CRM · Analyze — Financebot

A portfolio-specific passage was removed from the public build.

The event. FQ2 FY2027 (quarter ended 2026-07-31), reported 2026-08-26 after close: revenue $11.345B, +10.8% YoY, adjusted EPS $5.90 (beat consensus by ~80%), GAAP net income $4.844B, +73.3% YoY, non-GAAP operating margin 34.1%, GAAP operating margin 20.5%. cRPO +14% cc to $33.5B — a point ahead of guidance and, per management, the strongest net-new annual order value in four years. Agentforce ARR crossed $1.5B, +240% YoY (3.2B agentic work units in the quarter, +97% sequentially; 6,000+ paying Agentforce customers). FY27 revenue guide raised to $46.1-46.4B ($100M organic, $200M from the pending Contentful/Fin deals) and FY27 non-GAAP EPS guide raised to $16.67-16.71, from $14.06-14.12. Stock closed +22.6% on the day, +21.5% on the week, and sits at $252.05, still 6.3% below its 52-week high of $269.11.

1. Health (Fundamentals)

Metric	Value	Note
Revenue TTM (sum of quarters)	~$42.8B	Q3FY26 $10.26B + Q4FY26 $11.20B + Q1FY27 $9.83B... (see reconciliation below)
Revenue growth (FQ2 YoY)	+10.8%	Accelerating vs. FY26 full-year +9.5%
Revenue CAGR (3yr, annual)	9.8%	fin.py, FY2023→FY2026
Net income CAGR (3yr)	229.7%	Distorted — FY2023 net income was $208M (near-zero base); not a clean trend signal
FCF TTM (sum of 4 quarters)	~$15.16B	Q3FY26 $2.18B + Q4FY26 $5.32B + Q1FY27 $6.56B + Q2FY27 ~$1.10B (reported +81% YoY vs. $0.605B prior-year Q2)
FCF CAGR (3yr, annual)	31.6%	fin.py — flattered by margin expansion off a low FY2023 base
FY27 FCF growth guide	cut from 9-10% to 4-5%	Management's own words: cut "to reflect the cost of that debt" (see leverage below)
Gross margin	77.6-78%	Stable, high-quality software economics
Operating margin (GAAP / non-GAAP)	20.5% / 34.1% (FQ2)	Wide GAAP/non-GAAP gap — SBC-heavy comp structure
Debt/Assets (latest annual, FY26)	15.3%	Stale — see leverage event below
Total debt	$17.18B (Jan-2026) → $41.88B (Apr-2026)	+$24.7B in one quarter
Net debt	$7.11B (Jan-2026) → $30.35B (Apr-2026)
Stockholders' equity	$59.14B (Jan-2026) → $34.24B (Apr-2026)	−42%, from the ASR
Tangible book value	−$5.6B (Jan-2026) → −$31.7B (Apr-2026)	Negative and worsening — Informatica goodwill + ASR
Shares outstanding (ordinary)	929M (Jan-2026) → 819M (Apr-2026)	−11.8% in one quarter, −10% YoY
Dividend	$1.76/sh, 0.86% yield, 20% payout	New (started FY25), well covered

The dominant Health event is not the earnings beat — it is a debt-funded, $25B accelerated share repurchase that closed in March 2026, the largest ASR in software history, funded by a $25B senior-notes issuance (maturities 2028-2066) plus a $6.0B term loan. In FQ1 FY27 alone, Salesforce delivered ~103M shares (~80% of the total buyback) and returned $27.5B to shareholders in a single quarter ($27.1B buybacks + $365M dividends). Diluted share count fell 10% YoY — a real per-share tailwind, but mechanically financed, not FCF-funded. Management explicitly cut its own FY27 FCF-growth guide from 9-10% to 4-5% to absorb the new interest burden (interest expense went from ~$68-70M/quarter to $317M/quarter in the data available). Net debt is now $30B+, D/A jumped from 15% to an estimated ~39% on the April balance sheet, and tangible book value is deeply negative (Informatica goodwill layered on top).

