NOW › analyze
NOW · Analyze
Price: $118.14 · Market cap: $122.1B · Sector: Technology — Software (Application) · Beta 0.93 Verdict: 🟡 HOLD — conviction [7.0] (downgraded from [8.0] on 2026-03-29)
A portfolio-specific passage was removed from the public build.
Why this is a downgrade even though the stock went up
The March file was right about the business and right about the direction — the stock is +18.8% since. But its verdict rested on one explicit mitigating argument, and that argument has broken.
March 2026: "SBC is the single largest knock on the investment case… However, the trend is clearly improving — SBC/revenue down from 19.3% to 14.7%, SBC/FCF from 64.5% to 43.1%."
| Period | SBC $B | FCF $B | SBC / FCF | SBC / revenue |
|---|---|---|---|---|
| FY2022 | 1.40 | 2.17 | 64.6% | 19.3% |
| FY2023 | 1.60 | 2.70 | 59.4% | 17.9% |
| FY2024 | 1.75 | 3.38 | 51.7% | 15.9% |
| FY2025 | 1.96 | 4.53 | 43.2% | 14.7% |
| ttm Jun-26 | 2.18 | 4.57 | 47.7% ⬆ | 14.8% |
SBC grew +11.5% while FCF grew +0.9%. The four-year improvement stopped and reversed, because retention equity from the Moveworks and Veza acquisitions is now flowing through the P&L. Management's own 10-Q expects SBC to be "relatively flat as a percentage of revenue" in 2026 — i.e. the improvement is over for at least a year.
That is a thesis input, not a detail. It is the difference between "a compounder with a shrinking dilution problem" and "a compounder with a permanent one."
1. General Health (§1)
The business is genuinely executing — this must be said first
Q2 FY2026 (reported 2026-07-22) beat on every guided line:
| Metric | Q2 FY26 | |
|---|---|---|
| Subscription revenue | $3,877M, +24.5% | 150bps above the top of guidance |
| Total revenue | $3,987M, +24.0% | |
| cRPO | $13.20B, +21% | >2pts above guidance |
| RPO | $29.0B, +21% | |
| Renewal rate | 98% | best-in-class, unchanged |
| Customers >$5M ACV | 658, +23% | +32 crossed $20M |
| Deals >$1M net-new ACV | 123, +~40% | |
| AI ACV | crossed $1B, net-new +40% QoQ | 2026 target raised $1B → $1.5B |
| Non-GAAP operating margin | 29.5% | 300bps above guide |
FY26 subscription guidance was raised to $15.76–15.78B (+22.5%). A 98% renewal rate with 125% net expansion is not the signature of a business being disrupted.
🚩 But GAAP gross profit went backwards
| Quarter | Revenue | Gross profit | Gross margin | Op income | Op margin | D&A | SBC |
|---|---|---|---|---|---|---|---|
| Jun-25 | $3.22B | $2.49B | 77.5% | $0.36B | 11.1% | $0.172B | $0.50B |
| Sep-25 | $3.41B | $2.63B | 77.3% | $0.57B | 16.8% | $0.194B | $0.49B |
| Dec-25 | $3.57B | $2.73B | 76.6% | $0.44B | 12.4% | $0.212B | $0.49B |
| Mar-26 | $3.77B | $2.83B | 75.1% | $0.50B | 13.3% | $0.258B | $0.55B |
| Jun-26 | $3.99B | $2.82B | 70.7% | $0.16B | 4.1% | $0.407B | $0.65B |
Five consecutive quarters of gross-margin decline, −680bps, accelerating. In Q2 revenue rose +$217M while cost of revenue rose +$229M — gross profit fell on a 24%-growth quarter.
The 4.1% GAAP operating margin reconciles cleanly to the 29.5% non-GAAP figure:
| Q2 FY26 add-back | $M |
|---|---|
| Stock-based compensation | 655 |
| Amortization of purchased intangibles | 219 |
| Business-combination costs | 75 |
| Severance | 62 |
| Total | 1,011 |
$162M GAAP + $1,011M = $1,173M → 29.4%. So the collapse is not a mystery charge.
But the split matters enormously, and this is the central finding of the report. Roughly $356M is acquisition/restructuring (non-cash or one-time). The rest is not:
Management attributes the non-GAAP subscription gross margin decline — 83.0% → 80.5%, −250bps — to "increased customer utilization of hyperscaler partnerships and accelerated AI adoption."
