CRM · Q3 2026 Earnings

Bullish

Salesforce

Reported December 3, 2025

30-second summary

30-second take: Q3 revenue grew 8.6% YoY to $10.26B with non-GAAP EPS of $3.25 — a $0.39+ beat versus the $2.84–2.86 guide — and Agentforce + Data 360 ARR crossed $1.4B with paid Agentforce deals above 9,500 (up from 6,000+ last quarter). Management raised FY26 revenue to $41.45–41.55B (+$0.30B midpoint), non-GAAP EPS to $11.75–11.77 (+$0.42), GAAP EPS to $7.22–7.24, and operating cash flow / free cash flow growth to ~13–14% (from 12–13%). The next-quarter guide of +11–12% nominal revenue growth (10–11% CC) is the cleanest acceleration signal since the AI cycle began — management is finally telegraphing reacceleration in print, not just commentary.

Headline numbers

EPS

Q3 FY2026

$3.25

Revenue

Q3 FY2026

$10.26B

+8.6% YoY

Gross margin

Q3 FY2026

78.0%

Free cash flow

Q3 FY2026

$2.18B

Operating margin

Q3 FY2026

21.3%

Key financials

Q3 FY2026

Metric	Q3 FY2026	YoY	Q2 FY2026	QoQ
Revenue	$10.26B	+8.6%	$10.24B	+0.2%
EPS	$3.25	—	$2.91	+11.7%
Gross margin	78.0%	—	78.0%	+0bps
Operating margin	21.3%	—	22.8%	-150bps
Free cash flow	$2.18B	—	$0.60B	+259.8%

Guidance

Guidance is issued for both next quarter and the full year. Both may appear below.

Actuals vs prior guidance

Metric	Period	Prior guide	Actual	Δ	Result
Revenue	Q3 FY2026	$10.24 - $10.29 billion	$10.259 billion	in-line (midpoint $10.265B guided, $10.259B actual)	Beat
Revenue (YoY growth)	Q3 FY2026	8% - 9% YoY	8.6% YoY	+0.6 to +0.9pts above high end of guide range	Beat
EPS (GAAP)	Q3 FY2026	$1.60 - $1.62	$2.19	+$0.57 to +$0.59 above high end of guide	Beat
EPS (non-GAAP)	Q3 FY2026	$2.84 - $2.86	$3.25	+$0.39 to +$0.41 above high end of guide	Beat
Operating Margin (GAAP)	Q3 FY2026	—	21.3%	—	Beat
Operating Margin (non-GAAP)	Q3 FY2026	—	35.5%	—	Beat

Changes to prior guidance

Metric	Period	Prior guide	New guide	Δ	Result
Revenue	FY 2026	$41.10 - $41.30 billion	$41.45 - $41.55 billion	+$0.15 to +$0.45 billion raise at both ends	Raised
Revenue (YoY growth)	FY 2026	8.5% - 9% YoY (8% constant currency)	9% - 10% YoY (9% constant currency)	+0.5pts raise at both nominal and CC low end	Raised
EPS (GAAP)	FY 2026	$6.99 - $7.03	$7.22 - $7.24	+$0.23 to +$0.25 raise at both ends	Raised
EPS (non-GAAP)	FY 2026	$11.33 - $11.37	$11.75 - $11.77	+$0.42 to +$0.44 raise at both ends	Raised
Operating Margin (GAAP)	FY 2026	21.2%	20	—	Raised

Segment performance

Q3 FY2026

Segment	Q3 FY2026	YoY
Agentforce Sales	$2.297B	+8.4%
Agentforce Service	$2.495B	+9.0%
Agentforce 360 Platform, Slack and Other	$2.18B	+19.4%
Agentforce Marketing and Commerce	$1.361B	+2.0%
Agentforce Integration and Analytics	$1.393B	+6.1%
Subscription & Support Revenue	$9.726 billion	—

Platform metrics

Q3 FY2026

Segment	Q3 FY2026
Agentforce and Data 360 ARR	$1.4 billion
Agentforce ARR	>$0.5 billion
Agentforce Paid Deals	>9,500
Agentforce Tokens Processed	3.2 trillion
Current Remaining Performance Obligation (cRPO)	$29.4 billion
Total Remaining Performance Obligation (RPO)	$59.5 billion

Profitability

Q3 FY2026

Segment	Q3 FY2026
Non-GAAP Operating Margin	35.5%

Other KPIs

Q3 FY2026

Segment	Q3 FY2026	YoY
Americas	$6.703B	+7.8%
Europe	$2.47B	+10.9%
Asia Pacific	$1.086B	+9.0%

Management tone

Narrative arc: Pilot-to-platform (Q1) → Agentic enterprise has arrived (Q2) → Consumption flywheel earning the print (Q3).

