AI Chip Selloff Spreads Across Global Markets | Bloomberg Tech 7/07/2026

Samsung’s quarterly profit surged on stronger memory demand, but the beat still wasn’t enough for investors, pointing to expectations risk despite improving fundamentals. Amazon returned to the bond market with plans to raise at least $25B to fund AI infrastructure spending, supporting credit and growth sentiment. SpaceX also added Nasdaq 100 inclusion and is seeing new sell-side coverage, which is incremental positive for positioning but not a near-term catalyst on its own.

Analysis

Samsung's earnings rebound matters less as a headline than as evidence that memory pricing has finally turned, but the market is telling you it wants proof of duration, not just one quarter of supply discipline. If AI-driven HBM demand is real, the higher-beta winner is MU, not SSNLF: Micron has more operating leverage and should rerate first, while Samsung's conglomerate discount and weaker transparency mute upside. If the next memory upcycle is only a cyclical pop, the move will fade within 1-2 months; if HBM contract pricing keeps rising into next quarter, this becomes a 6-12 month equity story.

Amazon's bond takeout is a financing signal, not a demand signal. Debt issuance can smooth the cash drag from AI capex and keep equity dilution off the table, but it also resets the hurdle: investors will now judge AMZN on AWS margin and incremental ROIC, not on the size of the buildout. Second-order winners are the infrastructure enablers—VRT, ANET, ETN/PWR—while the risk is that the market caps AMZN as a capex-first, FCF-later story for several quarters if spend outruns monetization.

The SpaceX index/coverage story is mostly sentiment, but it can spill into the public space-exposure complex and keep speculative flows alive. That tends to help names like RKLB/ASTS on narrative, though the effect is usually short-lived unless it translates into backlog or launch cadence. Falsifier for the bullish AI-infra trade: if Amazon's bond comes at tight spreads and AWS growth re-accelerates, the market will treat this as constructive balance-sheet management rather than a capital-intensity warning.

Key Decisions for Investors

• Relative value: long MU / short SSNLF for 1-3 months; enter on any post-print weakness, with the thesis invalidated if DRAM spot prices roll over or Micron guides HBM mix lower.
• Buy VRT or ANET on AMZN pullbacks as the cleaner AI capex beneficiaries; 6-12 week horizon, looking for order backlog confirmation from hyperscaler spend.
• No outright short AMZN yet; if post-deal spreads are tight and AWS commentary stays firm, capex fear is overdone. Reassess only if 2Q/3Q capex-to-sales rises without revenue acceleration.
• Watch RKLB/ASTS only as sentiment trades; fade any first-day spike unless accompanied by a quantified contract/backlog catalyst.