The SaaS earnings message: AI changes the rules, not the whole game

Key takeaways

• AI is boosting infrastructure and data platforms, but pressuring traditional software expectations.

• Investors are rewarding clear usage growth and questioning seat-based software models.

• The key test is not AI features, but whether customers pay more for them.

Artificial intelligence (AI) is starting to split the technology market into two camps. On one side are the companies selling the picks, shovels and data roads. On the other are software companies that must prove their old toll booths still work. The latest earnings round showed that split clearly.

The old software bargain is being tested

Software as a service (SaaS) has been one of the great business models of the past two decades. Companies paid regular subscriptions to use cloud-based tools, often priced by the number of employees, or “seats”. More employees meant more seats. More seats meant more revenue. Very civilised. Almost suspiciously so.

AI complicates that model. If an AI agent can perform work that used to require several users clicking through software screens, then pricing by seats may become less powerful. A customer may still need the software, but may not need to pay for as many human users in the same way. This is why our AI SaaS disruption list focuses on companies where AI could either strengthen the product or weaken the old seat-based pricing model.

Salesforce sits right at the centre of that debate. The company reported first-quarter fiscal 2027 revenue of 11.13 billion USD and adjusted earnings per share of 3.88 USD, both ahead of expectations, as compiled by Bloomberg. It also highlighted momentum in Agentforce, its AI agent platform. Yet its second-quarter revenue forecast was slightly below market expectations, and investors focused on the risk that AI could change traditional software demand.

The lesson is simple. Adding AI features is no longer enough. Investors want evidence that AI increases customer spending, improves retention and protects margins. In simple terms: does it make the product more valuable, or just more expensive to run?

Data platforms look like the new control room

Snowflake had a very different market reaction. Snowflake helps companies store, manage and analyse large amounts of data in the cloud. That matters because AI is only as useful as the data it can safely use. A very clever model with messy data is like a sports car with cooking oil in the engine.

Snowflake reported first-quarter revenue of 1.39 billion USD, up about 34% from a year earlier, and raised its full-year product revenue guidance. It also expanded its partnership with AWS through a five-year, 6 billion USD commitment. The market liked the combination: stronger growth, clearer AI demand and deeper access to cloud infrastructure.

Snowflake closed at 239.20 USD on 28 May 2026, up 36.4%. That move was large because it changed the narrative. Snowflake had been treated as a high-growth software company facing tougher scrutiny. The new results suggested it may also be a key data layer for enterprise AI.

That is an important distinction. Some software companies may be disrupted by AI. Others may become the plumbing that makes AI useful inside companies. Plumbing rarely sounds exciting, but investors have learned not to laugh at pipes.

Hardware is getting paid first

Dell shows the other side of the story. Dell is best known for personal computers, but its more important AI role today is in servers, storage and infrastructure. These are the physical systems companies need to train, run and deploy AI models.

Dell reported first-quarter revenue of 43.84 billion USD and raised its full-year revenue forecast to 165 billion USD to 169 billion USD. The company also now expects 60 billion USD of AI server revenue for fiscal 2027. Shares rose sharply after hours as investors reacted to the stronger AI server outlook.

This is why hardware and infrastructure have looked more straightforward than SaaS in the AI trade. Companies need computing power before they can redesign workflows. The bills arrive early. The productivity gains may arrive later, possibly with a calendar and a polite apology.

For investors, that creates a timing gap. Infrastructure companies can show demand now. Software companies must show that AI becomes a paid product, not just a feature customers expect for free.

Risks to watch

The biggest risk is overconfidence. AI infrastructure demand is strong, but server businesses can have lower margins and more cyclical demand than pure software. If customers pause spending, hardware earnings can slow quickly.

For SaaS, the risk is pricing pressure. If AI reduces the need for traditional user licences, companies must shift towards usage-based pricing, where customers pay based on activity or value delivered. That can work, but transitions are rarely neat. Investors should watch renewal rates, large customer growth, margins and whether AI products create real extra revenue.

A third risk is customer fatigue. Every company now says it has an AI strategy. Some do. Some have a chatbot wearing a small hat. The market will become less forgiving when AI claims do not show up in sales, cash flow or customer behaviour.

Investor playbook

• Separate AI beneficiaries by layer: chips, servers, cloud, data, applications and services.
• Watch whether AI revenue is new spending or just a renamed old product.
• Compare share price moves with earnings quality, not only headline growth.
• Keep position sizes sensible when expectations move faster than the business.

The useful question

The latest earnings round does not say that SaaS is broken. It says the easy version of the software story is over. In the old world, investors could often reward recurring revenue, high margins and steady customer growth. In the AI world, they must ask what the software actually does, how essential it is, and whether AI makes it stronger or easier to replace.

That is a healthier question. Dell shows that infrastructure is getting paid first. Snowflake shows that trusted data may become more valuable. Salesforce shows that even strong software leaders must keep proving the model. AI is not killing software. It is turning the lights on, and a few old business models may prefer the room slightly darker.

This material is marketing content and should not be regarded as investment advice. Trading financial instruments carries risks and historic performance is not a guarantee of future results.