This is a genuine change in capital-allocation posture, and it predates the FQ2 print by one quarter — the FQ2 balance sheet had not yet propagated into Yahoo's feed as of this writing (latest quarterly balance-sheet/income-statement data available is FQ1 FY27, ended 2026-04-30; the FQ2 print is captured here only through the press release and secondary sources, not the raw statements). Re-verify the FQ2 (Jul-2026) balance sheet at the next touch — confirm whether net debt is stabilizing or still climbing, and whether the credit rating held.

Capital allocation of FCF (last 4 reported quarters): overwhelmingly buybacks (~$27B in one quarter alone, debt-funded) and one large acquisition (Informatica, ~$8.2B in FQ4 FY26, plus smaller deals). Organic FCF itself remains strong and growing (~$15.2B TTM), but the balance sheet took on real leverage to fund the return-of-capital program on top of, not instead of, organic FCF.

2. Moat & Competitive Advantage

Quantitative base. Gross margin stable at 77-78% for years; ROE 16.9% (TTM); ROIC is structurally suppressed by ~$59B+ of acquired goodwill/intangibles (Informatica, Slack, Tableau, MuleSoft history) sitting on the balance sheet — the reported return numbers understate organic returns on invested capital but confirm the moat has been bought as much as built.
Moat source: CRM data gravity. Salesforce holds an estimated 43.6% share of the CRM market (~$25B of category revenue). Agentforce's differentiation is that its agents reason over decades of a customer's own interaction, opportunity and case history inside the same data model — a platform-agnostic agent cannot replicate this without replicating the CRM itself. This is a real, structural switching-cost moat, not a marketing claim.
Adversarial stress-test. The two live attackers are asymmetric:
Microsoft Copilot attacks on distribution, not depth — 30M+ Copilot seats bundled into Office/365 relationships Salesforce doesn't own. Microsoft doesn't need CRM-specific data depth to win the low end of agentic workflows if it's already inside the account.
ServiceNow attacks on platform velocity — its AI-platform revenue grew at a ~182.5% CAGR (CY2022-25) vs. Salesforce's ~92.8%, i.e. roughly double the growth rate, from ITSM outward into adjacent workflows. ServiceNow is smaller in absolute CRM terms but compounding the "agentic platform" category faster.
A well-funded rival cannot easily replicate the CRM-native data model quickly, but can commoditize the seat-based pricing layer* — and management conceded on the call that Salesforce itself is "still trapped in some ways in old per-user pricing models" for headless/agentic configurations. That is a live pricing-model vulnerability, not a hypothetical one.
Revenue-stream map. Core CRM clouds (Sales/Service/Marketing/Commerce) — mature, durable, high switching cost. Data 360/MuleSoft — integration layer, durable. Slack — collaboration, moderate durability, exposed to Teams bundling. Agentforce — the highest-growth, least-proven stream; ARR of $1.5B against $46B+ total revenue is still ~3%, i.e. the growth narrative is real but not yet financially dominant.
Disruption forecast. Near-term (1-2yr): pricing-model transition risk as agentic/headless usage grows faster than per-seat monetization keeps up. Medium-term (5-10yr): the CRM data moat should hold against pure AI-native entrants (an LLM alone cannot replicate 25 years of a specific enterprise's case history), but is genuinely contestable by Microsoft's distribution advantage and ServiceNow's platform-growth rate if Salesforce doesn't out-execute on agentic pricing.
Evergreen assessment. Yes, with a caveat: the core CRM business is a forever business by switching-cost economics. The Agentforce layer is not yet proven to carry the same durability — it is being built on top of a still-unresolved pricing model.
Moat-defense evaluation. Management is defending the CRM-data thesis with real product motion (Agentforce Operations, Data 360, Claudeforce with Anthropic launching Sept 2026) and real bookings evidence (cRPO acceleration, record net-new AOV). The open question the market hasn't fully priced is whether the pricing-model transition captures the value Salesforce is creating, or lets a distribution-advantaged rival capture it instead.

