Why are CRM, NOW, and WDAY all down sharply in 2026 even though they're growing?

Two reasons stacked. First, the 10-year Treasury at 4.40% and the FOMC's 8-4 hawkish hold on April 29 keep the discount rate on long-duration software valuations elevated. Second, generative AI threatens the per-seat licensing model that built each company. Both pressures hit valuation multiples regardless of operational performance — Salesforce, ServiceNow, and Workday all raised guidance into the selloff.

Which of the three has actually proven AI monetization?

Salesforce, by a wide margin. Agentforce ARR closed FY26 at $800 million, up 169% year over year, and Agentforce + Data 360 ARR hit $2.9 billion (+200%). ServiceNow does not disclose Now Assist ARR separately but reported Now Assist customers above $1 million ACV grew 130%+ in Q1 2026. Workday disclosed cumulative emerging-AI ARR above $400 million with $100 million of net new ACV in Q4 FY26.

What is the Rule of 40 and how do these three rank?

Rule of 40 is revenue growth percentage plus operating margin percentage — the SaaS industry's standard growth-and-profitability metric. On non-GAAP margins: ServiceNow leads at 54 (22% growth + ~32% non-GAAP op margin), Salesforce is at 44 (10% + 34.1%), and Workday lags at 39 (13% + ~26%). A score above 40 generally signals sustainable balance; below 40 typically commands lower multiples.

How much rate sensitivity does each stock have?

Highest to lowest: Workday > ServiceNow > Salesforce. WDAY's 47.3 P/E with sub-40 Rule of 40 means duration risk and growth-quality risk stack together. NOW's 52.6 P/E is duration-sensitive but cushioned by 54 Rule of 40. CRM at 22.6 P/E has the largest cushion — even a 10-year drift to 5.0% leaves the multiple defensible relative to S&P 500 averages.

Is the SaaS selloff over?

Probably not yet, but it's late stage. The April 29–30 FOMC + GDP/PCE combination was a worst-case macro print and the stocks held their lows. Sector EV/sales has compressed from 5.6x to roughly 4.2x. Goldman's framework now pays 1.1x EV/sales per 10-point Rule-of-40 improvement, up from 0.8x earlier in 2025 — meaning the market is starting to discriminate by quality rather than re-rating the entire sector down.

CRM vs NOW vs WDAY: Three SaaS Bets, Three Risks

ServiceNow closed April 30 at $88.31, down 58% from its post-split high of $211.48. Workday closed at $122.40, off 56% from $276. Salesforce got off comparatively easy at $176.53, down only 40% from its $296 peak. The three together have shed roughly $200 billion of market value in fifteen months while every single one of them grew revenue, raised guidance, and printed record cash flow.

The market is not stupid. It is pricing two real risks at once. The 8-4 FOMC hold on April 29 — the most divided FOMC vote since 1992 — kept the federal funds rate at 3.64% with the 10-year Treasury at 4.40% and Core PCE Q/Q SAAR re-accelerating to 4.3% in the BEA's April 30 advance estimate. That is a duration-unfriendly rate regime. On top of that, generative AI threatens to reprice the per-seat licensing model that built every one of these companies. Both threats are real. Neither is fatal.

The useful question is not which of these stocks is cheapest. It is which one is positioned to survive a stagflation-lite macro and an AI-capex arms race at the same time. By that test, the three are not interchangeable. Salesforce is generating $800 million in Agentforce ARR (up 169% year over year) at a 22.6 P/E. ServiceNow is selling AI workflow automation at a 52.6 P/E with 22% subscription growth and the deepest moat. Workday is the cheapest growth story but spends 27% of revenue on R&D into a market — back-office HR — that AI agents may genuinely commoditize. Three different stocks. Three different rate-regime risks. Three different bets.

Revenue Scale and Growth: The Gap Between $41B and $9.5B Matters More Than It Looks

Start with the numbers each company actually reported, not the ones the market is pricing.

