Three Earnings, One Verdict: AI Capex Will Decide Whether This Week's Beats Actually Matter

Five S&P 500 names report inside seven sessions, and the order itself shapes the trade. Netflix opens things on Tuesday, Oct. 20 (results at approximately 1:01 p.m. PT). Tesla and Meta Platforms follow on Wednesday, Oct. 21, both with after-close prints (Tesla at 5:30 p.m. ET, Meta at 1:30 p.m. PT / 4:30 p.m. ET). American Express closes the consumer leg on Friday, Oct. 23 at 8:30 a.m. ET. Then the capex decision arrives: Microsoft and Alphabet on Oct. 28, Apple Oct. 30, Amazon Oct. 29.

Tesla's setup is more demanding than the headline suggests. Deliveries of 486,532 beat consensus by roughly 22,000–29,000 vehicles — a 5% gap — and shares already moved on Oct. 2. The Oct. 21 release therefore trades on the print's components, not the headline. Consensus sits at $0.43–$0.45 in EPS on roughly $27.8B of revenue, down from $0.50 a year earlier; gross margin in Q2 2026 was already 16.8%, well below the 18%+ the street models in a normal year.

Netflix goes into its Oct. 20 print already wearing a Q2 statement on its back: management guided Q3 2026 to $12.86B of revenue and $0.82 of diluted EPS, a forecast that fell short of the Wall Street bar and drove shares down more than 8% in after-hours on July 17. Full-year 2026 revenue was reaffirmed at $51.0B–$51.4B. Q2 2026 itself delivered $12.56B of revenue and $0.80 of EPS — close to the original guide but not the upside the bull case requires.

The cleanest read on U.S. premium-consumer health in the week comes from American Express, which reports pre-market on Oct. 23. Q2 2026 already showed the consumer cohort is intact: billed business grew 9% on an FX-adjusted basis — the strongest quarter in three years — with U.S. consumer spending up 11% and travel bookings jumping 22%. Revenue rose to $21.60B and diluted EPS hit $4.58, both records for the franchise at that point.

The thesis of the week is not whether Tesla, Netflix and AmEx beat consensus — they have all set up reasonable enough bars — but whether the consumer narrative can survive the hyperscaler verdict on Oct. 28–29. The July precedent is the cleanest evidence: when Alphabet raised its 2026 capex guide from $180–$190B to $195–$205B after a Q2 print of $44.9B in single-quarter capex, GOOGL dropped about 5% on the day and pulled Amazon, Microsoft and Meta with it.

The suppliers are pricing the AI capex thesis before the hyperscalers confirm it on the call. NVIDIA's Q2 FY2027 revenue of $96.2B (+106% YoY) and data-center revenue of $89B (+117%) make clear the order book is full through2026 — the Q3 FY2027 consensus of $103.8B would extend the run. TSMC is on the other end of the same pipeline, and SK Hynix supplies more than half of the world's HBM (high-bandwidth memory) for AI accelerators, which makes it the cleanest single-name beneficiary of any re-acceleration in 2026/2027 hyperscaler spend.

Short-term (days–quarters): the Tesla print is the highest-conviction beat candidate (deliveries already ahead, energy storage at a record 13.7 GWh), and AmEx is the cleanest read on premium consumer health at a19.0x P/E with a 1.13% dividend yield. Netflix is the most binary — ad-revenue traction versus a stock that has already given back some of its post-Q2 selloff's recovery. The cross-asset risk is Microsoft/Alphabet/Amazon on Oct. 28–29 re-anchoring capex at or above the current $610–$730B run-rate.

Long-term (1–3 years): the bull case for the consumer names is that hyperscaler spending remains a 2026–2027 phenomenon, and the 10-year Treasury drifts back below 4.5% as the Fed eventually eases — which would let Tesla, Netflix and AmEx re-rate higher on multiple expansion. The bear case is that hyperscalers hold capex at $200B+ through 2027 (Bank of America sees 90% of operating cash flow absorbed in 2026), the 10-year stays above 5%, and consumer-discretionary multiples stay compressed while the AI-infrastructure trade absorbs index-level capital.