3. Valuation

Data trap encountered before any model could run — see Data Traps below. The vendor forwardPE (15.98x) and epsCurrentYear (14.67, implying 17.18x) disagree with each other and both are stale relative to the FY27 non-GAAP EPS guide the company raised yesterday ($16.67- 16.71). Using the company's own just-raised guide as the primary-source anchor:

Forward P/E (company guide, FY27 non-GAAP EPS $16.69 mid): $252.05 / $16.69 = 15.1x — genuinely cheap in absolute terms for a company with re-accelerating bookings, if the guide's growth is trusted at face value. It should not be trusted at full face value: the ~50%+ YoY jump implied (FY26 non-GAAP EPS was materially lower) is inflated by (a) the mechanical 10% share-count reduction from the debt-funded ASR, not organic earnings power, and (b) FQ2's unusually large net income ($4.84B) likely includes non-operating items (the same income statement shows recurring "Gain on Sale of Security" lines of $0.4-1.0B/quarter in adjacent periods) that were not fully decomposed here.
FCF yield. TTM FCF ~$15.16B against a market cap of ~$219B (spot × ~870M diluted shares, pre-full-ASR-settlement) = ~6.9% FCF yield — a reasonable, not extreme, starting point for a low-double-digit revenue grower with a real (if narrowing) moat. Note the FCF growth rate going forward is explicitly guided to just 4-5%, materially slower than the trailing 3yr 31.6% CAGR — the deceleration is real and self-inflicted by the leverage decision, not a demand problem.
Graham's Intrinsic Value — largely inapplicable, and the vendor input is also stale. √(22.5 × EPS × BVPS) using Yahoo's TTM EPS ($10.92) and BVPS ($63.66) = ~$125, far below spot. But Yahoo's BVPS has not caught up to the ASR-driven equity collapse — true BVPS post-buyback is closer to $34.24B equity / ~819M shares ≈ $41.80, which drops Graham IV to ~$101. Either way, Graham is the wrong tool here: it penalizes a company that converted equity into debt to retire shares, and it can't see the SaaS moat or the growth. Weight this near zero per the framework's own rule (Graham is for value plays with real book value; CRM's book value was deliberately levered away this year).
Bogle's Expected Return. Dividend yield 0.86% + guided non-GAAP EPS growth (~19% off the restated FY26 guide comparison, though the raw guide-to-guide jump looks larger) suggests a double-digit expected return if the market holds today's multiple — directionally supportive, but low-confidence given the EPS-growth input above is distorted.
DYT / DDM — N/A. Yield is small (0.86%) and the dividend is only 2 years old; not a dividend-driven thesis.

Fair value range: $230-320 (mid ~$270). The low end reflects a conservative 14-16x on the company's raised FY27 non-GAAP guide, discounted for the buyback-inflated EPS growth and the new leverage; the high end reflects 19-22x on the same guide if the bookings re-acceleration (cRPO +14% cc, strongest net-new AOV in 4 years) proves durable and the pricing-model transition resolves in Salesforce's favor. Spot ($252.05) sits in the lower half of that range — reasonable, not compellingly cheap, one day after a 22.6% pop.