Salesforce closed fiscal 2026 (year ended January 31, 2026) with $41.5 billion in revenue, up 10% year over year, and $11.2 billion in Q4 alone, up 12%. Management guided fiscal 2027 to $45.8–$46.2 billion, another 10–11% growth print. ServiceNow reported Q1 2026 (calendar) on April 22 — subscription revenue of $3.67 billion, up 22% year over year, with current remaining performance obligations of $12.64 billion (+22.5% YoY) and total RPO of $27.7 billion (+25%). Workday closed fiscal 2026 (year ended January 31, 2026) on February 24 with $9.55 billion in revenue, up 13%, including a Q4 of $2.53 billion (+14.5%). Q1 FY27 reports May 21.

The scale gap matters because of what it implies for AI-product economics. Salesforce already has 150,000+ enterprise customers paying for CRM. Layering Agentforce on top — at $2 per conversation or seat-based add-ons — converts an existing distribution moat into AI revenue at near-zero customer-acquisition cost. ServiceNow has roughly 8,400 enterprise customers but every one of them sits on workflow infrastructure that is far stickier than CRM. Workday has approximately 11,000 customers concentrated in HCM and finance — the back-office functions where AI agents are most plausibly substitutive, not complementary.

Growth is not slowing the way the share-price moves suggest. ServiceNow's 22% subscription growth at a $14.7-billion run-rate is the most impressive number on this page. Salesforce's 10% growth on a $41.5-billion base is roughly $4 billion of incremental revenue per year — more than Workday's entire fiscal-2026 net income twenty times over. Workday's 13% growth is decelerating to a 12–13% guide for FY27, which is the only one of the three trending the wrong way.

Profitability: Salesforce Has Crossed the Profitability Chasm. The Other Two Are Still Investing.

Operating margins separate these companies more cleanly than valuations do.

Salesforce printed a 20.1% GAAP operating margin and 34.1% non-GAAP operating margin for FY26. Operating cash flow hit $15.0 billion, up 15%. Net income was approximately $7.5 billion. Activist pressure from Elliott, Starboard, and ValueAct in 2023–2024 reset the cost base, and the company has held the discipline. ServiceNow's GAAP operating margin runs around 12–13%; its non-GAAP figure is closer to 30%. Workday is at roughly 12% GAAP and 26% non-GAAP. Both NOW and WDAY are still in build-mode profitability, prioritizing R&D investment over margin expansion.

The quiet number is stock-based compensation. Workday's SBC ran 15.8% of revenue, ServiceNow's 13.9%, Salesforce's 9.7%. SBC is real shareholder dilution dressed up as a non-cash expense. On Workday's $9.55 billion revenue base, 15.8% SBC is roughly $1.5 billion of annual dilution against $693 million of GAAP net income — meaning shareholders are funding more than half of the dilution out of every dollar of reported earnings.

This is the post-2022-rate-shock SaaS playbook. Companies that cut to print profit got rewarded (CRM, MSFT, ORCL). Companies still spending to grow (NOW, WDAY) face a much harder valuation environment when the 10-year sits at 4.40% and the discount rate on 2030 free cash flow keeps drifting up. Salesforce already paid the discipline tax. Its peers have not.

AI Monetization vs. AI Capex: Where the Market Is Wrong on All Three

The clearest signal that AI-disruption fears are mispriced is that one of these three companies is already shipping disclosed AI revenue at scale — and the market is not crediting it.

Salesforce closed FY26 with $800 million in Agentforce ARR, up 169% year over year. Combined Agentforce and Data 360 ARR hit $2.9 billion (+200% YoY). The company closed 29,000 Agentforce deals in Q4, up 50% sequentially, and disclosed that more than 60% of Q4 bookings came from existing-customer expansion. That last number matters: it means Agentforce is not a customer-acquisition motion. It is a unit-economics-positive upsell on top of seats Salesforce already has. The capex required to monetize the next dollar is essentially zero.

ServiceNow does not disclose Now Assist ARR directly, which is itself informative. It does disclose that customers spending more than $1 million in annual contract value on Now Assist grew over 130% year over year in Q1 2026, and that 36 deals included five or more Now Assist products. Subscription growth accelerated from 21% in Q4 2025 to 22% in Q1 2026 — accelerating into the AI cycle, not decelerating. Management raised FY26 subscription guidance by $205 million to $15.735–$15.775 billion. The company is selling AI without yet quantifying its ARR contribution because — by ServiceNow's own admission — Now Assist is being deployed inside existing workflow contracts rather than as a separate SKU.