That is a cash, structural, recurring cost. It does not add back and it does not go away. Professional-services gross margin swung from +14% to −14%. This is the same dynamic identified at Alphabet in analyze-2026-08-04.md: AI is an extension of revenue and an attack on the cost structure — except Alphabet owns TPUs and ServiceNow rents. The load-bearing word in management's own explanation is utilization: it scales with usage. Classic SaaS earned ~90% incremental gross margin because the marginal cost of a seat was zero. AI-delivered SaaS pays an inference bill per agent action — the revenue recurs and so does the cost, denominated in someone else's compute. SBC is non-cash; purchased-intangible amortization is non-cash; hyperscaler COGS is a wire transfer.
Per analysis_notes.md §2.1 — "Margin compression = moat erosion" — this is the quantitative moat base, and it is pointing the wrong way.
🚩 The most damaging single fact in this file: management's FY26 guide of 81% full-year subscription gross margin implies H2 prints below Q2's 80.5%. The company is guiding to further decline while describing the pressure as a short-term investment. Anyone holding the "transient" view is overriding the company's own forecast in the company's favour.
🚩 The FY26 free-cash-flow guide requires a large H2 acceleration
| FY2025 | FY2026 | |
|---|---|---|
| Q1 | $1.43B (derived) | $1.53B |
| Q2 | $0.53B | $0.47B |
| H1 total | $1.96B | $2.00B → +2.0% |
| Q3 | $0.57B | — |
| Q4 | $2.00B | — |
| H2 total | $2.57B | must be $3.75B → +46% |
| Full year | $4.53B | guide ~$5.75B (35% margin) → +27% |
H1 FCF grew +2.0% while revenue grew ~24%. To hit the full-year guide, H2 FCF must grow +46%. Q4 is genuinely ServiceNow's seasonal peak and the guide already absorbs a −200bps FCF-margin hit from Armis — but this is the single most checkable claim in the story, and it resolves at the Q3 print.
The valuation is highly sensitive to it:
| FY26 FCF outcome | SBC-adjusted FCF/share | Multiple at $118.14 |
|---|---|---|
| Guide met ($5.75B) | $3.12 | 37.8x |
| H2 grows +25% ($4.71B) | $2.61 | 45.3x |
| H2 flat ($4.57B) | $1.99 | 59.4x |
🚩 $11.6B of debt-funded M&A, and no organic bridge
| Target | Price | Status |
|---|---|---|
| Armis (OT/IoT security) | $7.75B | closed Q2 FY26 — largest ever |
| Moveworks (agentic AI) | $2.85B | closed 2025-12-15 |
| Veza (identity) | ~$1B | closed H1 2026 |
| Logik.io, data.world | ~$0.9B | 2025 |
Goodwill was $1.27B → $3.58B by end-FY2025, before Armis closed. Armis closed 2026-04-20 for $7.6B cash → $5.3B of goodwill (non-deductible) + $2.5B of intangibles — roughly 70% of the price is goodwill, and that $2.5B is the engine of the $219M/quarter amortization add-back. Interest expense went $6M → $66M in one quarter (~$264M annualized); $4.0B of senior notes were issued in May 2026 at 4.25–6.30%. Current ratio is now 0.70.
🚩 And the accounting is circular: spend cash on an acquisition → book the acquired revenue inside the headline growth rate → add back the amortization of the very thing you bought → report a higher non-GAAP margin. You count the revenue and not the cost. Do that three times in 24 months and non-GAAP EPS stops being an earnings measure.
Management quantified only Armis (~125bps of Q2 growth, ~175bps in Q3), implying ~21.75% organic constant-currency growth against 23% reported. Moveworks and Veza were never quantified. True organic growth is therefore below 21.75% and cannot be computed from public disclosure. For a company that spent $11.6B on acquisitions in 24 months while reporting "subscription growth," that is a deliberate accountability gap — and it is the second-most important disclosure omission in the file.
🚩 Buybacks paused at the lowest valuation ever
Zero shares repurchased in Q2 2026 — all $2.225B of YTD buyback was Q1. Cash went to the Armis close instead, at the stock's lowest-ever EV/EBITDA (~14x NTM). And even with $3.41B of ttm repurchase, the diluted share count still rises ~+0.75%/yr. ServiceNow spends more on buybacks than it books in SBC and still dilutes.
Balance sheet and returns
Net cash ~$1.3B, debt/assets 9.2% — still conservative, though weakening (current ratio 0.70, interest expense 11x'd). ROE 14%. ROIC on GAAP operating income: FY2022 4.4% → FY2023 6.7% → FY2024 9.8% → FY2025 10.1% — low for a business described as wide-moat, precisely because GAAP operating income is small.