4. Sentiment

FQ2 FY27 was unambiguously a blowout on every headline metric — revenue, EPS, cRPO, Agentforce ARR, guidance (raised twice this fiscal year already) — and it landed inside a software-sector rally that has already re-rated TEAM (+35% on its print), CRWD (+18.6%) and OKTA (+28.2%) in the same week. Benioff's tone was combative and specific: he called the "SaaSpocalypse" narrative "nonsense," pointing to record net-new AOV and near-record-low attrition as the rebuttal, and unveiled Claudeforce (a joint interface with Anthropic, GA September 2026) as the next AI storyline. Analyst response was broadly positive but not universally converted — BMO, UBS, Wells Fargo and Citigroup raised price targets through mid-August, several still hold Neutral/Equal-Weight ratings (targets $204-210, below spot), while Morgan Stanley downgraded to Equal-Weight in July (pre-print). Insider signal is mixed-to-positive: activist investor G. Mason Morfit (ValueAct) bought 96,000 shares (~$25M) in December 2025 at $260.58 — a real, sizable open-market purchase, not an options exercise. Several directors made routine ~$500K buys in the Feb-Mar 2026 window (standard director-ownership-program purchases, mildly positive). CEO Benioff's daily 2,250-share sales through Oct-2025 were systematic 10b5-1 option-exercise-and-sell activity (routine, not opportunistic) and stopped entirely after early Nov-2025 — no further Benioff sales recorded through late Aug-2026.

Key Risks

Leverage is new and material. Net debt went from $7.1B to $30B+ in one quarter to fund a buyback, not to fund growth. If bookings momentum stalls, Salesforce now carries real interest expense with less balance-sheet flexibility than a year ago; tangible book value is deeply negative.
Pricing-model transition risk on the core growth driver. Management's own admission — "still trapped in some ways in old per-user pricing models" for headless/agentic workloads — is the single biggest threat to Agentforce converting bookings growth into durable, capturable revenue, and it's the exact seam a distribution-advantaged rival (Microsoft) could exploit.
The FY27 EPS guide's growth rate is not clean. A meaningful share of the jump to $16.67-16.71 is mechanical (10% fewer shares from a debt-funded buyback) rather than organic earnings expansion — don't treat the headline growth rate as pure operating performance.
Data lag on the balance sheet. The FQ2 (Jul-2026) 10-Q was not yet reflected in vendor quarterly statements at the time of this analysis — confirm net debt, interest expense and goodwill trajectory directly at the next touch rather than assuming the FQ1 trend held flat.

Data Traps Hit

[[pitfall-vendor-forward-eps-is-the-wrong-fiscal-year]] — fired, with a twist. Yahoo's forwardPE (15.98x, implied EPS $15.78) and epsCurrentYear/priceEpsCurrentYear (17.18x, implied EPS $14.67) disagree with each other, and both are stale against the company's own FY27 non-GAAP EPS guide raised to $16.67-16.71 the day before this analysis. The standard fix in this pitfall (trust epsCurrentYear over forwardPE) does not fully resolve it here — neither vendor field had caught up to a guidance raise from less than 24 hours earlier. Primary-source guidance beat both vendor fields.
New candidate trap — Yahoo's bookValue/BVPS lags a mid-year buyback-driven equity collapse. Snapshot BVPS ($63.66) reflects the pre-ASR balance sheet; the true post-buyback book value (equity $34.24B / ~819M shares ≈ $41.80/share) is 34% lower. Any Graham IV or P/B computed off the raw snapshot field materially overstates book-value support. Recommend a new pitfall note — see below.
[[pitfall-single-quarter-fcf-read-as-ttm]] — guarded against, not fired. Reconstructed TTM FCF as the sum of the four most recent quarters ($2.18B + $5.32B + $6.56B + ~$1.10B ≈ $15.16B) rather than reading any single column as trailing-twelve-month.
Balance-sheet/income-statement data lag (not yet a named pitfall) — Yahoo's quarterly financial statements for CRM stopped at FQ1 FY27 (2026-04-30) as of this writing, one full quarter behind the FQ2 print this analysis is centered on. All FQ2-specific figures here come from the press release and secondary sources, not raw MCP statement data — flagged throughout rather than presented as reconciled.

