At SAP Sapphire 2026, SAP consolidated eighteen months of M&A, product development, and AI strategy into a single integrated vision it dubbed the “autonomous enterprise.” JPMorganChase CFO Jeremy Barnum confirmed live on stage that the bank is migrating its general ledger to SAP’s latest version. Bayer, Novartis, Takeda, Ericsson, and H&M anchored the production credibility.

Top Developments At SAP Sapphire 2026

  • The model layer consolidated into a portfolio architecture. SAP Business AI Platform unifies BTP, Business Data Cloud, and Business AI, anchored by: 1) Anthropic’s Claude as the primary reasoning model; 2) the SAP’s-RPT-1 tabular foundation model (which shipped today); 3) the pending Prior Labs acquisition for non-SAP tabular workloads; and 4) partnerships with Mistral and Cohere for sovereign options. Large language models, tabular foundation models, and domain-specific models each handle the workloads they are structurally best at. Claude as an anchor creates ecosystem bifurcation where Microsoft-OpenAI and SAP-Anthropic become the two dominant alignments, leaving Salesforce, Workday, and Oracle as multimodel-neutral. Important caveat: The portfolio strategy depends on continued partnership stability with Anthropic and on Prior Labs closing on schedule in Q2–Q3 2026.
  • Agent governance became the decade’s control point. SAP AI Agent Hub, built on the SAP LeanIX foundation, which is already frequently deployed across the Fortune 500, governs SAP and non-SAP agents with verified-agent enforcement, Cloud ALM telemetry, Signavio agent mining, and SuccessFactors workforce mapping. It will be generally available in Q3 2026, bundled at no additional charge. Microsoft made the same move earlier in May with the launch of Agent 365. Important caveat: AI Agent Hub’s ability to govern competitor agents at parity with native SAP agents is a marketing claim until demonstrated.
  • Company memory introduces the customer-specific intelligence flywheel. Built on SAP Signavio, company memory ingests process models, policies, Teams chats, and email approval chains into structured process atoms that govern agent behavior; every exception updates the memory, and every override teaches the next agent. Horizontal AI platforms cannot replicate this without 50-plus years of ERP context. Important caveat: zero public production references at announcement.
  • Joule Work makes screen-based ERP architecturally obsolete. The conversational front door generates application experiences on demand (mobile general availability now, desktop GA H2 2026, full agent-to-agent interoperability Q4 2026). SAP is no longer competing on application design; it is competing on dynamic context generation. Screen-by-screen ERP design becomes obsolete over 36 months, forcing enterprises to rethink user training, change management, and information architecture. The Q4 2026 agent-to-agent GA date is the architectural milestone that matters. Important caveat: Earlier GA dates cover demo-ready capabilities, not full enterprise readiness at Fortune 500 scale.
  • SAP made the agent runtime free through year-end 2026. Joule Studio 2.0 is free, the Joule agent runtime is free through December 31, 2026, and agent-to-agent interoperability is free with no cap. This is the most aggressive commercial move SAP has made in a decade. The Microsoft Teams 2017 playbook applied to enterprise agents. Important caveat: The 2027 pricing cliff is unmodeled in most customer 2026 budgets because no post-promotion pricing has been disclosed.
  • Hybrid landscape support reverses three years of cloud-only positioning. A significant subset of Joule assistants and agents will operate on S/4HANA on-premises and ECC for RISE customers. ECC and on-prem customers committed to a cloud migration path gain access to select AI scenarios during transition, resolving the “wait until we are in the cloud” objection to AI investment. This is the largest concession SAP has made to its installed base in three years. The DSAG pressure won.

A Calibrated Caution For Tech Leaders

The vision is credible. The execution is partial. Most of the 224 agents and 51 assistants sit in mixed GA, early-adopter, and preview status. Claude as the anchor creates concentration risk that becomes board-level in regulated industries within 24 months. Forrester research finds that 21% of enterprise SaaS decision-makers cite vendor lock-in as a top commercial concern, a risk that compounds with each architectural layer of consolidation. Commit at the architectural pattern level. Pilot at the product level. Define explicit go and no-go criteria before declaring SAP your strategic AI architecture for 2030.

Forrester clients can engage with our analyst team to pressure-test SAP within their broader enterprise AI and applications strategy, including model portfolio design, agent governance architecture, and vendor concentration risk. Schedule a guidance session to align your 2027–2030 enterprise architecture with the post-Sapphire reality.