Workday is the contradiction. It generated more than $100 million of new annual contract value from emerging AI products in Q4, growing over 100% year over year, with cumulative AI-product ARR above $400 million. That is real. But Workday also spends 27.3% of revenue on R&D — the highest of the three by a wide margin — and FY27 subscription guidance decelerated to 12–13% growth from 14.5% in FY26. The arithmetic is: Workday is investing the most aggressively in AI, monetizing the least relative to its growth profile, and seeing the steepest growth deceleration. That is not a universally positive story.

The market is treating all three as victims of AI commoditization. The data says: Salesforce is monetizing AI better than any enterprise-software peer, ServiceNow is selling AI bundled into a workflow moat that AI is more likely to deepen than erode, and Workday is the one where the capex-vs-monetization math actually deserves scrutiny.

Rate-Regime Sensitivity at 10Y 4.40%: The Hidden Driver Behind the Selloff

Long-duration software is the most rate-sensitive equity exposure in the S&P 500, and the post-FOMC tape on April 29–30 confirmed it.

The 10-year Treasury closed at 4.42% on April 29 — the highest yield since the Iran-war re-escalation cycle began in mid-April. The 2-year sat at 3.92%, the 30-year at 4.98%. The fed funds rate stayed at 3.64% but the FOMC voted 8-4 to hold — the most divided vote since 1992, with three governors dissenting for cuts and one (Miran) dissenting for a 25bp cut. The market's read: a hawkish hold dressed in dovish dissent geometry. The next day's Q1 GDP advance estimate (Real GDP +2.0% vs +2.3% expected) printed alongside Core PCE Q/Q SAAR of 4.3% — the highest non-COVID quarterly inflation print in two years.

That is a stagflation-lite combination, and it is bad for long-duration software in three specific ways:

Discount rate. Every SaaS DCF model values out-year free cash flow heavily. A 10-year at 4.40% versus the 3.5–4.0% range that prevailed when these stocks set their 52-week highs implies roughly 15–20% downside to fair value before any operational concerns. NOW at 52.6x earnings and WDAY at 47.3x are pricing forward growth at multiples that require a discount-rate environment that no longer exists.

Customer-side capex. Higher rates compress IT-budget growth at Workday and ServiceNow's customer base — large-cap enterprises and federal contractors. Both companies have flagged subscription headwinds tied to public-sector deal slippage and Iran-war spending diversion. ServiceNow's stock fell 14% on Q1 print despite raising guidance because subscription revenue did take a hit from federal sequencing. Workday flagged DIA-contract non-continuation as a Q1 FY27 headwind.

Real rates. Fed funds at 3.64% against headline CPI of 3.3% YoY (CPI index 330.293 in March 2026 versus 319.785 in March 2025) means real rates of roughly +30–40bps — narrowly restrictive territory and the first sustained stretch since 2007. That is precisely the rate regime under which growth-tech historically de-rates and value-tech outperforms.

The rate-regime test sorts the three cleanly. Salesforce's 22.6 P/E gives it the largest cushion — a 10-year at 5.0% would still leave CRM cheap. ServiceNow's 52.6 P/E means a 50bp move in the 10-year carries roughly 8–10% multiple compression even before any earnings adjustment. Workday is the most exposed: its 47.3 P/E is paired with the slowest growth and the largest SBC dilution, so a hawkish-hold extension into 2026 H2 directly threatens the multiple.

Valuation: P/E, Rule of 40, and Why the Market Has Picked a Lazy Yardstick

Comparing P/E ratios alone is the wrong move for SaaS. The right metric is the Rule of 40 — revenue growth plus operating margin — which captures the growth-vs-profitability trade-off these companies are managed against.