Synthesis & Verdict

ACCUMULATE, conviction 7.0. Two forces are in real tension and the framework requires naming both rather than netting them into a false confidence:

The demand signal is genuinely strong and getting stronger — cRPO acceleration, the best net-new bookings quarter in four years, Agentforce tripling-plus YoY, and a guidance raise landing on top of an already-raised guide. This is Fundamentals + Sentiment pulling toward a higher score.
The balance sheet took on real, new risk this year, funded by debt rather than by the business's own cash generation, and management's own FCF-growth guide cut (9-10% → 4-5%) is the company itself telling you the leverage has a cost. This is Fundamentals pulling the other way, and it is the reason this isn't scored higher despite the blowout print.

Weighting: for a high-growth SaaS name, Moat + Sentiment normally outweigh a static Graham valuation — and they do here, since Graham is structurally broken on this name this year. But the leverage event is not a valuation-model quirk, it's a real capital-structure fact, and it caps conviction below where the earnings beat alone would put it. This is a close call because the same $25B decision that weakened the balance sheet also bought back 10% of the company near multi-year lows — a decision that will look shrewd if bookings momentum holds and look reckless if it doesn't. The next FQ3 print (~early December 2026) is the test: does cRPO growth hold ≥14% cc, and does the FY27 FCF guide stay at 4-5% or get cut again.

A portfolio-specific passage was removed from the public build.

Watchlist Write-Back (for orchestrator to apply)

Ticker: CRM · Conviction: [7.0] (from [—]) Entry: $210-235 (strong buy <$210) Trim: 22x fwd (~$367, off the company's own FY27 non-GAAP EPS guide $16.69 mid) — replaces the stale fixed-dollar Trim $250+ One-line thesis: FQ2 FY27 blowout (revenue +10.8%, cRPO +14% cc — best bookings quarter in 4 years, Agentforce ARR $1.5B +240%) confirms the CRM-data-gravity moat is still compounding, but the same year saw a $25B debt-funded buyback that erased tangible book value and forced management to cut its own FY27 FCF-growth guide in half — cheap on the raised EPS guide (15.1x fwd), not cheap enough to ignore the new leverage. One-line break: cRPO growth decelerates below 12% cc at FQ3, OR the FY27 FCF-growth guide is cut again below 4%, OR net debt is still rising (not stabilizing) at the next reported balance sheet.

Pitfall Notes for Orchestrator

New pitfall candidate — "Yahoo bookValue/BVPS lags a mid-year buyback-driven equity collapse." Claim: when a company executes a large accelerated share repurchase mid-fiscal-year (especially a debt-funded one), Yahoo's snapshot bookValue field can continue reporting the pre-buyback book value for at least one quarter after the equity actually shrank, because the field is not recomputed on the same clock as price/market cap. On CRM, snapshot BVPS was $63.66 (implicitly Jan-2026, pre-ASR) while the reconstructed Apr-2026 book value was ~$41.80/share — a 34% overstatement. This corrupts Graham IV and P/B for any company that ran a large buyback funded by anything other than steady-state FCF. Evidence: CRM quarterly balance sheet (Jan-2026 equity $59.14B/929M sh vs Apr-2026 equity $34.24B/819M sh), Yahoo get_stock_info snapshot bookValue field, 2026-08-27. Suggested scope: [accounting, valuation, data-quality, buybacks]. Related to [[pitfall-vendor-forward-eps-is-the-wrong-fiscal-year]] (same failure family — a vendor field lagging a real corporate event) and [[pitfall-weighted-average-share-count-lags-a-step-change]] (the share-count analog of the same lag).

Possible amendment to [[pitfall-vendor-forward-eps-is-the-wrong-fiscal-year]]: the standing guard ("if forwardPE and priceEpsCurrentYear disagree, trust the current-year field") did not fully resolve CRM, because both vendor fields were stale against a guidance raise issued less than 24 hours before the data pull. Worth adding a line: when a name reported earnings within the last 1-2 trading days, treat both vendor forward-EPS fields as potentially stale and check the company's own just-issued guidance before trusting either one.

Report file: Output/Stocks/Technology/CRM/analyze-2026-08-27.md