Is SaaS dead

SaaS is not dead. Cloud software remains a large and growing business model. The pressure is falling on the older version of SaaS: products priced per seat, built around basic workflows and differentiated mainly by feature count.

The next phase of SaaS will combine subscription software with artificial intelligence, automation, proprietary data, usage-based pricing and measurable business outcomes.

SaaS at a Glance

Question	Answer
Is SaaS dead?	No
Is traditional SaaS under pressure?	Yes
What is causing the pressure?	AI agents, lower software creation costs, vendor consolidation and investor demands for profitable growth
Which SaaS companies are most exposed?	Undifferentiated tools with weak retention, limited data advantages and simple features that AI can reproduce
What is growing?	Cloud infrastructure, enterprise platforms, vertical software, AI-enabled applications and workflow automation
Should a company still build SaaS?	Yes, if it solves an expensive problem and creates defensible value beyond a thin AI interface

Gartner forecast worldwide spending on cloud application services, the category most closely associated with SaaS, to reach approximately $299.1 billion in 2025, up 19.2% from the previous year. The market is still expanding. What is changing is how SaaS products are built, priced and defended.

Why People Think SaaS Is Dead

The argument usually brings together several separate concerns:

AI can generate software faster. If a small team can build a basic application in days instead of months, more companies can enter previously crowded categories.
AI agents can perform tasks directly. Users may ask an AI agent to complete work instead of opening several separate applications.
Seat-based pricing may fit less well. Companies may prefer to pay for completed tasks, automated transactions or business outcomes rather than for every human or AI agent using the software.
Investors are demanding better economics. Growth is less convincing when customer acquisition costs are high, retention is weak and operating losses remain large.
Many SaaS products are easy to copy. Basic project management, note-taking, CRM, support and productivity features can be bundled into larger products or reproduced with general-purpose AI tools.

These pressures affect particular SaaS products. They do not remove cloud software as a business model.

What the Market Data Shows

Large software companies continue to generate significant revenue from cloud and subscription products.

Microsoft reported that its Microsoft Cloud business exceeded $168 billion in fiscal 2025 revenue, growing 23% year over year. Azure surpassed $75 billion in annual revenue, while Microsoft 365 Commercial Cloud grew 15%.

Salesforce reported $41.5 billion in fiscal 2026 revenue, up 10% year over year. Its remaining performance obligation reached $72.4 billion, providing a large base of contracted future revenue. Salesforce also reported growth from Agentforce and Data 360, showing how an established SaaS company is adding AI to its existing software model.

ServiceNow reported $3.877 billion in subscription revenue during the second quarter of 2026, representing 24.5% year-over-year growth. ServiceNow also said its AI business had passed $1 billion in annual contract value.

These figures do not describe a disappearing industry. They describe a market moving toward AI applications, business workflows, data platforms and software that can complete work on a customer's behalf.

What Is Changing About SaaS?

1. SaaS Is Moving From Systems of Record to Systems of Action

Traditional SaaS stored information and helped people manage workflows. Newer software is beginning to perform more of the workflow itself.

For example:

A CRM system stores customer records.
An AI-enabled CRM can qualify leads, draft follow-ups, update records and recommend next actions.
A service platform tracks support tickets.
An AI-enabled service platform can classify requests, retrieve knowledge and resolve routine cases.

The value shifts when software controls part of a business process instead of only displaying information about it.

2. Pricing Is Moving Beyond Per-Seat Subscriptions

Per-user subscriptions will remain common for collaboration and productivity products. AI adds costs and creates value that do not always align with the number of human users.

Potential pricing models include:

Subscription plus usage fees
Charges per automated task
Charges per API call or token
Revenue-share models
Outcome-based pricing
Enterprise platform commitments

Salesforce's reporting of "agentic work units" shows where this model may be heading. Software companies want to measure the work their systems complete, not only the number of people logged into them.

3. The Application Layer Is Becoming More Competitive

AI coding tools reduce the cost of building basic software. That makes it easier for companies to launch similar products quickly.

A simple interface or a thin wrapper around a large language model is unlikely to remain defensible for long. Stronger SaaS companies will need advantages such as:

Proprietary customer data
Deep workflow integrations
Regulatory or industry expertise
High switching costs
Trusted distribution
Reliable automation
Strong security and compliance
A large network or ecosystem

The application may be easier to build. The surrounding business system is still difficult to copy.

Which SaaS Companies Are Most at Risk?

The most exposed SaaS products often share several of these traits:

They solve a low-value problem.
Customers can replace them with a spreadsheet or general-purpose AI tool.
The product has weak retention.
The company depends on expensive paid acquisition.
Competitors offer nearly identical features.
The product does not own important data or workflow infrastructure.
Customers use the product infrequently.
Pricing is based on seats even though the value comes from automation.

A small SaaS company is not automatically at risk. A focused vertical product can remain durable when it becomes essential to a particular industry or workflow.

A payroll platform, healthcare compliance system or construction estimating tool may have stronger protection than a generic productivity application because it handles specialised data, processes and regulations.

What Types of SaaS Are Likely to Grow?

The strongest opportunities are likely to come from five areas.

AI-Enabled Enterprise Software

Large businesses need secure software that connects AI to internal data, permissions and business processes. Google reported that nearly 75% of its Google Cloud customers had used its vertically optimised AI products. It also reported that AI customers used more Google Cloud products on average than non-AI customers.

Vertical SaaS

Industry-specific software can address problems that general AI tools do not fully understand. Examples include healthcare, financial services, logistics, legal services, manufacturing and government software.

Systems That Control Workflows

Products become harder to replace when they manage approvals, payments, compliance, customer operations or other mission-critical processes.

Data and Integration Platforms

AI systems need reliable access to business data. Platforms that clean, govern, connect and secure that data may become more important as companies deploy multiple AI agents.

Usage-Based and Outcome-Based Applications

Software that charges for measurable work completed may align more closely with the value created by AI automation.

Should Entrepreneurs Still Build SaaS?

Yes, but "build a SaaS product" is no longer enough of a strategy.

A new SaaS company should answer four questions:

What expensive problem does the product solve?
Why can't a general-purpose AI assistant solve it?
What proprietary data, workflow or distribution advantage will improve over time?
How will pricing reflect the value created for the customer?

The strongest products may combine SaaS, an AI agent, a data platform and a managed service. Customers may not care what label the product carries. They care whether it reliably produces a valuable result.

That changes the founder's job. The goal is not to add another set of features to a crowded category. It is to own a difficult problem, connect deeply to the customer's work and make the product harder to replace as it learns more about that work.

Final Verdict

SaaS is not dead. Traditional SaaS is being rebuilt.

Cloud delivery, recurring revenue and software subscriptions remain viable. The exposed products are interchangeable tools with weak retention, limited data advantages and little connection to business outcomes.

For founders, the practical rule is simple: build software that AI makes more valuable, not software that AI can easily replace.