The throughline across three quarters has been management progressively shifting from forward claims to behind-the-print evidence. Q1 was about deal counts (4,000 paying), Q2 was about declared arrival ("the agentic enterprise has arrived"), and Q3 is built around production metrics — 70% QoQ growth in customers in production, 3.2 trillion tokens, 9,500+ paid deals. The anchor: "This isn't your clippy. This is not your kind of a good AI demo. This is real enterprise adoption of software agentic AI and capability at scale globally." The signal is that management now believes the proof points are dense enough to address bear-thesis pushback directly rather than via rhetoric.

The data layer was reframed from "strategic business" to "$10 billion business." Q2 framed Data Cloud as "probably the most strategic and most important business for Salesforce going forward... already a $7 billion business." This quarter Benioff went further: "when you look at Data360 plus MuleSoft plus Informatica... the data layer... I think it's about a $10 billion business for us next year now... that's the first layer. That's data." $7B → $10B in one quarter reflects Informatica closing (three months ahead of schedule) and a deliberate elevation of the data layer above the application layer in the strategy narrative. This frames the Informatica deal not as M&A but as the foundational layer of the AI moat.

Employee agents emerged as a parallel monetization vector to customer agents. Through Q1 and Q2 the AgentForce story was almost entirely customer-facing (service, sales). This quarter Mark explicitly redrew the strategy: "employee agents... deeply integrated into our Slack product... it's the core of every demonstration we give to our customers." Slack Bot is now described as "the heart of our employee agent strategy." This matters because customer agents had a clear headcount-displacement bear case; employee agents reframe Slack from a strategically underwhelming acquisition into a critical monetization surface — and provide a second AOV expansion path independent of seat reduction.

Pricing flexibility evolved from per-conversation to a four-model architecture. Last quarter pricing was a vulnerability — investors flagged confusion. This quarter Miguel laid out a complete pricing system: Agentic Enterprise License Agreement (AELA, 16 closed in Q3, 100+ in pipeline, multimillion-dollar deals), seed-based SKUs (doubled YoY), per-conversation usage-based, and Flex agreements that let customers redeploy headcount-savings into agent credits. "362 customers 'refilled the tank' versus 3 in Q1." The Flex agreement is the key innovation — it directly answers the "but what if seats decline?" bear case by converting seat reduction into prepaid agent consumption.

Sales capacity now positioned as the binding constraint, not demand. Q1 cited the +22% AE capacity build as ambitious; this quarter Miguel framed it as already deployed (20% more capacity in place, 15% ramped by year-end) with double-digit pipeline growth matching capacity growth. The implied message: bookings could be higher if reps could be ramped faster. Combined with the Q4 cRPO guide of +15% nominal, the operational picture is one where management has prebuilt distribution ahead of an inflection it now expects to print.

Recurring themes management leaned on this quarter:

AgentForce as fastest-growing product ever, driving deal mix and consumptionEmployee agents (Slack Bot) emerging as parallel adoption vector alongside customer agentsData harmonization/integration (Data360 + MuleSoft + Informatica) as competitive moat against hallucinationsEnterprise adoption at scale: 3.2 trillion tokens processed, 1.2B LLM calls, 540B tokens in October aloneMulti-cloud strategy and Growth360 Playbook driving 70%+ of top 100 wins with 5+ cloudsPublic sector and life sciences as new high-momentum verticals (IRS, UK police, pharma takeover)

Risks management surfaced:

Faster-than-anticipated mixed shift to cloud for Tableau creating on-prem revenue timing unpredictabilityWeakness in marketing and commerce segments offsetting agent force gainsAsia Pacific constrained, particularly in Australia and IndiaComms/media, manufacturing, automotive, energy vertical performance more measuredForeign exchange headwinds ($25M FX impact since last quarter, though partially offset by $200M CRPO tailwind)

Q&A highlights

Keith Weiss · Morgan Stanley

Addresses the marketplace narrative that generative AI will enable enterprises to build their own solutions instead of using vendors like Salesforce. Asked what Miguel is hearing from customers about their appetite to build DIY solutions versus adopting Salesforce's generative AI capabilities.

