SaaSpocalypse Unlikely as CLSA Says AI Spares Core Software

Fears of a SaaSpocalypse, the worry that artificial intelligence (AI) will gut demand for software-as-a-service (SaaS, subscription software delivered over the cloud) and the firms that deploy it, are running ahead of the evidence. CLSA, the Asia-focused brokerage, told clients that most major SaaS vendors either held or raised revenue and margin guidance for the coming year and beat profit forecasts last quarter, showing no visible dent in demand yet.

Beneath that headline the brokerage drew a sharper map. AI is not pressing on every kind of software at the same speed, and the latest results from Salesforce show both the demand holding up and the business model underneath it starting to move.

Why CLSA Is Pushing Back on the SaaS Doom Trade

The doom trade has been brutal. India’s technology benchmark has been one of the year’s worst sectoral performers, and the same anxiety that AI tools can automate offshore work has battered enterprise software valuations worldwide. Against that backdrop CLSA’s note reads as a deliberate lean against the crowd.

The brokerage’s core finding is plain: across the major SaaS names that reported in the latest quarter, earnings per share (EPS, a company’s profit divided by its shares) came in ahead of consensus, and forward guidance moved up or held rather than down. If AI were already eating into subscription demand, CLSA argues, it would be showing in those numbers first. It is not. The note circulated as the Nifty IT index constituents and weights ticked up about 2% in early trade, even with the index sitting down roughly 26% for the year.

That gap between a wrecked tape and steady fundamentals is the whole story. Investors have priced a collapse that the operating results have not yet delivered.

The Three-Tier Map of What AI Can and Cannot Replace

CLSA’s framework is what makes the note more than a cheer. It splits SaaS into three buckets and asks a single question of each: can AI substitute what this software produces, or only sit on top of it?

The split matters because it tells investors where to be brave and where to be careful. A blanket bet against software misses that one of these tiers is built to resist exactly the kind of disruption the bears are pricing.

Where the Moat Holds and Where It Cracks

Systems of Record Keep Their Lock

The argument for the record-keepers rests on a single technical fact. AI is probabilistic; it produces a likely answer, not a guaranteed one. Systems of Record need deterministic output, the same correct number every time a ledger is read or an insurance claim is priced. That mismatch is the moat.

So AI does not replace SAP, Snowflake, Databricks, Guidewire or Duck Creek. It layers a conversational interface on top of them, which can actually deepen the lock-in by making the underlying data more useful. The platform that owns the source of truth gets stronger, not weaker.

Engagement and Workflow Face the Knife

The exposed tiers are the ones whose value is the output itself. When the deliverable is a screen, a workflow or a generated app, AI can recreate that deliverable directly, and the case for a separate subscription thins out. That is where CLSA sees genuine pressure building on ServiceNow, Adobe, Pega, Mendix, OutSystems, Datadog and Sitecore.

None of that means those firms are doomed. It means their moat is the part of the stack AI is best placed to copy, and they have to keep moving up the value chain to stay ahead of the very models they now embed. The bear case, captured well in the argument that traditional seat-based SaaS is dying, lands hardest on this group.

Salesforce’s Numbers Cut Both Ways

Salesforce, the customer-relationship-management vendor, is the cleanest live test of the thesis, and its latest quarter supports the demand side hard. The company posted record figures and lifted its outlook rather than trimming it.

• $11.1 billion in first-quarter fiscal 2027 revenue, up 13% year on year.
• Up about 50% in non-GAAP diluted EPS, to $3.88 a share.
• More than $1 billion in Agentforce annual recurring revenue (ARR), inside $3.4 billion of combined AI and data ARR.

The adoption signals are just as strong. Salesforce booked 3.8 billion agentic work units, the discrete tasks its AI agents complete in production, up 111% quarter on quarter, and Agentforce customers in production grew by half. The full detail sits in Salesforce’s record first-quarter fiscal 2027 results, which raised the full-year revenue range to $45.9 billion to $46.2 billion.