At non-GAAP operating margins, the 2026 Rule of 40 looks like this:

Salesforce: 10% revenue growth + 34.1% non-GAAP op margin = 44
ServiceNow: 22% subscription growth + ~32% non-GAAP op margin = 54
Workday: 13% revenue growth + ~26% non-GAAP op margin = 39

That re-orders the comparison. ServiceNow leads, Salesforce is well clear of the 40 threshold, Workday is below it. Goldman Sachs's most recent enterprise-software framework noted that each 10-point improvement in Rule of 40 was historically associated with roughly 1.1x increase in EV/Revenue multiple, up sharply from 0.8x earlier in 2025 — meaning the market is paying more for growth-quality at the margin than it did six months ago. By that test, NOW's 52.6 P/E is partly justified, CRM is meaningfully cheap, WDAY is the outlier on the wrong side.

On EV/sales, the sector average has compressed from 5.6x at the end of 2025 to roughly 4.2x by mid-March 2026. Salesforce now trades around 3.5x EV/sales (cheap to the sector), ServiceNow around 6.3x (expensive but defensible given Rule-of-40 leadership), Workday around 3.3x (cheap to sector but with the worst Rule-of-40 score). Workday cheap-to-sector with sub-40 Rule-of-40 is exactly what a value trap looks like in software.

The simpler framing: pay for what you get. Salesforce gives you 44 Rule-of-40 quality at a 22.6 P/E and the best disclosed AI ARR in enterprise software. ServiceNow gives you 54 at 52.6x — a fair price for premium growth at scale. Workday gives you 39 at 47.3x — paying premium-growth multiples for sub-threshold growth quality.

Cash Flow and Capital Returns: Salesforce Is Now Buying Back Stock at 22 P/E

Cash conversion is what turns AI ARR into shareholder returns, and the gap between these three is wider than the income statements suggest.

Salesforce produced $15.0 billion in operating cash flow in FY26 (+15% YoY) on $41.5 billion of revenue — a 36.1% OCF-to-sales ratio. That is the highest absolute cash-flow figure in enterprise software outside Microsoft and Oracle. Salesforce returned approximately $9.3 billion to shareholders in FY26 through dividends and buybacks, with the buyback running aggressively since the stock dropped through $200. Buying back 22-P/E stock with cash flow growing 15% is precisely the capital-return profile that should attract value capital — and the fact that the share price keeps drifting lower despite that activity tells you the marginal seller is rate-driven, not fundamentals-driven.

ServiceNow's OCF-to-sales runs at the top of the industry, with management guiding free-cash-flow margin around 32% for 2026. The board authorized an additional $5 billion buyback in February 2026 — but ServiceNow is not buying back at the pace Salesforce is, partly because management views the current price as a temporary reset rather than a structural opportunity. At a 52.6 P/E, every dollar of buyback reduces share count less than at CRM, so the capital-return arithmetic is genuinely worse even with stronger conversion.

Workday's free-cash-flow margin is around 30% on a non-GAAP basis. The company holds roughly $20.83 per share in cash and has authorized buybacks, but the SBC drag eats most of the optical share-count reduction. On Workday's $9.55 billion revenue base, 15.8% SBC (~$1.5 billion) plus modest buybacks has produced essentially flat shares-outstanding for two years — the buyback is a SBC-offset program, not a true capital return.

The practical implication: at current prices, Salesforce's buyback is the only one of the three that's genuinely accretive to per-share intrinsic value. ServiceNow's buyback is defensive. Workday's is offsetting dilution. Three very different shareholder-yield profiles inside what looks like one trade.

AI Strategy and Competitive Moats: Three Different Bets on the Agentic Enterprise

All three companies have AI strategies. The question is which moat AI deepens versus which one it threatens.

Salesforce — Distribution-First. Agentforce sits on top of an installed base of 150,000+ customers. The company can deploy agentic AI as a feature inside existing seat-based contracts, then upsell to consumption pricing as agent-call volume grows. The $800 million Agentforce ARR is direct evidence of monetization, but the more important number is 50% sequential deal growth in Q4 — which means the sales motion is working. Marc Benioff has framed this as the biggest product cycle in Salesforce's history. The risk: if AI agents truly disintermediate CRM workflows (a real but lower-probability scenario), Salesforce's distribution advantage erodes. The base case is that distribution wins.