Miguel emphasized a major secular trend toward the 'agentic enterprise' where companies want conversational, autonomous-capable systems. He argued customers recognize that LLMs alone are insufficient—they need context (data, metadata, deterministic workflows) embedded where humans work. Only Salesforce can provide this integrated combination of data, apps, and AI, leading to strong demand and bookings growth. Stressed that 'the last mile is hard' and requires the full platform.

Miguel visited 400 customers across 3 continents and 12 countries in the past quarterCustomers progressed from experimentation to frustration to wanting to scaleMulti-trillion dollar market opportunity identifiedLLMs cannot succeed alone; context and deterministic workflows are critical

Remo Lensha · Barclays

Asked about sales capacity expansion and how Salesforce thinks about ramp productivity for new representatives.

Miguel detailed a proactive capacity investment initiated a year ago by Mark. Salesforce has 20% more capacity in place today and will finish the year with 15% more ramped capacity. Takes 6-12 months to ramp AEs on average. Pipeline generation in Q3 was strong with double-digit growth, and open pipeline for next year shows double-digit growth matching enabled capacity growth. Highlighted the exponential monetization multiplier from agentic enterprise (3x-10x), with examples of customers seeing bookings/AOV double, triple, or multiply 4-5x.

20% more sales capacity in place today; 15% enabled/ramped by year-end6-12 months average ramp time for AEsQ3 pipe gen growth: double-digit, above expectationsNext year open pipeline: double-digit healthy growth

Brad Zelnick · Deutsche Bank

Asked about Salesforce's competitive advantage in infrastructure (now over $10B with Informatica acquisition), how Salesforce avoids building data centers, and how this infrastructure business is leveraged to drive overall company success.

Mark clarified Salesforce is not building data centers but will leverage third-party builds and benefit from lower costs. Described the 'data foundation' (Informatica, Data360, Data Cloud, MuleSoft) as ~$10B business and fundamental for customer transformation. Key differentiator is federation capability—e.g., federating Data360 to IBM mainframe, allowing AgentForce to run on both platform data and legacy system data simultaneously. Infrastructure is deeply integrated with all apps (AgentForce, customer agents, employee agents), not independent.

Data foundation composed of Informatica, Data360, Data Cloud, MuleSoftExpected ~$10B business from data infrastructure next yearFederation allows Data360 to federate to IBM mainframe (demonstrated in Tokyo)Infrastructure is critical for accurate, reliable, low-hallucination AI

Brent Thill · Jefferies

Asked about the halo effect of AgentForce on sales and service (stable at high single-digit growth) and Slack's acceleration. Requested elaboration on AgentForce's impact on other clouds and Slack's resurgence.

Mark affirmed AgentForce as an accelerator on core apps, refuting the false narrative that large language models jeopardize core business. Emphasized humans + agents + apps + data as the driver. Srini provided deep product context: Data360 ingest up 38% QoQ, zero-copy up 52% in records, unstructured data up 109%. The foundation requires context (data, metadata, deterministic reasoning), tools (actions, APIs via MuleSoft), and day-two operations (eval, audit, compliance). Explained the infrastructure rebuild (hyperforce layer, data layer, metadata layer, application rewrites). Contrasted DIY attempts with platform approach; customers who tried DIY are now moving to Salesforce platform.

Data360 ingest: +38% QoQ growthData360 zero-copy: +52% QoQ growth in recordsUnstructured data growth: +109% QoQMulti-year infrastructure rebuild completed (hyperforce, data, metadata, apps)

Kirk Matern · Evercore Partners

Asked about AgentForce pricing clarity (confusion at launch) and investor concerns about how Salesforce monetizes when AgentForce helps customers reduce headcount. Requested explanation of how AOV grows even in stable or declining headcount scenarios.

Miguel highlighted Q3 booking momentum: 70% more AgentForce customers in production, 362 customers 'refilled the tank' (vs. 3 in Q1). Described evolution from per-conversation/usage-based pricing to flexible models. Now offers: (1) Agentic Enterprise License Agreement (AELA)—flat-fee, all-in approach for determined customers (16 closed in Q3, 100+ in pipeline, multimillion-dollar deals); (2) Seed-based SKUs (Agent Force for Sales/Service)—predictability-focused, doubled YoY; (3) Per-conversation/action usage-based; (4) Flex agreements allowing customers to redeploy headcount savings into credits for AELA or seed licenses. Emphasized future is 'humans and agents working together,' noting both headcount and price per seat increasing for core clouds. Flex agreements provide full flexibility if headcount decreases in some areas.