Look closer and the same quarter carries a warning. Salesforce headcount grew mostly in sales, while engineering headcount has stayed broadly flat for two years on AI-led efficiency. The company also stopped publishing role-by-role certification data for its system-integrator partners, a quiet tell that AI is reshuffling who does what inside services firms. The adoption curve, mapped in Salesforce’s Agentforce adoption milestones, is real; so is the staffing message underneath it.

The Indian IT Names With the Most at Stake

For the services industry that builds and runs these platforms, the framework reads as a watchlist. CLSA ranks who is most levered to SaaS implementation work, and therefore most tied to which tiers hold.

• Among largecap IT services firms, Accenture, Capgemini, Wipro and Cognizant carry the highest SaaS exposure.
• Persistent Systems ranks highest on SaaS exposure among the mid-caps.
• LTIMindtree holds the strongest ServiceNow capability in the largecap pool, while Infosys stands out on SAP competency.
• Hexaware leads on Guidewire; Snowflake, Birlasoft and Persistent rank best among the mid-caps.

The nuance is in the laggards. CLSA reads recent SAP project delays at HCL Technologies and Wipro as client-specific issues rather than a sign of trouble with the SAP platform itself. That distinction is the difference between a stock-specific wobble and a sector-wide crack, and the brokerage is firmly on the wobble side for now.

The read-through is that firms anchored to Systems of Record work, the SAP, Snowflake and Guidewire deployments, sit on steadier ground than those whose revenue leans on building front ends and workflows AI can increasingly generate on its own.

The Pricing Shift From Seats to Consumption

The most durable change in the note is not about demand at all. It is about how software gets billed. AI has forced vendors away from charging per user seat toward charging for usage and outcomes, the seat-based to consumption-based move that touches record-keepers and engagement tools alike.

That shift cuts in two directions for the services layer. Consumption pricing rewards firms that can drive real usage, expanding implementation and product-engineering work as agents scale. It also hands buyers a lever to pay only for what they consume, which can compress spend if AI does the same job with fewer seats and fewer hands.

If the switch lifts usage as fast as vendors hope, the integration work that IT services firms live on grows with it. If buyers use consumption pricing mainly to trim budgets, the same shift starts squeezing the services tier before it touches the platforms.

Frequently Asked Questions

What is the SaaSpocalypse?

SaaSpocalypse is the market’s name for the fear that artificial intelligence and AI agents will hollow out demand for software-as-a-service and the IT services firms that deploy it. CLSA argues the fear is overblown so far, because major SaaS vendors beat earnings and held or raised guidance in the latest reported quarter.

Which software is most exposed to AI substitution?

CLSA flags Systems of Engagement and Systems of Workflow platforms, whose output AI can recreate directly. Named examples include ServiceNow, Adobe, Pega, Mendix, OutSystems, Datadog and Sitecore. Systems of Record such as SAP and Snowflake are seen as better protected because they need deterministic output.

Which IT services companies have the highest SaaS exposure?

Among largecaps, CLSA names Accenture, Capgemini, Wipro and Cognizant. Persistent Systems leads the mid-caps. LTIMindtree ranks strongest on ServiceNow work, Infosys on SAP, and Hexaware on Guidewire.

Did Salesforce’s latest results show AI hurting demand?

No. Salesforce reported record first-quarter fiscal 2027 revenue of $11.1 billion, up 13%, raised full-year guidance to $45.9 billion to $46.2 billion, and crossed $1 billion in Agentforce annual recurring revenue, all signs that AI is adding to demand rather than subtracting from it.

What is the shift from seat-based to consumption-based pricing?

AI is pushing software vendors away from charging per user seat toward charging for usage or outcomes, such as tokens consumed or tasks completed. The change reshapes vendor revenue and the implementation work that IT services firms sell.

Disclaimer: This article is for informational purposes only and is not investment advice. It covers equities and securities, which carry risk of loss, and the views attributed to CLSA are the brokerage’s own. Readers should consult a qualified financial professional before making investment decisions. Figures are accurate as of publication.