Salesforce (Minutes): The "Holy Trinity" of Applications, Data, and Intelligence

The following is the minutes of the Q4 2025 earnings call for $Salesforce(CRM.US) . For the earnings interpretation, please refer to Salesforce: Do intelligent agents need to spend money first to make money later? .

1. Core Financial Information Review

Revenue Situation: The total revenue for FY 2025 was $37.9 billion, a year-on-year increase of 9% (nominal and fixed exchange rates), with subscription and support revenue growing over 10% at fixed exchange rates. Q4 revenue was $10 billion, a nominal year-on-year growth of 8%, including about $75 million in foreign exchange headwinds, with a 9% growth at fixed exchange rates.

By region, revenue in the Americas grew 8% both nominally and at fixed exchange rates, EMEA grew 6% or 7% (fixed exchange rates), and APAC grew 10% or 14% (fixed exchange rates); by industry, Q4 saw strong performance in health and life sciences, communications and media, while technology, manufacturing, automotive, and energy industries performed relatively steadily. In Q4, the average of the top 100 deals involved 6 cloud services each. All of the top 10 deals included AI, Data Cloud, service platforms, and industry clouds.

Cash Flow: Q4 operating cash flow was nearly $4 billion, a year-on-year increase of 17%; free cash flow was $3.8 billion. The total operating cash flow for the year reached a record $13.1 billion, a year-on-year increase of 28% (including an estimated 10-point cash tax headwind); free cash flow was $12.4 billion, a year-on-year increase of 31%.

Remaining Performance Obligations (RPO): The company surpassed $60 billion for the first time in its history, reaching $63.4 billion in Q4, a year-on-year increase of 11%, representing customers' long-term commitment to Salesforce and the durability of its business model. CRPO was $30.2 billion, nominally increasing by 9% (including $300 million in foreign exchange headwinds, with an 11% growth at fixed exchange rates), benefiting from strong performance in Data Cloud, AI, and Slack, as well as early renewals.

Capital Return: In FY 2025, $7.8 billion was spent on stock buybacks and $1.5 billion on dividends, fully offsetting the stock-based compensation dilution for FY 2025 through the capital return program, which has returned over $21 billion to shareholders since its inception.

Revenue Guidance: For FY 2026, revenue is expected to be between $40.5 billion and $40.9 billion, a year-on-year growth of approximately 7% - 8% (nominal and fixed exchange rates), with subscription and support revenue expected to grow approximately 9% at fixed exchange rates. This considers about $200 million in foreign exchange headwinds from a stronger dollar, headwinds from professional services business growth, partial contributions from Data Cloud and Agentforce, as well as weakness in marketing and business operations and a slowdown in exploratory foundational growth The expected annual customer churn rate is slightly above 8% and remains stable.

Profit and cash flow guidance: The non-GAAP operating profit margin for fiscal year 2026 is expected to be 34%, an increase of 100 basis points year-on-year, with intentional investments in high-growth opportunities (such as Agentforce and Data Cloud), and a gradual increase in profit margins throughout the year is anticipated. The GAAP operating profit margin is expected to be 21.6%, an increase of over 250 basis points year-on-year. The GAAP diluted earnings per share are expected to be between $6.95 and $7.03, while the non-GAAP diluted earnings per share are expected to be between $11.09 and $11.17. Operating cash flow is expected to grow by approximately 10% - 11%, capital expenditures are expected to account for about 2% of revenue, and free cash flow is expected to grow by approximately 9% - 10%.

Q1 guidance: Revenue is expected to be between $9.71 billion and $9.76 billion, a year-on-year increase of 6% - 7% (nominal and fixed exchange rates), with CRPO nominal growth of approximately 10% year-on-year (including a $100 million foreign exchange headwind, slightly over 10% under fixed exchange rates). GAAP earnings per share are expected to be between $1.49 and $1.51, while non-GAAP earnings per share are expected to be between $2.53 and $2.55.