Q3 AgentForce bookings: 70

Answers to last quarter's watch list

Whether GAAP margin cuts continue or stabilize. Cuts continued — FY26 GAAP operating margin was lowered another 90bps to 20.3% (from 21.2%), the second consecutive quarterly cut. The driver is Informatica closing and associated integration costs. Non-GAAP margin guide held at 34.1%, so the issue is purely deal-related, but two consecutive cuts mean GAAP cleanup is taking longer than initially framed.

Resolved negatively

Marketing & Commerce growth crossing 5%. Got worse, not better — Marketing & Commerce decelerated to +2.0% YoY from +4.0% in Q2 and +3.3% in Q1. Four consecutive quarters under 5%, with the trend pointing down. This is now clearly structural rather than cyclical, and management still hasn't addressed strategic positioning.

Resolved negatively

Data Cloud + AI ARR cadence above $1.5B at >100% growth. The disclosure framework changed: management now reports "Agentforce + Data 360 ARR" at $1.4B and breaks out Agentforce ARR separately at >$500M (+330% YoY). The metric is no longer directly comparable to last quarter's $1.2B Data Cloud + AI ARR. The split disclosure is arguably more useful but breaks the trend line management drew.

Not resolved

Agentforce paid deal count above 8,000. Cleared — paid Agentforce deals at >9,500, up from 6,000+ in Q2 (~+50% QoQ, matching the prior quarter's pace). Customers in production jumped 70% QoQ. The production-readiness bottleneck is decisively cleared.

Resolved positively

cRPO retiring the FY23-cohort drag commentary. The Q4 cRPO guide of ~+15% nominal (+13% CC) is materially above the +10–11% pace of recent quarters, and management on this call emphasized Agentforce momentum and "path to $60B+ organic revenue target" rather than dwelling on the FY23 cohort. The drag commentary is being replaced by reacceleration commentary, though management hasn't explicitly retired it.

Resolved positively

Informatica close timing slipping past early FY27. Informatica closed three months ahead of schedule — Mark explicitly called it out: "completed three months ahead of schedule." This is the opposite of slippage. The Q3 segment rename to "Agentforce 360 Platform, Slack and Other" and the $10B data-layer reframing both reflect the closed deal.

Resolved positively

What to watch into next quarter

Whether Q4 revenue prints inside the $11.13–11.23B guide and cRPO clears +15% nominal. This is the inflection guide. A Q4 print at the midpoint would be +11.5% YoY versus 8.6% just delivered — the clearest revenue acceleration since FY23. Watch both the revenue print and whether cRPO actually lands at ~15% nominal; cRPO is the harder metric to manage to and the cleaner signal.

Marketing & Commerce stabilization or strategic review. Four quarters at +2–4% with the trajectory worsening. A fifth quarter under 5% should force the question of divestiture or rationalization; management has been silent on strategic options.

GAAP operating margin cuts stopping. Two consecutive quarterly cuts to FY26 GAAP margin (now 20.3% from 21.6% at Q1). Watch for stabilization in Q4 — a third consecutive cut would suggest Informatica integration costs are larger or more persistent than disclosed.

Agentforce ARR trajectory from the new $500M+ baseline. First-time standalone disclosure at +330% YoY. Watch whether next quarter discloses the same metric on the same basis (it likely won't be +330% indefinitely) and whether $1B+ in Agentforce ARR specifically is reached in FY27 — that's the threshold where Agentforce stops being a feature story and becomes a P&L line.

AELA pipeline conversion. 16 closed Q3, 100+ in pipeline. Watch whether the 100+ converts at >50% rate over the next two quarters — AELA is the highest-AOV pricing model and its conversion rate is the cleanest indicator that "agentic transformation" is a real budget category for CIOs versus a sales narrative.

Q4 token volume and any disclosure of LLM gateway economics. 3.2 trillion tokens cumulative through Q3, 540B in October alone. If Salesforce becomes one of the largest LLM consumers globally, the unit economics of gateway arbitrage (routing across foundation models for cost) become material. Watch for any quantification of LLM cost per token or gross margin attribution.

Sources

Salesforce Q3 FY26 Press Release (Form 8-K Exhibit 99.1), filed December 3, 2025 — https://www.sec.gov/Archives/edgar/data/1108524/000110852425000234/crm-q3fy26xexhibit991.htm
Salesforce Q3 FY26 earnings call commentary (management prepared remarks and Q&A)
Tapebrief Q1 FY26 and Q2 FY26 CRM briefs — prior watch lists and guidance baselines

Get the next brief, free.

We publish analyst-grade earnings briefs the same day or morning after every call — headline numbers, segment KPIs, Q&A highlights, and tone analysis. Free during beta.

This is not investment advice.