Salesforce's $3.9B Agentforce and Data 360 ARR: How Much Was Bought, Not Built

Salesforce says its Agentforce and Data 360 annual recurring revenue reached nearly $3.9 billion in the quarter to 31 July 2026, up more than 210%. About $1.1 billion of that number arrived on 18 November 2025, when the Informatica acquisition closed. Strip it out and organic growth in the Salesforce Agentforce ARR line looks closer to 130%. Still fast. Not the same claim.

What Salesforce actually reported for Agentforce and Data 360 ARR

Salesforce published its second quarter fiscal 2027 results on 26 August 2026. The headline AI metric read that Agentforce and Data 360 annual recurring revenue reached nearly $3.9 billion, up over 210% year on year. Agentforce ARR on its own exceeded $1.5 billion, up over 240%.

Annual recurring revenue, or ARR, is the annualised value of subscription contracts active at the end of a period. It is a company-defined metric, not an audited accounting line. Each vendor chooses what goes into it, and can change that choice.

Read the number in sequence and the jump becomes visible. One year earlier, for the quarter to 31 July 2025, Salesforce reported Data Cloud and AI ARR of over $1.2 billion. In the quarter to 31 October 2025 it was nearly $1.4 billion. Then, one quarter later, it was more than $2.9 billion.

All figures are Salesforce's own wording from its quarterly results releases. The metric was called Data Cloud and AI ARR until fiscal 2026 and Agentforce and Data 360 ARR afterwards. "Over" and "nearly" are the company's rounding, not ours.

The Agentforce definition moved in the same quarter

The Q2 fiscal 2027 release also changed what counts. It states that beginning in Q2 fiscal 2027, Agentforce ARR includes Slackbot and Headless 360. No restated prior-year figure was published alongside it.

So the 240% Agentforce growth rate compares a wider current basket against a narrower old one. We do not think that is improper. We do think a growth rate that crosses a definition change should carry the change in the same sentence, and this one did not.

Where Informatica sits inside the Salesforce ARR number

Salesforce agreed to buy Informatica on 27 May 2025 for $25 a share in cash, about $8 billion of equity value. The deal closed on 18 November 2025. On that day Salesforce drew $4 billion on a 364-day credit facility and $2 billion on a three-year facility to fund it.

The first two releases after the close were explicit. Q4 fiscal 2026 reported ARR that exceeds $2.9 billion, including $1.1 billion Informatica Cloud ARR. Q1 fiscal 2027 reported nearly $3.4 billion, again including $1.1 billion of Informatica Cloud ARR. Salesforce did not hide the acquisition. It put it in the same bullet.

The Q2 release dropped the split

The Q2 fiscal 2027 release does not repeat the Informatica clause. It defines Agentforce and Data 360 ARR as Data 360 plus certain generative AI subscriptions, including Agentforce, generative AI products and features, and Headless 360. It defines Informatica Cloud ARR separately. It never says whether one sits inside the other, and gives no Informatica Cloud ARR figure for the quarter.

We read the $3.9 billion as still including Informatica, for an arithmetic reason. If Informatica had been removed, the remaining business would have gone from about $2.3 billion to $3.9 billion in a single quarter. Nothing else in the release, including revenue, supports a $1.6 billion organic quarter. The series only holds together if Informatica is still in it.

Rebuilding organic Agentforce and Data 360 growth

There are two honest ways to remove an acquisition from a growth rate. They answer different questions, so publish both. A single number hides which question you chose.

Method 1: subtract the acquired ARR

Take the current figure, remove the acquired ARR, and compare against the original base. Nearly $3.9 billion less about $1.1 billion leaves roughly $2.8 billion. Against just over $1.2 billion a year earlier, that is growth of about 133%.

This method answers one question: how much did the business Salesforce already owned grow? It is the closest thing to a sales scorecard. It still overstates slightly, because the base is "over" $1.2 billion and the current figure is "nearly" $3.9 billion, and both roundings push growth up.

Method 2: put Informatica in both years

The pro forma method adds the acquired company's ARR to the prior-year base. Informatica's last public report, for the quarter to 30 September 2025, showed Cloud Subscription ARR of $968.6 million, up 29.5%. That is two months after Salesforce's base date, so it is a slightly generous stand-in.

Add it to the $1.2 billion and the pro forma base is about $2.17 billion. Nearly $3.9 billion against that is growth of roughly 80%. Use a smaller July 2025 Informatica figure and the result edges toward 85%. This method answers a different question: how fast did the combined business grow?

Why the reported rate is the least useful of the three

The reported 210% compares a current figure that includes Informatica with a prior-year figure that does not. It measures neither the sales motion nor the combined business. It measures the sales motion plus a closing date.

We would not call the headline misleading in a legal sense. The Q4 and Q1 releases named the Informatica amount clearly. Our objection is narrower: a growth rate that mixes bought and built ARR should never be the first number a reader sees, and in this release it was the only one.

Scenarios, not disclosures. If Informatica's cloud ARR kept its pre-deal pace of about 30% a year, it would sit near $1.2 billion by July 2026, which puts organic growth near 125%. Salesforce could settle this with one number.

What Informatica's own filings say about its cloud ARR

The Informatica slice deserves a second look, because it is labelled "cloud" and sits inside an AI metric. Informatica's last results tell you how much of that cloud ARR was new demand.

In the year to 30 September 2025, Informatica's total ARR rose only 3.9%, to $1.75 billion. Cloud ARR rose $221 million. Self-managed subscription ARR fell $97 million and maintenance ARR fell $58 million over the same period.