II. Detailed Content of the Earnings Call

2.1 Key Information from Executives

1. Product Performance: Data Cloud and Agentforce have performed exceptionally as two major new products, with Data Cloud serving as the driving fuel for Agentforce. Customers are increasing their investments, with data volume exceeding 50 trillion records, doubling year-on-year, and nearly a quarter of records being ingested from outside Salesforce through the zero-copy network. Agentforce had 3,000 paying customers within 90 days of launch, demonstrating strong productivity, efficiency improvements, and cost-saving capabilities across multiple industries. Sales Cloud and Service Cloud achieved double-digit growth in Q4.

In the fourth quarter, we completed over 400 transactions worth more than $1 million. Slack was included in over one-third of transactions exceeding $1 million. Tableau and MuleSoft are critical to customers, appearing in nearly half of the transactions over $1 million, with the integration of Tableau with Data Cloud and Agentforce set to change the way customers handle data, while MuleSoft is transforming enterprise integration through agent-driven integration.

2. Investments and Partnerships: With the continuous development of artificial intelligence, we will see increased investments in software to build intelligent agent layers and handle a higher proportion of workloads. We are leveraging the massive investments of these infrastructure companies. We are collaborating with two major infrastructure providers, Amazon and Alibaba. Both companies have made substantial investments in infrastructure. The third is Google, which is also making large-scale infrastructure investments. We are receiving very favorable pricing from these companies 3. Customer Cases: OpenTable processed 73% of restaurant web inquiries through Agentforce in 3 weeks, a 50% improvement over previous tools. Goodyear collaborated with Salesforce to automate and enhance sales efficiency using Agentforce, reducing on-site service repair time. Salesforce itself uses Agentforce in service, sales, business technology organization, and customer support, having autonomously handled 380,000 service requests since its launch on the Salesforce Health portal in October, with a resolution rate of 84%, and only 2% of requests requiring manual escalation. It also accelerated the quoting cycle by over 75%, increased AE capacity in Q4 while productivity improved by 7% year-on-year, interacting with over 50 potential customers daily in lead development. Pfizer has fully adopted Salesforce's Life Sciences Cloud, and Singapore Airlines hopes to provide services through Agentforce.

2.2 Q&A

Q: What are the current impacts and progress of the transition from a seat-based pricing model to a usage-based pricing model? Will it expand the overall contract size? Does the shift to a usage-based pricing model have a positive impact on Salesforce? Are there any obstacles to be aware of during the process?

A: The company initially adopted a user-based pricing model, while also having usage-based products, such as sandbox products, including Commerce Cloud, Marketing Cloud, etc. Agentforce is also a usage-based product. More specifically, it would be a hybrid model, depending on the specific circumstances of our customers, such as how many human employees they have and how many intelligent agents they have deployed. For example, Tableau was previously designed only for human users, but now it is deeply integrated into the data cloud and has also developed a deep intelligent agent layer.

This quarter, we reached a deal with a large telecommunications company, with an annual recurring revenue (ACV) of about $20 million and a total contract value (TCV) of about $60 million, of which the Agentforce component is about $7 million, and the value of products used by human employees is about $13 million. The pricing model will change over time. Customers understand the return on investment (ROI) that digital labor brings, which drives the speed of transactions. In the fourth quarter, we completed 3,000 transactions, and in the future, we may shift from the initial pricing method to credit pricing, which will have a significantly positive impact on the pricing structure.

Q: Agentforce is rapidly developing, but not all companies are ready to adopt AI agents. Has Agentforce created a halo effect on other products, meaning has Agentforce helped increase the usage or activity of other products?

A: Yes, we have seen this halo effect. From the way customers use technology, there are new ideas, new workflows, and new interactions. For example, they are now able to handle potential customer leads that could not be responded to in a timely manner after hours, and these leads subsequently flow into their Salesforce automation system This halo effect is reflected in our core technology, which enhances each of our core applications by providing intelligent support for them. The concept of the integration of applications, data, and intelligence that we convey is very attractive to customers.