2. Moat (§2) — WIDE, trend stable: core widening, perimeter eroding
The bear case, stated at full strength first
ServiceNow charges per fulfiller — the human who resolves or configures a record — at roughly $70–100/user/month. Requesters are free. So the billing unit is the population of humans doing service work. And management said on the Q2 call that Now Assist closes "80% to 85% of service requests without any human interaction."
ServiceNow sells the tool that destroys its own meter. Licensing consultancies now publish fulfiller-to-requester reclassification playbooks; one documented audit moved 220 of 600 help-desk users off fulfiller licences and cut the bill by $165K/yr. ServiceNow itself cut ~1,000 jobs (~3%) on 2026-07-30 — demonstrating the very headcount compression it tells customers not to worry about.
The bull rebuttal — and it is currently winning on evidence
ServiceNow has already re-cut the meter, and the transition is roughly half done:
- April 2026: five legacy tiers collapsed into three AI-native tiers. Now Assist, Moveworks, Workflow Data Fabric and AI Control Tower are bundled into every tier. Legacy SKUs went end-of-sale 2026-07-01.
- Price uplift on the new tiers is 20–30% (Pro Plus above 30%). A customer cutting fulfiller seats 20% and taking a 30% uplift is revenue-neutral to positive.
- A consumption layer sits on top ("Assist" pools — ~25 assists for a summary, ~150 for a large agentic action). Usage, not headcount, becomes the growth variable.
- "50% of our net new business is already non-seat-based." The most important sentence on the call.
- Active seats are still growing, renewal is 98%, and cRPO is +21%.
Moat sources
| Source | Rating | Evidence |
|---|---|---|
| Switching costs | Present — strong. The real moat. | Implementation runs $50K (SMB) to $4.5M+ for global estates. 658 customers >$5M ACV averaging ~$14.9M; 18 of the top 20 deals carry 8+ products. 98% renewal, ~125% net expansion. The moat analyst searched specifically for a documented large-enterprise rip-out of a multi-product ServiceNow estate and found none. |
| Cost advantage | Present — newly relevant | The 450-connector Workflow Data Fabric and 20 years of CMDB schema are genuinely expensive to reproduce. McDermott: "it will cost 5 to 10x to build an agent versus run one on ServiceNow" — self-serving, but directionally supported. |
| Intangibles / brand | Present | 20+ years of Gartner leadership; the default safe CIO choice, reinforced by audit/governance positioning. |
| Network effects | Weak | No true cross-customer network. The partner/SI ecosystem is a switching cost wearing a network-effect costume. |
| Efficient scale | Absent | Large, growing, openly contested market. |
The lock-in is not the ticketing workflow — that is commoditised and is exactly what Atlassian is taking. It is the CMDB as the system of record for enterprise assets, dependencies and entitlements, plus the compliance evidence trail on top. An AI-native startup can beat ServiceNow at resolving a password reset. It cannot reproduce a five-year audited change-management record.
Adversarial stress-test — where the attack is actually landing
| Attacker | Landing? |
|---|---|
| Atlassian (JSM up-market) | Landing hardest, with named evidence. CEO, April 2026: "our largest ever quarter for competitive displacements from a major ITSM provider." Service collection >$1B ARR, +30%. Taking mid-market and departmental, not $5M+ estates. |
| Microsoft (Copilot Studio + Power Platform in E5) | Largest latent threat, least current evidence. Microsoft currently ships ServiceNow connectors and hand-off patterns — it is building on top, not underneath. That is a choice, not a constraint, and it could reverse. |
| Salesforce (Agentforce) | Landing in the new CRM stream, not the core. Agentforce passed $1B ARR — a dead heat with NOW's AI ACV. Blocks expansion; doesn't threaten ITSM/ITOM. |
| AI-native startups (Serval, Atomicwork, Ravenna) | Landing in employee support and mid-market, priced per resolution. Notably picking off Moveworks accounts unsettled by the acquisition — NOW's own M&A is feeding them. |
Pattern: every attack lands at the perimeter and none lands in the core. That is a real moat under pressure, not a moat breaking.
Evergreen: conditionally. The need — a governed, audited record of what work happened to which asset — is evergreen and regulation makes it more so. Evergreen at the CMDB/governance layer; explicitly not evergreen at the ITSM ticketing layer.