Cloud growth was mostly conversion

Roughly $155 million of the $221 million cloud gain was matched by declines in the on-premises lines. Not every lost on-premises dollar became a cloud dollar, and some cloud ARR was genuinely new. But most of the cloud line's growth was existing customers changing how they pay.

Informatica said as much. In its Q4 2024 results it reported that net new bookings missed forecast, due primarily to a higher-than-expected contribution from on-premises and self-managed migrations to the cloud. Its cloud net revenue retention was 120% in Q3 2025, down from 126% a year earlier.

This is the part we think most commentary misses. The question is not only whether Informatica belongs in an AI and data ARR figure. It is whether a line that grew by moving existing customers between billing models tells you anything about agent demand. We do not think it does.

The revenue lines tell the same organic growth story

ARR is a company-defined metric. Revenue is audited and reported in a consistent table, so it is the better check. Salesforce gave the Informatica revenue contribution every quarter, which makes the subtraction easy.

In Q2 fiscal 2027, subscription and support revenue was $10,820 million against $9,690 million a year earlier, including $440 million from Informatica. Total growth was $1,130 million, or about 12%. Without Informatica it was $690 million, or about 7%.

The category that holds Data 360

Salesforce now reports a line called Data 360, Headless Platform, and Other. Its footnote lists Data 360, Flex Credits, Headless Platform, Informatica, MuleSoft and Tableau inside it. The line grew from $3,008 million to $3,618 million, a gain of $610 million.

Remove the $440 million Informatica contribution and the gain is about $170 million, or roughly 6%. This is the revenue home of Data 360, the product carrying the AI story. MuleSoft and Tableau share the line, so the 6% does not measure Data 360 alone. It does cap how fast the whole basket grew without the acquisition.

Guidance carries the same split. Salesforce raised full-year fiscal 2027 revenue guidance to $46.1 billion to $46.4 billion, up 11% to 12%, including slightly above 3 points of Informatica contribution. The company explained the raise as $100 million of organic growth, $200 million from two pending acquisitions, and a $100 million currency headwind.

That disclosure is a model for how this should be done. Two thirds of the gross raise is bought, and Salesforce said so in the sentence that announced it. We wish the ARR line followed the same discipline.

Where this estimate is weakest

This analysis rests on one undisclosed number and one inference. Both could be wrong, and the case for Salesforce's framing is stronger than critics usually allow.

The missing Q2 disclosure cuts both ways

Salesforce did not publish Informatica Cloud ARR for the July 2026 quarter. We assumed it stayed near $1.1 billion. If it grew at its old pace, organic growth is lower, near 125%. If Salesforce moved some Informatica customers into Data 360 contracts, part of that ARR now reads as Data 360 and our subtraction undercounts it.

Our inference that the $3.9 billion still includes Informatica is also an inference. It rests on the series being otherwise implausible, not on a sentence in the release. If Salesforce confirms a different treatment, the organic figure changes and so does this post.

The case for counting Informatica

Salesforce's argument is that agents are only as good as the data they act on, and Informatica is that data layer. On that view, Informatica is part of the AI product, not a bolt-on. Buyers who move to Agentforce may well buy Informatica Cloud as a prerequisite.

That argument has merit. Cross-sell after an acquisition is real sales work, and some of the future growth in the Informatica line will be organic in every sense. Our point is limited to timing. ARR that arrives on a closing date is not evidence of demand for agents, whatever happens to it afterwards.

Finally, the base itself moved. The old metric covered Data Cloud plus Agentforce and standalone generative AI products. The new one adds generative AI products and features and Headless 360. Some of the 133% is therefore definitional, and nothing published lets anyone measure how much.

A five-question test for any vendor's AI ARR headline

Salesforce is the clearest current example, not the only one. Every large software vendor now publishes an AI revenue figure, and most of them are built from the same parts. We covered how loosely these numbers are reported across the market in the leaderboard of AI revenue figures and what sits behind them.

The test below works on any of them. It takes about ten minutes with the earnings release open.

Salesforce passes the first question in two quarters and fails it in the third. It fails the second. It passes the fourth only if you do the subtraction yourself. Its guidance passes the fifth, and its ARR line does not.

The same test applies to pricing claims. A vendor that blurs acquired and organic ARR will often blur list price and realised price too, which is the subject of the arithmetic behind Agentforce's per-conversation pricing.

You are not buying Salesforce stock. You may be deciding whether Agentforce is the safe choice because "everyone is adopting it". A 210% headline is a social proof argument. A 7% organic subscription growth rate is a different picture of how fast real customers are adding spend.

Whether Salesforce ends up a platform or a casualty of the agent shift is argued in the case for and against Salesforce in the agent era. This post does not settle that. It only removes one number from the evidence on the bullish side, and replaces it with a smaller one that is still impressive.

The wider point holds across the sector. The software selloff earlier this year was partly a market discovering that AI revenue lines were growing far faster than total revenue, as covered in the breakdown of the SaaS selloff. When a vendor's AI line and its revenue line disagree this sharply, believe revenue first.

Frequently asked questions

Where to start with the next earnings release

Pick one vendor whose AI number appears in your next renewal conversation. Open its latest earnings release and search for three words: "including", "acquisition" and "definition". Each hit is a place where the headline growth rate may be carrying something other than sales.

Then do one subtraction. Take the AI figure, remove anything the vendor says it acquired, and divide by the figure from a year earlier. Write that organic rate next to the headline in your vendor file. If the vendor will not give you the inputs, ask its account team for them in writing, and note who declined.

Related reading