Apple Slides 6%, Samsung Soars, SpaceX Braces: July 31

Apple’s record quarter couldn’t keep investors happy, Samsung delivered one of the most dramatic profit reversals in memory chip history, and SpaceX heads into a critical week with its stock already battered and a 911 million share unlock looming.

Key points

• Apple slid roughly 6% after its Q3 2026 beat, with analysts flagging AI-related margin pressure and supply chain risks as the forward concern.
• Samsung’s Q2 operating profit surged 1,800% year over year, driven by AI-fueled memory chip demand that pushed revenue to a record high.
• SpaceX closed at $112.55 on July 29, roughly 50% below its post-IPO peak, with analysts placing a $237 bull case and a $62 bear case ahead of its August 4 earnings report.
• HSBC Private Bank argues the recent Asian semiconductor sell-off reflects recalibrated earnings expectations, not a breakdown in the AI investment cycle.
• Tesla is reportedly preparing to separate its China business, citing a potential merger with SpaceX according to the Wall Street Journal.

What does Apple’s slide say about AI economics?

Apple beat on revenue in Q3 2026, yet the stock fell around 6%. The market’s reaction points to something specific: investors are increasingly focused on what AI costs Apple going forward, not what it earned last quarter. Margin compression tied to AI infrastructure spending and supply chain questions appear to be driving the concern.

This is a pattern worth watching across big tech. Strong reported numbers are being discounted when the cost structure for AI integration looks uncertain. Apple is not a pure-play AI stock, but it is spending heavily to embed AI features across its ecosystem, and those costs are now visible enough to move the stock on an otherwise good earnings day.

Samsung’s 1,800% profit jump: what it means for the memory trade

The headline figure is striking. Samsung’s operating profit rising 1,800% year over year reflects just how badly the prior-year comparison period was, but it also confirms that AI-driven demand for high-bandwidth memory is real and flowing through to revenue at scale.

This matters for investors tracking the broader AI infrastructure build. Memory is a core bottleneck in AI training and inference workloads, and Samsung is a primary supplier. A record revenue quarter validates the demand thesis that HSBC has been defending publicly.

HSBC Private Bank’s position is that recent volatility in Asian semiconductor names is a valuation reset rather than a signal that AI capex is slowing. Samsung’s results give that argument some grounding in actual financials rather than just sentiment.

SpaceX: the most watched earnings debut of the summer?

SpaceX is a week away from reporting its first earnings as a public company, and the setup is unusually high-stakes. The stock has already swung from an all-time low to a sharp rebound and back in a single week, and that churn is happening before the two events that could move it most dramatically.

August 4 brings earnings. August 6 brings the unlock of 911.5 million insider shares. The gap between the analyst bull case ($237) and bear case ($62) against a current price near $112 reflects genuine uncertainty about how to value a company that spans launch services, Starlink broadband, and now a potential merger with Tesla’s operations.

Tesla executives have reportedly been directed to prepare for a China business separation as a precondition for a possible SpaceX tie-up, per the Wall Street Journal. A Tesla-SpaceX combination would be one of the larger corporate events in recent memory, and it adds another layer of complexity to valuing either company independently right now.

Elsewhere in AI infrastructure

Taiwan Semiconductor is reportedly adopting a strategy with echoes of Intel’s playbook, with analysts suggesting it reveals where the real bottleneck in AI compute actually sits. The details are worth watching as TSMC’s manufacturing decisions ripple through every major AI chip supplier.

Kioxia announced the CM10 Series enterprise SSDs, built on its 10th-generation BiCS FLASH 3D memory and designed for AI inference and key-value cache workloads, including compatibility with Nvidia systems. Storage is often overlooked in AI infrastructure conversations, but it becomes a meaningful cost and performance factor at data center scale.

Nothing here constitutes investment advice. All figures are sourced from the articles cited and reflect conditions as of July 31, 2026.