Christian Klein opened the SAP Sapphire keynote on Monday with a question that no chief executive of Europe’s most valuable technology company should need to ask. “Will SAP be a software company in the future?” The answer, delivered by SAP’s own AI assistant Joule at the end of the presentation, was that SAP is becoming a business AI company. The question was rhetorical. The 41 per cent decline in SAP’s share price over the past six months was not.

SAP unveiled what it calls the Autonomous Enterprise, a unified platform comprising more than 50 domain-specific AI assistants orchestrating over 200 specialised agents across finance, supply chain, procurement, human resources, and customer experience. The company announced a partnership with Anthropic to embed Claude as a primary reasoning engine across its AI-enabled portfolio. It launched a 100 million euro partner fund to accelerate deployment. It introduced seven vertical Industry AI solutions. It revealed agent-led migration tooling that it claims can reduce ERP transformation efforts by more than 35 per cent.

The announcement is the largest AI product launch in SAP’s 53-year history. It is also, unmistakably, a survival strategy.

The context

The 💜 of EU tech

The latest rumblings from the EU tech scene, a story from our wise ol’ founder Boris, and some questionable AI art. It’s free, every week, in your inbox. Sign up now!

SAP’s stock has lost more than a third of its value since peaking at 306.60 euros in July 2025. The January 2026 earnings call triggered a 15 per cent single-day decline, the steepest since 2020, after cloud revenue guidance fell short of expectations. The Q1 2026 results in April showed cloud revenue growing 27 per cent at constant currencies to 5.96 billion euros, but total revenue of 9.56 billion missed analyst forecasts, sending the stock down another six per cent in after-hours trading.

The problem is not SAP’s cloud business. Cloud ERP Suite revenue grew 30 per cent at constant currencies. Current cloud backlog reached 21.9 billion euros. The problem is the market’s judgement of what cloud revenue will be worth when AI agents start replacing the human users who generate per-seat licence fees.

In February, Workday’s chief technology officer traded his C-suite title for a technical staff role at Anthropic, a defection that crystallised the talent drain from legacy enterprise software to the AI companies building tools to displace it. The same month, a wave of agentic AI product launches from Anthropic, Salesforce, and Google erased roughly 285 billion dollars from SaaS company valuations in 48 hours, an event the financial press now calls the SaaSpocalypse.

SAP’s market capitalisation has fallen from more than 300 billion dollars to roughly 200 billion. The company that runs the back office of the global economy is being repriced as though it might not run the back office of the future.

The bet

The Autonomous Enterprise is SAP’s answer. The architecture has three layers. The SAP Business AI Platform provides the infrastructure for building, contextualising, and governing AI agents. The Autonomous Suite embeds those agents into core business applications. And Joule Work, a new interface, replaces the traditional screen-by-screen navigation with a conversational layer in which users describe a desired business outcome and Joule orchestrates the workflows, data, and agents to deliver it.

The most concrete demonstration is the Autonomous Close Assistant, which SAP says can compress a financial close process from weeks to days by automating journal entries, reconciliation, and error resolution across the entire cycle. The assistant does not replace the finance team. It orchestrates the agents that execute the tasks the finance team currently performs manually, while the humans approve, override, and govern.

This distinction matters. SAP is not selling AI that eliminates enterprise software. It is selling AI that makes enterprise software do more of the work that humans currently do inside enterprise software. The agents run within the same approval workflows, compliance frameworks, and governance controls that already govern human decisions in SAP systems. The lock-in does not weaken. It deepens.

The partnership

The Anthropic deal makes Claude a primary reasoning and agentic capability embedded across SAP’s solution portfolio. The integration goes beyond a standard API arrangement. Anthropic and SAP will collaborate to build custom agents and agentic workflows optimised for industries including public sector, healthcare, education, life sciences, and utilities.

SAP also announced expanded partnerships with Microsoft, bringing RISE with SAP onto Azure with deeper integration; Amazon Web Services, enabling zero-copy data sharing between SAP Business Data Cloud and Amazon Athena; Google Cloud, for bidirectional agent-to-agent interoperability; and Palantir, whose AIP platform will handle data migration scenarios alongside SAP’s agent-led transformation toolchain.

