marketable securities and a $47.5 billion order backlog. That provides room to invest, but investors want proof that the spending can create durable cash flow.

Has Wall Street Been Too Short-Term Before?

Yes. Amazon faced years of criticism for investing heavily in fulfillment, Prime and cloud infrastructure before AWS became its leading profit engine. Meta, Alphabet, Microsoft and Tesla have faced similar resistance when long-term investment pressured near-term profits.

Sometimes Wall Street underestimated the opportunity; other times, its skepticism correctly exposed spending that failed to deliver acceptable returns. SpaceX must prove it is building essential infrastructure — not merely financing an expensive vision.

Starship could be decisive. Management believes it can reduce launch costs by as much as 99%, potentially transforming satellite deployment, global communications and space-based computing.

Musk also expects SpaceX to reach $1 trillion in annual revenue by 2030, possibly 2029. Compared with its current annualized revenue of approximately $31 billion, that is an extremely ambitious target and should be viewed as a vision — not a conservative forecast.

The outcome depends on three things: Starlink's margins, AI profitability and Starship's execution.

And while Musk continues carving America's path into the space age, I'll be on my balcony — as always — watching each launch and witnessing that future take flight.

After-Market Reports · Aug 4, 2026

Major U.S. Earnings Scorecard

Quick ranking: ALAB delivered the cleanest beat and guidance raise, followed by ANET and AMGN. AMD beat estimates, but lofty expectations drove the selloff. PINS guidance showed slowing growth. SpaceX spending overshadowed its revenue beat.

After-hours moves are approximate. EPS figures generally adjusted where consensus comparisons were available. Sources: Reuters, company releases, MarketBeat, Barron's, The Times.

Not Subscribed Yet?

Sign up for the newsletter here

The AI Trading Institute

AUGUST COHORT WAITLIST OPEN

The July cohort sold out. Add your name to the August waitlist and we'll notify you the moment enrollment opens.

The Market Note · AMD Earnings

AMD Beat Earnings. So Why Is the Stock Down?

Revenue surged, Data Center doubled, and guidance beat Wall Street. The selloff says more about sky-high expectations than AMD's business.

AMD delivered a strong quarter, but the stock still dropped more than 10% after hours.

Revenue climbed 50% year over year to $11.54 billion, beating expectations of $11.28 billion. Adjusted earnings came in at $1.66 per share versus $1.62 expected.

The real star was Data Center. Sales more than doubled to $6.7 billion as demand surged for AMD's Epyc CPUs and Instinct AI accelerators.

So, what went wrong? Nothing, really.

Investors Wanted More

AMD expects approximately $13 billion in revenue next quarter, ahead of Wall Street's $12.52 billion estimate. However, some analysts were hoping for guidance as high as $14 billion.

That's the problem with a stock that has nearly tripled in one year. Good numbers are no longer enough — the market wants perfection.

This looks like investors once again getting ahead of themselves and then punishing the company for failing to beat expectations that were never realistic to begin with.

AMD's AI Story Is Getting Stronger

CEO Lisa Su said demand for AMD's accelerators and CPUs is running well above previous expectations. She also expects Data Center sales to double again in 2027.

AMD will begin shipping Helios, its first rack-scale AI system, to customers including Meta, OpenAI, and Oracle. Helios combines AMD's CPUs, GPUs, and networking technology into one complete system, putting the company in more direct competition with Nvidia.

AMD is no longer competing for a small piece of the AI market. It wants a meaningful share of the entire AI infrastructure buildout.

Sometimes a stock falls because the business is weakening. Other times, it falls because investors expected too much. This looks like the second one. — Maria

The AI Money Question · AI & Semiconductors

Is Nvidia Going 'Buy Here, Pay Here'?

Nvidia may be preparing to do more than sell the systems powering the AI boom. It may also help make the financing behind them possible.

Yes, billion with a B

The Wall Street Journal reports that Nvidia is discussing a financial backstop of roughly $250 billion to help OpenAI lease a massive data-center campus in Ohio. The talks are not a completed agreement, and the final amount and terms could change — or the proposal may never be finalized.

A credit backstop is a financial safety net, not a $250 billion check written directly to OpenAI. Nvidia's backing could give lenders more confidence and make it easier to finance the lease and construction of the project. The reported guarantee would not directly pay for the Nvidia chips inside the facility.

Why would Nvidia do it?

OpenAI could become one of the world's largest buyers of AI computing power. Helping remove the financing bottleneck could lock in years of Nvidia sales, keep competitors from taking that business and give Nvidia better visibility into OpenAI's future computing needs.

Nvidia would not simply wait for the AI highway to be built. It may help clear the road — and then collect the tolls.

The question Wall Street should be asking

Is AI demand strong enough to finance itself, or does Nvidia need to help finance the demand?

There is a major difference between customers funding their own purchases and a supplier helping support the financing behind them. The demand may be real, but investors still need to know how much risk Nvidia must accept to turn that demand into revenue.

Demand for Nvidia's systems remains strong, and the data-center buildout continues. The concern is not whether customers want Nvidia's technology. It is whether these enormous projects can stand on their own financially — and how much responsibility Nvidia is willing to accept to keep them moving.

If this happens, Nvidia may need a collections department — and a few GPU repo men. The repo-men line is a joke; the need for financial safeguards is not.

The Week Ahead

Earnings Watch: Aug 5–7

A focused trading calendar for the most market-moving reports

Wednesday, Aug 5

Before Open · High risk

LLY, DIS, UBER, CVS, CRCL, GFS, FLUT

After Close · Highest remaining volume

APP, DASH, DUOL, XYZ, SNDK, WDC, IONQ, SOUN, SMR, JOBY, AXON, EBAY, EXPE, HUBS, FIG, ZG, OXY

Thursday, Aug 6

Before Open · High risk

DDOG, CEG, COP, WBD, QBTS, HWM, PTON, RL

After Close · Cloud/ad-tech/quantum focus

NET, ABNB, DKNG, ROKU, TTD, TWLO, TEAM, MP, RGTI, AAOI, LYFT, RKT

Friday, Aug 7

Before Open · Lighter session

OKLO, ACMR, ENB, FLR, UAA, WEN — no major market-moving reports after close.

Week of August 3 — Two SPX 0DTE trades

Two SPX 0DTE plays — late-morning and late-day. Different strategies at different entry times.

Free Community

Join a community of 900+ like-minded traders in our Discord who trade based on math

Trade ideas, screenshots, daily banter. Zero cost.