3. Valuation (§3)
DYT/DDM are N/A (no dividend). Graham gives $21.14 on GAAP EPS of $1.64 — recorded and set aside; it is the wrong instrument for a 24%-growth platform whose GAAP earnings are suppressed by SBC and purchase amortization.
The multiple you quote depends entirely on how you treat SBC
| Basis | Figure | Multiple at $118.14 |
|---|---|---|
| GAAP ttm EPS | $1.64 | 72.0x |
| Consensus forward non-GAAP EPS | $5.01 | 23.6x ← the number everyone quotes |
| Forward non-GAAP EPS less SBC/share (~$2.40) | $2.61 | 45.3x |
| ttm FCF/share | $4.44 | 26.6x |
| ttm SBC-adjusted FCF/share | $2.32 | 50.9x |
| FY26 guided SBC-adjusted FCF/share | $3.12 | 37.8x |
The 23.6x and the 45.3x are the same company. SBC is a real transfer from shareholders to employees — the share count proves it, rising 0.75%/yr after $3.41B of buybacks.
The repo's standing test: ADBE passes at SBC = 19.7% of FCF; FSLY failed it. NOW is at 47.7%. On that test NOW sits much closer to the failing side, and the trend just reversed. (Mitigating: the CFO set a public, dated target of sub-10% of revenue by 2029 at the 2026 Analyst Day. Real commitment, three years out, and year one went the wrong way.)
Fair value — on SBC-adjusted free cash flow
Base: FY26 guided SBC-adjusted FCF of $3.12/share (assumes the guide is met).
| 3yr growth | Yr-3 SBC-adj FCF/sh | 25x | 30x | 35x | 40x |
|---|---|---|---|---|---|
| 18% | $5.13 | $128 | $154 | $180 | $205 |
| 22% | $5.67 | $142 | $170 | $198 | $227 |
| 26% | $6.25 | $156 | $187 | $219 | $250 |
Reverse test: to justify $118.14 at 30x SBC-adjusted FCF requires only +8.0%/yr of growth for three years — against a company currently growing subscription revenue 22.5%. At 35x it requires +2.6%/yr.
That is the strongest single argument for the stock, and it is why this is a HOLD rather than a TRIM. The price already discounts a severe deceleration.
Forward-modelled to management's own 2029 glidepath
The static 50.9x is unfair to a company with a published plan (sub-10% SBC by 2029, Rule of 40 >60). So model the plan at face value — deliberately, so the bull case gets its best shot:
| Revenue | FCF margin | FCF | SBC % rev | SBC-adj FCF | /share | Multiple @ $118.14 | |
|---|---|---|---|---|---|---|---|
| FY26E | $16.5B | 35% | $5.78B | 14.5% | $3.39B | $3.26 | 36x |
| FY27E | $19.8B | 36% | $7.13B | 13.0% | $4.56B | $4.35 | 27x |
| FY28E | $23.6B | 37% | $8.73B | 11.5% | $6.02B | $5.71 | 21x |
| FY29E | $27.6B | 38% | $10.5B | 10.0% | $7.72B | $7.27 | 16x |
Discounting FY29 SBC-adjusted FCF at 10% over ~3.4 years, net of ~$3.8B net debt:
| Exit multiple | Value/share |
|---|---|
| 30x (glidepath and gross-margin recovery both land) | $156 |
| 25x (base) | $130 |
| 20x (discipline) | $103 |
| Gross margin never recovers (FY29 FCF margin 32% not 38%, 22x) | $92 |
Probability-weighted (40% base / 35% impaired / 25% bull): ≈ $124.
Cross-check on non-GAAP EPS: FY26E $4.07 → 29.0x; FY27E $5.01 → 23.6x. NOW has traded 45–60x forward non-GAAP for most of five years — 23.6x is the cheapest it has ever been, and that is the bull's single strongest fact. At a fair 28–32x for a ~20% grower with a wide moat → $140–160.
Reconciling the two methods: the gap is entirely the add-backs. FY26 non-GAAP EPS of $4.07 adds back ~$2.4B of SBC plus ~$0.9B of purchased-intangible amortization ≈ $2.40/share after tax. In substance FY26 non-GAAP EPS is ~$1.65 of economic earnings plus ~$2.40 of add-backs. The EPS method is right about the direction of travel; the FCF method is right about today — and it already grants the glidepath full credit in its terminal value.
Fair value: $105 – $150, central estimate $130. At $118.14 the stock trades ~9% below the central estimate, in the lower half of the range. Not a bargain, not expensive. This is genuinely close and the report will not pretend otherwise.
