KEYS

Current Thesis

Picks-and-shovels AI-infrastructure validation play. Every 800G/1.6T optical port, every GB300 / MI355X / TPU-v7 / MTIA-2 die, every PCIe Gen6 / CXL fabric in a hyperscaler buildout has to be validated on Keysight (or Anritsu / Spirent — and KEYS bought Spirent, deal closed 2025). Goldman raised PT to $384 on 2026-04-14, which is the first top-5 sell-side desk to publicly catch up to the AI-test narrative we want to be already long into. Trade is binary on the Q2 FY26 print (~2026-05-21): commentary needs to confirm 1.6T optical test orders are now a discrete material line item — not a sentence buried in Q&A. Sell-side median PT is still ~$200, meaning the upgrade cycle has 12-18 months of catch-up to do if Q2 confirms.

Bull Case

• Hyperscaler capex 2026 tracks $400B+ vs $325B 2025 — Microsoft, Meta, Amazon, Google calls Jan-Apr 2026 all reaffirmed or raised. Every dollar of GPU spend pulls ~1.5–3% of test/validation spend through KEYS / ANSS / CDNS. KEYS is the only listed pure-play with optical + digital + RF test under one roof.
• Spirent integration is the unlock. Goldman's 2026-04-14 PT raise to $384 explicitly cited Spirent cross-sell into hyperscaler network test (800G/1.6T Ethernet, AI fabric validation) as the Q2-Q4 FY26 catalyst. This is new — KEYS never had this revenue lever in prior cycles.
• Order book turning. Q1 FY26 print (2026-02-18) showed book-to-bill back >1.0 after a 5-quarter trough; commercial-comm (datacenter test) called out as "double-digit growth" segment by management. Sequential reacceleration is the signal that matters more than absolute level.
• **Goldman to $384 on 2026-04-14 is the first upgrade**, not the last. Median sell-side target ~$200 means the analyst herd still has to catch up. Catching narrative-velocity 1-3 weeks before broad upgrade-cycle is the entire edge.
• Custom-silicon proliferation. Every hyperscaler ASIC (TPU-v7, MTIA-2, Trainium-3, MAIA, Athena) requires bring-up validation that NVDA reference designs do not. KEYS sells the rack at first-silicon for each new chip family. As internal-silicon share rises against merchant GPU, KEYS captures more test dollars per dollar of compute.
• AI-test moat is non-trivial. Compliance suites for IEEE 802.3df (1.6T) and PCIe Gen6 take 18-24 months to develop. KEYS owns the de-facto reference suites; Anritsu lags ~12 months on optical, Tektronix retreated from high-speed digital after the Fortive carve-out.

Bear Case

• Premium multiple, low velocity. KEYS trades ~28x fwd vs 5y avg ~22x. Not stretched-meme, but zero margin for a guidance miss. A6 dynamics don't apply (this is a2), but P/E compression on an in-line print is the real risk.
• Lumpy capex exposure. ~35% of revenue is hyperscaler / telco / aerospace-defense — three customer cohorts that can pause orders for 1-2 quarters without warning. The 2024 cycle bottom proved this; book-to-bill went 0.85 → 1.05 in a single quarter and reversed in two.
• 5G/6G stall. Telco-test (~25% of revenue) has been flat-to-down for 6 straight quarters. If Q2 FY26 doesn't show telco re-acceleration, the AI-test bull case has to carry the entire P&L alone — and Spirent doesn't help on the telco line.
• Sell-side moving means we're not first. Goldman to $384 (2026-04-14) is the upgrade catalyst, but it also means we missed the earliest window. If Q2 is in-line rather than beat-and-raise, the upgrade cycle compresses into a single week and the trade is fully priced.
• Broad-market beta. Picks-and-shovels still trades like a cyclical — SPX -10% drags KEYS -15% even on perfect fundamentals. Macro regime is the dominant factor on any 5-day horizon.
• Commercial-comm vs aero/defense mix. If aero/defense over-indexes on the print and commercial-comm misses, the headline is fine but the AI-test thesis is broken even on a beat.

