Salesforce in the AI stack

The business and what the filings show

Salesforce is the largest application software company on this map: $43.9B of revenue over its latest four quarters, +11%, at a 77% gross margin and a 20% operating margin. It generated $15.2B of free cash flow and reports remaining performance obligations of $66.3B. No single customer accounted for 10% or more of revenue in fiscal 2026.

Its AI strategy is Agentforce, autonomous AI agents built into the sales, service and marketing applications its customers already use. The company describes them as digital labor, which makes Salesforce the clearest example of a software company trying to change what it charges for: from seats used by people to work done by agents.

Salesforce is a mature company now growing more slowly than most of the software layer on this map, so AI is being asked to reaccelerate it rather than to create it. That makes its revenue growth over the next few years an unusually clean test of whether AI agents produce new revenue for incumbents or only defend what they already have.

How AI shows up in the numbers

AI shows up at Salesforce as Agentforce and the data products it runs on, sold to existing customers. Its numbers do not yet separate AI revenue from the rest, so the test is in the totals: growth that reaccelerates while gross margin holds would mean customers are paying for agents; flat growth with rising AI costs would mean they are not. Its earnings releases show how often it talks about AI and how often it puts a number on it.

For scale, Salesforce’s revenue over the latest four quarters equals 8% of the $566B the five largest buyers spent on capital projects over their latest four quarters. Their capex rose +81% on the year before, while Salesforce’s revenue grew +11%: slower than the spending, so it is capturing a smaller share of each new capex dollar, or more of its business sits outside the build-out.

How it talks about AI

In its latest earnings release, furnished to the SEC on August 26, 2026, Salesforce mentioned AI 6 times in 1,483 words of narrative text, about 4.0 mentions per 1,000 words. Of its 6 sentences about AI, 1 included a number: a revenue figure, a growth rate, a customer count or another measurable result. Across the 8 releases we read, back to December 3, 2024, AI mentions per 1,000 words went from 2.9 to 4.0.

Specific claims are easier to check against the numbers than general enthusiasm, which is why the share of AI sentences that carry a figure is worth watching alongside how often AI is mentioned. The sentences below are quoted from the release as published, after removing safe-harbor text and company boilerplate.

• “We’re seeing incredible demand for our AI and data products, with ARR about to cross $4 billion.”
• “AI is delivering value across every layer of our platform.”

Source: earnings release (Form 8-K, Exhibit 99.1), August 26, 2026.

The trend in the numbers

Since the start of 2023, Salesforce’s largest quarter by revenue was Q2 2026 at $11.3B, and its highest gross margin was 78% in Q2 2025. The latest quarter set the revenue record for the period.

Quarter by quarter, Salesforce’s revenue grew +9% in Q3 2025, +12% in Q4 2025, +13% in Q1 2026, +11% in Q2 2026 on the same quarter a year earlier. The pace has been steady.

Gross margin in the latest quarter was 77%, against 78% a year earlier, little changed. Operating margin moved from 23% to 21% over the same span. Over the latest four quarters it turned 36% of its revenue into operating cash flow and spent 1% of revenue on capital projects, against 2% a year earlier.

Where it sits in the stack

Salesforce sits in the software and applications layer of the AI stack. It sits at the top of the stack, where revenue from end customers has to come from; its own spending flows to AI clouds.

Among the 27 companies in this layer on our map, it ranks first by revenue and number 22 of 27 by revenue growth. Its gross margin of 77% compares with 78% for the layer as a whole, close to the layer average. The full layer analysis is on the software page.

Latest four quarters for 10-Q filers; latest fiscal year for 20-F filers. Companies under $100 million of revenue are left out.

Against the 27 US-listed companies in its layer, Salesforce grew +11% over the latest year versus +19% for the layer combined, and converted 34% of revenue into free cash flow versus 33% for the layer. It is growing more slowly than its layer, so its share of the layer’s revenue is shrinking.

What has to go right, and what could go wrong

What has to go right. Customers have to pay for Agentforce in volumes large enough to lift Salesforce’s growth, without the inference costs eroding its gross margin.

What could go wrong. If AI agents reduce the number of people who need Salesforce licenses faster than agent revenue grows, or if customers build agents on other platforms, AI could become a cost of defending the business rather than a source of growth.

These describe the conditions the business depends on, as the filings show them. They are not a view on the stock.

Through the lens of Investing in AI

Rob May’s book offers six mental models for telling AI moats from traps. The one most useful for Salesforce’s layer is “AI challenges traditional SaaS assumptions”: Per-seat pricing, zero marginal cost, smooth ARR curves, and automatic land-and-expand break down when software does work on your behalf. The mistake is not using SaaS instincts. It is using them automatically.

The question it asks of any company is: What work is this product actually doing, and how does pricing map to it? Applied to Salesforce, the filings on this page are the place to start: the margin trend shows whether it holds scarce value or passes it through, the customer and financing disclosures show who it depends on, and the commitments show what it has promised before the revenue arrives.

Customers and concentration

US companies must disclose any customer that accounts for 10% or more of revenue, and foreign filers usually describe their largest customers. Concentration matters because a company’s growth is only as durable as the handful of buyers behind it, and in the AI stack those buyers are often the same few companies.

• “No single customer accounted for ten percent or more of total revenue during fiscal 2026 and 2025.” 10-K, March 2, 2026

Commitments and backlog

• Backlog. Salesforce reports $66.3B of remaining performance obligations, 1.5 times its revenue over the latest four quarters: contracted revenue not yet earned.
• Purchase commitments. $4.2B of unconditional purchase obligations as of July 31, 2026 (10-Q): what it has contracted to buy from its own suppliers.

What to watch

Salesforce’s next quarterly report will cover the quarter ending around October 30, 2026. Large companies must file a 10-Q within 40 days of a quarter’s end and a 10-K within 60 days of the fiscal year’s end, so the new figures should arrive by about December 29, 2026 at the latest. This page, the layer pages and the next quarterly report all update from that filing.

• Revenue growth: reacceleration is what Agentforce has to deliver.
• Gross margin, where the cost of running AI agents would appear.
• Remaining performance obligations, especially the part due in the next twelve months.
• Its place in the next quarterly AI Stack Report, which updates every company on the map from the new filings.

Figures from Salesforce’s SEC filings, latest period ended July 31, 2026. Other companies: all company pages.