US stock futures fall as tech remains skittish; nonfarm payrolls on tap

Broadcom fell 12.6% after guidance failed to meet high expectations and the company held its 2027 AI chip sales outlook at $100 billion, triggering profit-taking across semiconductors. The Philadelphia Semiconductor Index dropped 2.2%, while the Nasdaq ended slightly lower and the Dow hit a record close as investors rotated into more economically sensitive sectors. Markets now await Friday's May nonfarm payrolls report, which could influence Fed rate expectations amid sticky inflation and cooling growth.

Analysis

The important read-through is not simply “tech down, cyclicals up,” but that the AI complex is shifting from multiple expansion to proof-of-demand and margin durability. Broadcom’s miss matters because it weakens the credibility of the “every earnings print can be a raise” regime that has supported semi multiples; that usually triggers factor de-leveraging across the entire infrastructure stack for several sessions, even in names with different end-markets. NVDA holding up is telling: the market is discriminating between pure AI demand leaders and second-tier beneficiaries whose narratives are more easily capped by guidance.

Second-order effects favor the picks-and-shovels with the cleanest pricing power and balance sheets, while exposing the more cyclical semi names to inventory and order-normalization risk over the next 1-2 quarters. AMD, INTC, and MU are more vulnerable if payrolls come in soft, because weaker labor data would reinforce the “growth cooling” narrative and compress forward multiples just as the sector loses its earnings momentum. If rates stay higher for longer, the discount rate headwind is enough to keep semis under pressure even if the AI capex story remains intact.

The contrarian setup is that the pullback may be shallow if payrolls surprise to the upside: a resilient labor print would revive the “higher for longer” rates trade, but also support cyclicals and value versus long-duration tech, extending the rotation rather than reversing it. Conversely, a weak payroll number could help the Fed cut expectations, which is typically supportive for tech beta after an initial risk-off knee jerk; that creates a narrow window where the first move lower in semis may be the best entry for longs in quality names. The key is to separate narrative damage from fundamental damage — one bad guide is not the same as a demand rollover.

Key Decisions for Investors

• Short AVGO vs long NVDA for 2-6 weeks: broadcom-style guidance disappointment should keep pressure on the broader AI infrastructure basket, while NVDA retains the strongest demand signal; target 1.5-2.0x gross return if the semi rotation persists.
• Buy AMD/INTC downside via 1-2 month puts or put spreads into payrolls: these names have the most fragile multiple support if labor data softens and the market prices weaker growth plus still-elevated rates.
• Use any post-payrolls flush to add to NVDA on weakness, not strength: preferred entry is a 3-5% selloff on a hot payroll print, where the first reaction is rates-driven and usually mean-reverts faster than the secular AI bid.
• Pair long XLI / short SOXX for the next 2-4 weeks if payrolls are firm: this isolates the rotation away from long-duration growth and should benefit from a broader “real economy over AI-beta” factor move.
• Hold off adding to MU until the next inventory/channel check: memory names are the most exposed to a delayed capex pause if the Broadcom move proves to be the first sign of a broader ordering reset.