4. Sentiment — fear on the stock, optimism on the business
These have decoupled, and that decoupling is the whole story. The business beat every guided metric, raised full-year guidance, raised the AI target, holds 98% renewal, and grew $1M+ deals ~40%. The stock is −39% off its high and was down ~38% YTD going into the print.
- Analysts: 44 of 49 positive, mean target $140.25 (+18.7%). But the tail is deteriorating — CLSA initiated Underperform with a $72 target (2026-07-20), Erste downgraded to Hold, and a first strong-sell appeared this month. Almost nobody has abandoned the fundamental thesis; nearly everybody has cut the target. The debate is about the multiple, not execution.
- The fear is specific and structural, not noise: the market is pricing a genuine risk that per-seat SaaS is the wrong model for the agentic era. The −250bps cash gross-margin compression is the first hard evidence that the bears are directionally onto something.
⚠️ Correction — the CEO bought $3.0M, not $20M
The watchlist recorded "CEO bought $20M." That figure is wrong. Verified from the Form 4:
William McDermott, 2026-02-27: 28,682 shares at a weighted-average $104.597 = $3,000,057. Filed 2026-03-02, executed under a non-Rule 10b5-1 arrangement adopted 2026-02-13.
Context that strengthens the signal even at the smaller size: ServiceNow executives cancelled their 10b5-1 plans at the same time the CEO committed to the purchase. The buy is currently +13% in the money.
🚩 Insider data — the Yahoo trap again
Yahoo reports 547,524 shares "purchased" across 48 transactions, +40% net. This is false. The "purchases" are almost entirely RSU vests and $0.00 stock awards (McDermott 120,020 on 2026-02-17; Zavery 41,074 on 2026-02-06), plus a stock gift. The only genuine open-market purchase in the visible history is McDermott's $3.0M.
Every other priced transaction is a sale — ~$4.4M in 2026 YTD, net open-market flow ≈ −$1.4M. Most notable: director Anita Sands sold 16,445 shares at $90.14 on 2026-05-14 — near the lows. A director selling into weakness is a genuine negative that Yahoo's "+40% net purchased" would have completely inverted.
This is the third confirmed instance today of the same Yahoo miscount (ADBE, AMZN, NOW) — see pitfall-yahoo-insider-purchases-counts-vests-and-gifts.
5. Verdict
🟡 HOLD — conviction [7.0], downgraded from [8.0]
Add on weakness below ~$105. Do not chase here.
The March file's 8.0 rested on two legs. Leg one — durable 20%+ growth — is intact and arguably stronger: AI ACV $1B with the target raised mid-year, cRPO two points above guide, 98% renewal, $1M+ deals +40%. Leg two — an improving cash-conversion story — broke in two places at once. SBC/FCF reversed 43.2% → 47.7% after four straight years of improvement, and subscription gross margin fell 250bps in cash — a phenomenon the March file never contemplated because it did not yet exist.
And the price moved the other way: the stock is +18.8% since that verdict. Both terms of a rating — business quality and price attractiveness — moved against the holder simultaneously. That is the textbook case for a downgrade. That the stock is −39% off its high is irrelevant; the relevant comparison is the price at which the prior verdict was struck (see principle-down-a-lot-is-not-cheap).
Why 7.0 and not lower. The moat evidence is genuinely strong and it is what holds the rating: 98% renewal with a 20–30% price uplift on the new AI tiers (you do not raise price 20–30% and hold 98% renewal in a business being commoditised), 50% of net new business already non-seat-based, and $1B of contracted AI ACV — the empirical refutation of the disruption thesis arrived while the stock fell 39%. Same tier as INTU. It was close between 7.0 and 7.5.
Why not ACCUMULATE. Three things: the stock is already +15% off its $102.50 pre-print level of 2026-07-20, so the catalyst is in the price; management's own FY26 subscription-GM guide implies H2 is worse than Q2, leaving the central bear claim unresolved by the company's own forecast; and Q3 decelerates against a Federal pull-forward it already absorbed. 9% upside to a central estimate is not enough margin of safety when the marginal gross margin printed negative.
A portfolio-specific passage was removed from the public build.