Setup & Price Structure

• No live price context delivered with this dossier — operator must re-pin EMA / RSI levels against live quote at entry.
• Anchor: Goldman PT $384 issued 2026-04-14 implied ~15-20% upside from spot at the time (i.e. spot was ~$320-$330 region on 2026-04-14).
• Watch levels — re-pin at entry:
• Above weekly 20-EMA = trend intact, accumulate on dips toward EMA
• Weekly close below 20-EMA on >1.5x avg volume = first warning; cut size by half
• Weekly close below 50-EMA = thesis broken; full exit, no averaging down
• Volume signature to confirm post-print: rising 50-day OBV through Q2; if 2026-05-21 print is met with gap-and-fade on heavy volume = distribution = narrative is fully priced and we sell into the upgrade-cycle pop.
• RSI guard (archetype-2): start trimming on weekly RSI >78; full review on weekly RSI >85. Daily RSI >80 just means short-term froth — do not trim on daily alone for a2.
• Earnings blackout: do NOT initiate within 3 trading days of 2026-05-21. If we miss the pre-print window, defer to clean post-print setup (gap-up + first 5-day consolidation = re-entry, gap-down = thesis re-eval, not buy-the-dip).

Catalyst Calendar (next 30 days)

• 2026-05-21 (est.) — Q2 FY26 earnings print. The binary. Watch: total revenue vs consensus ~$1.32B, commercial-comm segment YoY growth (need >8% YoY, double-digit confirms), explicit Spirent revenue contribution callout, book-to-bill commentary (need >1.0 sustained). Management tone on 1.6T optical orders is the qualitative tell.
• 2026-05-12 to 2026-05-19 — Pre-print sell-side noise window. Watch for second top-5 desk PT raise (Morgan Stanley / JPM / BofA most likely candidates given Goldman's lead). A second upgrade in this window = sizing-up signal.
• 2026-05-19 to 2026-05-23 — Hyperscaler capex tape from MSFT/META/GOOGL/AMZN reiterations at industry conferences. Any cut to 2026 capex guide is a thesis kill regardless of KEYS print.
• OFC / Hot Chips read-throughs ongoing through May. Optical-networking customer commentary (Coherent, Lumentum, MRVL, AVGO) is the leading indicator on 1.6T order timing.

What Would Change Our Mind

• Bullish (size up to HIGH): Q2 FY26 revenue >$1.38B AND commercial-comm growth >12% YoY AND explicit 1.6T optical test order callout AND book-to-bill >1.05 → adds 2-3% to position toward 5% cap. Second top-5 desk PT raise pre-print = same signal, ~1% earlier add.
• Bearish (full exit, no debate): Q2 revenue <$1.30B OR commercial-comm growth <8% YoY OR weekly close below 20-EMA on >1.5x avg volume OR any of MSFT/META/GOOGL/AMZN cuts 2026 capex guide. Each of these alone is enough to flip the trade — do not wait for confirmation, do not average down.
• Defer (DEFER, revisit): Q2 print in-line (revenue $1.30-1.36B, commercial-comm 8-12%) without explicit Spirent or 1.6T callout = thesis intact but velocity uncertain. Wait for next sell-side upgrade or order-book commentary at 2026 Investor Day.

Correlation Notes

• Direct co-trades (correlated, NOT diversifying): ANET (1.6T optical client), ANSS (simulation/validation peer), CDNS (digital design test peer), SPT (the acquired Spirent assets — no longer listed but track via KEYS segment disclosure). Stacking KEYS + ANET + ANSS + CDNS is one bet, not four.
• Hyperscaler capex tape: MSFT, META, GOOGL, AMZN earnings + capex guidance directly drive KEYS order book on a 2-quarter lag. Their commentary is leading; KEYS print is lagging.
• Optical components leading indicators: COHR, LITE, AVGO, MRVL — when they pre-announce 1.6T order strength, KEYS test-equipment orders follow within 1-2 quarters.
• Inverse / rotation pair: Defensive-staples basket (XLP, KO, PG). KEYS underperforms severely on any flight-to-quality regime — size down on VIX >25 sustained.
• Macro overlay: 10Y yields >5% sustained = multiple compression risk on KEYS premium fwd P/E. Watch alongside the chart, not in isolation.