Anthropic is already embedding Claude into accounting software through its partnership with Xero, bringing AI-powered financial intelligence to millions of small businesses. The SAP deal extends that logic to the enterprise. Claude will power agents that take action for hundreds of thousands of SAP customers across finance, HR, procurement, and supply chain. A treasury manager can ask Joule to prepare a CFO briefing for a bank meeting and receive a completed presentation populated with live data, flagged risks, and analysis within minutes.

The question is whether the AI partner is also the AI competitor. Anthropic’s enterprise revenue has grown to the point where more than 1,000 businesses spend over a million dollars a year on its services. Its marketplace sells Claude-powered tools that perform functions SAP’s own applications handle. SAP is embedding the technology of a company whose long-term trajectory is to make SAP’s traditional product unnecessary.

The migration

SAP holds one card that no AI startup can match. Roughly 17,000 companies are still running SAP ECC, the legacy ERP system whose mainstream maintenance ends in December 2027. Extended support runs to 2030, but at higher cost and with diminishing returns. Every one of those companies must migrate to S/4HANA Cloud or find an alternative. Most will migrate.

The Autonomous Enterprise announcement converts that forced migration into an AI upsell. RISE with SAP customers will receive three Joule Assistants activated within their first year. SAP GROW customers get access to the full assistant portfolio at onboarding. The agent-led transformation tooling, built with Palantir, automates system analysis, code remediation, configuration, and testing at scale, reducing the effort, cost, and risk that have kept thousands of companies on the legacy platform.

SAP is using the deadline it created to sell the AI platform it just built. The 17,000 holdouts are not just a migration challenge. They are a captive market for the most expensive AI product launch in enterprise software history.

The market

The five largest technology companies are collectively spending more than 650 billion dollars on AI infrastructure in 2026, and the enterprise software companies that sit on top of that infrastructure are racing to prove that AI agents generate revenue rather than destroy it. Salesforce’s Agentforce has reached 540 million dollars in annual recurring revenue across 18,500 enterprise customers. ServiceNow is positioning itself as the AI control tower for IT and HR workflows.

Oracle has assembled more than 16 billion dollars in data centre financing to pivot toward AI infrastructure, a bet that the future of enterprise technology is measured in compute capacity rather than software licences. SAP’s approach is different. It is not building data centres. It is embedding agents into the business processes that the data centres ultimately serve.

The strategic logic is that AI will commoditise software interfaces but not business process logic. Anyone can build a chatbot. Not anyone can build a chatbot that understands the intercompany elimination rules in a multinational financial close, or the procurement compliance requirements of a German automotive manufacturer, or the lot-tracing regulations in pharmaceutical supply chains. SAP’s 53 years of accumulated process knowledge is the moat. The AI agents are the means of monetising it.

The question

Anthropic has reached a one trillion dollar implied valuation on secondary markets, roughly five times SAP’s current market capitalisation. The company that SAP just made its primary AI partner is worth more than SAP. The company that builds the reasoning engine is valued higher than the company that owns the business processes the engine reasons about.

That valuation gap is the market’s current answer to Klein’s question. The market believes that AI companies will capture more value than the enterprise software companies AI is embedded into. SAP is betting that the market is wrong, that the value accrues to whoever owns the process, the data, and the governance layer, not whoever builds the model.

The Autonomous Enterprise will take years to validate. The 200 agents and 50 assistants are launching in phases through 2026 and into 2027. The Industry AI solutions roll out quarterly. The Anthropic integration is in its early stages. The migration deadline will force millions of decisions over the next 18 months about whether to adopt SAP’s AI stack or look elsewhere.

Klein asked whether SAP will be a software company in the future. The honest answer is that SAP does not know. What it knows is that 300,000 customers run their most critical business operations on SAP systems, and that the only way to keep them is to make those systems do things that used to require the people who operate them. The Autonomous Enterprise is not a product launch. It is a wager that the company which automates the work will remain more valuable than the companies whose workers it automates away.