Break triggers (falsifiable, dated)
| # | Trigger | Date | Falsifies | Action |
|---|---|---|---|---|
| 1 | Non-GAAP subscription gross margin <80.0% (a 6th straight decline), or the FY26 sub-GM guide cut below 81% | Q3 FY26, ~2026-10-22 | "Transient investment" — the bull's load-bearing claim | Downgrade to 6.0, FV → ~$95 |
| 2 | cRPO growth <18.0% cc | Q4 FY26, late Jan 2027 (Q3 is contaminated by the Federal pull-forward — a Q3 miss is ambiguous, Q4 is not) | The demand leg carrying the entire multiple | TRIM — the wide-moat premium is unearned |
| 3 | FY26 SBC/revenue ≥14.5%, or diluted share count +>1.0% despite buybacks | FY26 10-K, ~Feb 2027 | The sub-10%-by-2029 glidepath, i.e. the terminal value | FV → ~$95, TRIM |
Secondary watch: the first quantified Moveworks + Veza revenue contribution. If still unbridged at the Q4 print, permanently haircut headline growth by 2–3pts. Also FY26 FCF margin — the 35% guide needs H2 FCF +46% against H1's +2.0%.
Upgrade condition, stated in advance: if Q3 prints subscription gross margin at or above 81% with cRPO ≥20% cc, move straight back to 7.5. The bull's argument — that a 20–30% price uplift and 98% renewal cannot coexist with genuine erosion — is a strong one. It simply has not been confirmed by a single quarter of data, and the one quarter we have shows gross profit going backwards.
The single number that would most change this answer
Non-GAAP subscription gross margin in Q3. It resolves whether the −250bps is an investment or the new business model, and everything else in the file is downstream of it.
Data-quality notes
- ✅ RESOLVED — the $272M "Other Non-Operating Income" in Q2 is a mark on private AI stakes. The 10-Q discloses $273M of net upward adjustments on non-marketable equity investments (measurement-alternative, triggered by observable price changes in later funding rounds). Other income, net was $206M vs −$3M a year earlier. Strip the mark and GAAP Q2 net income falls from $298M to roughly break-even on $3,987M of revenue.
This is the fourth confirmed instance today of the same mechanism — after AMAT, GOOGL and AMZN — and the first at a company this size. See
pitfall-unrealized-equity-marks-break-headline-pe. - 🚩 Yahoo's
insider_purchasesis false for NOW — see above. Third instance today. - 🚩 Yahoo's
operatingMarginsis the LATEST QUARTER, not ttm. NOW's reported 4.06% equals the Jun-2026 quarter exactly (4.01%); ttm is 11.3%. Verified against MU and AMZN as controls the same day. Do not mix this field with ttm figures. - 🚩 Yahoo's
freeCashflowfield for NOW ($5.15B) implies capex of only ~$160M against actual ttm capex of ~$730M from the statements. Use statement-derived FCF ($4.57B), not the summary field. Consistent withpitfall-roicai-fcf-field-returns-ocf— vendor FCF summary fields are unreliable across providers. - ⚠️ $272M of "Other Non-Operating Income" in Q2 FY26 (vs $88M in Q1, $3M a year earlier) carried GAAP net income to $298M. Nature unverified — likely a strategic-investment fair-value mark. Without it the GAAP quarter looks materially worse.
- ⚠️ Organic growth cannot be computed. Only Armis (~125bps) is quantified; Moveworks and Veza are not. True organic is below the implied 21.75%.
- ⚠️ Product-level revenue is not disclosed. All product mix is management anecdote or third-party estimate.
- ⚠️ Q2 subscription revenue was helped by US Federal on-premise revenue pulled forward from Q3.
- ⚠️ Sources conflict on the immediate post-earnings stock reaction; unresolved and immaterial to the verdict.
Sources
Yahoo Finance MCP (quarterly income statement, cash flow, balance sheet, recommendations, holder info) and .mcp/fin.py, retrieved 2026-08-04 · ServiceNow Q2 2026 press release · Q2 2026 call transcript (Benzinga) · Investing.com — Q2 slides, margins face AI pressure · McDermott Form 4, $3.0M purchase · Executives cancel 10b5-1 plans · Armis acquisition $7.75B · Moveworks close · $12B deal spree · Atlassian ITSM displacements (The Register) · Atlassian Q3 FY26 transcript · $30B by 2030 target · CLSA Underperform $72 · Layoffs July 2026 · Fulfiller vs requester licensing · Copilot Studio ServiceNow hand-off (Microsoft Learn) · SBC history (Macrotrends)
Related agency knowledge: pitfall-yahoo-insider-purchases-counts-vests-and-gifts · pitfall-roicai-fcf-field-returns-ocf · principle-primary-source-beats-vendor · prior file analyze-2026-03-29.md