Q: What are the considerations for merging the roles of Chief Operating Officer and Chief Financial Officer, and how will this evolve in the future?

A: We are pleased to have Robin serve as the new COO-CFO (Chief Operating and Financial Officer). Robin possesses the unique capabilities to excel in both positions. During the recruitment process, we found candidates from other companies who had held similar merged roles, and the combination of a CFO and COO forms a powerful partnership, making it a perfect match for us. Robin also believes that Agentforce currently presents exciting opportunities, and joining the operations team is very gratifying.

In this new era, our integrated and deeply unified platform can support customer growth and productivity, bringing long-term sustainable and profitable revenue growth to Salesforce. Additionally, the company's management team has been rebalanced, with Miguel Milano serving as Chief Revenue Officer, reporting directly to me and responsible for the global sales organization; Srini serving as Chief Engineering and Services Officer, responsible for customer service and support; Steve Fisher promoted to Chief Technology Officer; and David Schmaier becoming Chief Product and Influence Officer, reporting directly to me. We are very satisfied with the new management team and look forward to an excellent start in fiscal year 2026.

Q: From the perspective of labor arbitrage and the scale of job evolution opportunities, can a company with 1,000 employees slow down or stop hiring and introduce hundreds of Agentforce robots to accelerate growth? Can this concept be applied to a federal government agency with 3 million employees, introducing hundreds of thousands of Agentforce robots to improve efficiency after some employees voluntarily leave?

A: Taking our own company with 75,000 - 76,000 employees as an example, we have 9,000 customer service personnel and have seen significant efficiency improvements through help.salesforce.com, potentially reallocating some personnel to other functions such as sales and marketing. In engineering, efficiency has significantly improved after using new tools, especially high-performance coding tools, and we will not be hiring new engineers this year; engineering productivity has increased by 30% and will continue to improve. Sales are expected to grow significantly this year, and the sales team may expand by another 10% - 20%. Many companies are seeing opportunities to enhance their workforce with intelligent agents, and in the global economy, productivity can increase even without adding human labor. The company's goal is to become the world's leading digital workforce provider, which will be a focus for fiscal year 2026.

Q: How will the management model of future CEOs change?

A: I believe that this generation of CEOs is the last to manage only human employees; in the future, every CEO will manage both human employees and intelligent agents, which has become a trend. This was a topic of discussion among a group of large CEOs from various industries that I communicated with last week Q: How does the SaaS technology stack adapt to the world of agents during the transition from model building to inference layers, and to agent technology? Is there a risk that SaaS could become similar to a role database in the AI era? (i.e., the execution layer being replaced by general AI)

A: I believe there is a "holy trinity" of AI CRM, which consists of applications, data, and agents, and these three must work in synergy. In the past 90 days, we have had 380,000 conversations in service, with a resolution rate of up to 84%. You can check this at help.salesforce.com. Microsoft's Copilot has been launched for about two years; they are a distributor of OpenAI, repackaging ChatGPT. However, how well they have performed in agent delivery, whether they have best practices, whether they enable collaboration between humans and agents for customer success, and whether they balance human and agent workforce allocation are all considerations. The agent layer is certainly important, but it cannot operate in isolation; it needs to work with data and the support of connecting to all data sources through the company, as well as human collaboration. Our applications, data cloud, and agent layer are currently integrated and are creating value for our number one customer—Salesforce (itself). Whether other vendors have truly achieved this, we must be wary of false agents. Salesforce has achieved this; we are the number one AI CRM and a leader in the digital workforce revolution, currently having over 3,000 paying customers, with transaction volumes in the data cloud reaching trillions. Fiscal year 2026 will be the "year of agent power."

Risk disclosure and statement of this article: Dolphin Research Disclaimer and General Disclosure