ServiceNow — Workflow Lock-In. ServiceNow is the system of record for IT-service management at most large enterprises, and Now Assist embeds AI inside that workflow rather than alongside it. The metric to watch is Now Assist customers spending more than $1 million annually — up 130% YoY, with 36 deals including five or more products. That is consolidation behavior: customers are buying broader Now Assist surface area, not narrower. ServiceNow's moat is the deepest of the three because workflow-orchestration is exactly the layer where AI agents need a system of record, not a substitute for one. The risk: NOW's premium multiple leaves no room for a single bad quarter; the 14% drop on Q1 print despite a guidance raise was the proof point.

Workday — The Squeeze. Workday's HCM and financial-management platform is deeply embedded but operates in the part of enterprise software most plausibly disrupted by AI agents. HR workflows — performance reviews, scheduling, payroll exception handling — are exactly the use cases where an AI agent can plausibly replace a chunk of Workday-licensed work rather than augment it. Workday's response is to spend 27.3% of revenue on R&D — the most aggressive AI-investment ratio of the three — and to ship products like Sana Core and Sana Enterprise that position conversational AI as a new interaction layer. It is the right strategy. It is also a very expensive insurance policy on a moat that may genuinely narrow.

For the broader AI-capex thesis, see also our analyses of Microsoft's 29% pullback, Adobe at 7-year lows, and Alphabet's AI-capex bet paying off.

Three Stocks, Three Investor Profiles: A Framework for Decision-Making

The right answer is not to pick one. It is to match the rate-regime you expect to the stock that best survives it.

If you believe rates stay in the 4.0–4.5% 10Y range through 2026 H2 — the base case under the FOMC's hawkish 8-4 hold and the BEA's 4.3% Core PCE print — Salesforce is the cleanest expression. The 22.6 P/E gives it the most room for multiple compression already absorbed. The $800 million Agentforce ARR validates the AI thesis. The $15 billion of operating cash flow funds an aggressive buyback at attractive prices. CRM is the stagflation-lite survivor pick.

If you believe Miran's 25bp dissent presages a cut cycle in 2026 H2 — meaning the 10-year rallies back toward 3.75–4.00% — ServiceNow is the highest-beta way to express that view. The 52.6 P/E expands meaningfully on rate relief, and the 22% subscription growth and 54 Rule-of-40 score justify premium pricing as the macro overhang fades. NOW's 58% drawdown from $211.48 has created a rare entry into a best-in-class platform asset; the price is paying for the rate risk, not for the business.

Workday is the contrarian deep-value bet most likely to underperform either path. At 47.3 P/E with sub-40 Rule-of-40 and 27.3% R&D spend, WDAY needs both rate relief and successful AI re-platforming to justify its multiple. Either alone is not enough. The 56% drawdown looks compelling, but the AI-disruption case is most legitimate in HR/finance back-office workflows, and the SBC drag means even a successful transition leaves shareholders funding the survival cost. The right way to play WDAY is half-position, hedged against either failed AI execution or rate-cut delay — and that is too complex an expression for most portfolios.

For portfolio context: see our work on valuing stocks with P/B ratios, accretion/dilution analysis for SaaS M&A, and why dollar-cost averaging beats lump-sum entries in expensive markets.

Conclusion

Three enterprise-software stocks. Three completely different rate-regime risks. The market is treating them as one trade.

Salesforce at $176.53 is the clearest survivor pick: 22.6 P/E, $15 billion in operating cash flow, $800 million in Agentforce ARR validating the AI-monetization thesis, and the largest distribution moat in enterprise software. The 40% drawdown has already absorbed most of the rate-regime stress. ServiceNow at $88.31 is the premium-quality compounder priced for a very long duration of high rates — buy it if you believe the FOMC's four dovish dissents matter more than the four hawkish holds. Workday at $122.40 is the trap: cheapest to its own history, but stuck on the wrong side of the AI-capex-vs-monetization split, with a 47.3 P/E it cannot justify on a sub-40 Rule of 40 score and a back-office moat that is genuinely vulnerable.

The one trade that doesn't work is buying all three equally. The market is mispricing the rate-regime sensitivity that separates them. CRM is the position. NOW is the call option on rate cuts. WDAY is the one to leave alone.