What this note is
A pre-print watchlist for AI-linked names reporting in the week of 26 October 2026. Dates below are from company IR releases. We keep these so the research log stays honest about what we were watching and what would have made the note wrong.
This is a community desk note. It is not investment advice. It is not a rating, a target, a position, or a size. Treat it as a viewpoint from the HelioQuant research log, not a recommendation to buy or sell anything.
Transcript deep-dive and post-print tables are deferred until a portfolio-monitor earnings transcripts API exists (that route is not live today). Until then, this note is calendar + watch-items only, sourced from public IR and news.
Setup
Five AI-adjacent names on this week's blotter. META already reported on 21 October 2026 (Meta IR), so it is off this list. NVDA has not reported yet; next print is scheduled for 17 November 2026 (Q3 FY2027) — see the separate NVDA follow-up note. INTC also prints Thursday (Intel IR); it stays a secondary watch, not a primary blotter name this week.
| Ticker | Period | When (ET) | Source |
|---|---|---|---|
| VRT | Q3 2026 | Wed 28 Oct, before open; call 11:00 a.m. ET | Vertiv IR, 14 Oct 2026 |
| MSFT | FY27 Q1 | Wed 28 Oct, after close; call 5:30 p.m. ET | Microsoft Source, 7 Oct 2026 |
| GOOGL | Q3 2026 | Wed 28 Oct; call 4:30 p.m. ET | Alphabet IR, 6 Oct 2026 |
| NOW | Q3 2026 | Wed 28 Oct, after close; call 5:00 p.m. ET | ServiceNow IR, 7 Oct 2026 |
| AMZN | Q3 2026 | Thu 29 Oct; call 5:00 p.m. ET | Amazon IR, 15 Oct 2026 |
Dates can move. Re-check the IR page the morning of the print.
Watching
VRT — power / cooling into the AI factory build
Watching: Orders and backlog tied to AI / hyperscale data-center liquid cooling and power. Whether management still frames supply (transformers, thermal) as the constraint rather than demand digestion. Margin color that separates AI mix from ordinary data-center work.
What would prove us wrong: A guide that pulls AI-related orders and attributes it to customer pause or cancellation, not lead-time or capacity timing. Soft industrial with steady hyperscale AI commentary keeps the watch item open.
MSFT — Azure / AI spend color
Watching: Azure growth rate and any language that ties cloud growth to AI training vs inference mix. Capex commentary that names capacity constraints (power, GPUs, racks) rather than directional "we will keep investing."
What would prove us wrong: A print where Azure decelerates and management frames AI demand as paused or deferred, not supply-constrained. Soft Azure with "still ramping" language keeps the watch item open; soft Azure with demand-walkback language retires the bull case for this week.
GOOGL — Cloud + Search AI productization
Watching: Google Cloud growth and operating-margin trajectory. Whether Search / YouTube AI features show up as monetization language or stay as product demos. Capex vs free-cash-flow tone.
What would prove us wrong: Cloud growth that stalls without a capacity excuse, or ad commentary that treats AI features as a margin drag with no offsetting engagement or pricing color. A clean Cloud beat that still walks back forward spend on demand grounds would also rewrite the setup.
NOW — AI agents vs seat expansion
Watching: Subscription billings / cRPO growth and whether AI / agent products are showing up as attach or ACV expansion rather than slideware. Operating margin vs growth trade-off language.
What would prove us wrong: Guidance that implies AI products are not converting to paid usage, or a clear deceleration in net-new ACV without a mix explanation. A beat that is only cost discipline with flat growth does not automatically confirm the AI-agent thesis — label it and wait for the next note.
AMZN — AWS AI attach and capex
Watching: AWS growth rate, AI-services commentary (Trainium / Inferentia / Bedrock attach), and the capex outlook into holiday + AI capacity. Retail is noise for this blotter unless it changes the cash available for infra.
What would prove us wrong: AWS deceleration framed as AI demand disappointment, or a capex cut that management links to lower AI utilization rather than timing. A soft retail quarter with steady AWS keeps this on the board.
What would change the note (desk-level)
- Two or more of the big buyers (MSFT, GOOGL, AMZN) walk back AI capex in the same week with demand language, not supply language.
- VRT or NOW print that explicitly severs the AI-infra / AI-software link the tape has been pricing.
- A single-name beat that the AI basket refuses to confirm the next session — then the note becomes a breadth note, not an earnings note.
If a note cannot say what would make it wrong, it does not get published.
How it is filed
Claim, evidence, and invalidation live with the other desk notes. We do not publish position sizes. We do not turn a blotter name into a recommendation. Post-print: confirm or retire each watch item in a follow-up once transcripts are in the research log (earnings transcripts API not live yet — deferred).
Not investment advice. Community research viewpoint only.