SAP (NYSE: SAP) bets €1bn on Prior Labs and Dremio as enterprise AI strategy widens

SAP SE (NYSE: SAP) announced on May 4, 2026 that it has signed two definitive acquisition agreements on the same day, taking over Freiburg-based Prior Labs to build a European frontier AI lab focused on tabular foundation models and acquiring Austin-based Dremio to unify SAP and non-SAP data inside its Business Data Cloud. SAP has committed more than €1 billion over four years to scale Prior Labs alone, while financial terms of the Dremio transaction were not disclosed. Both deals remain subject to regulatory approval, with Prior Labs expected to close in Q2 or Q3 2026 and Dremio in Q3 2026. The dual announcement lands as SAP shares trade near $172 on the NYSE, roughly 45 percent below a 52-week high of $313.28 set in July 2025, sharpening the question of whether this is a moment of strategic conviction or a defensive response to mounting investor doubt about the company’s AI trajectory.

Why is SAP buying Prior Labs and Dremio together and what does the dual deal signal about its enterprise AI roadmap?

The two transactions are not independent. They form a coordinated stack. Prior Labs supplies the model layer in the form of tabular foundation models, a category SAP has already validated internally with SAP-RPT-1. Dremio supplies the data layer in the form of an Apache Iceberg-native lakehouse that can serve both SAP and non-SAP data into agentic workflows. Joule, SAP AI Core and SAP Business Data Cloud are the productization channels. By announcing both acquisitions on the same day, SAP is signaling that it has accepted a thesis that enterprise AI does not stall on model quality but on the readiness of structured business data to be reasoned over at scale.

This matters because the dominant narrative around enterprise AI has been built around large language models, and SAP has been a relative laggard in that conversation. Tabular foundation models are a genuinely different category. They are purpose-built for structured business data such as payment delays, supplier risks, customer churn, and upsell propensity, where statistical reasoning matters more than language fluency. By framing the Prior Labs acquisition around this category, SAP is attempting to redefine the playing field rather than compete head-on with Microsoft, Salesforce and Oracle on generative AI assistants alone. The strategic intent is to own a category before it becomes obvious.

The competitive implication is that SAP is choosing to build, not rent. Most enterprise software vendors have so far positioned themselves as customers of frontier model providers. By committing more than €1 billion to scale Prior Labs into an independent frontier AI lab in Europe, SAP is taking a different stance, one that resembles the early Anthropic and Mistral playbooks rather than the application-layer integrators. The risk in this approach is well known. Frontier AI capital intensity has historically swallowed budgets larger than €1 billion across four years, and SAP will need to demonstrate that tabular foundation models are durable enough as a category to justify a model-building stance rather than a partnership stance.

How does Prior Labs strengthen SAP’s position in tabular foundation models against hyperscaler competition?

Prior Labs brings TabPFN, a model series with more than three million open-source downloads, the leading position on the TabArena benchmark with TabPFN-2.6, and a research credential established through publication in Nature. The cofounders Frank Hutter, Noah Hollmann and Sauraj Gambhir lead a team SAP describes as recruited from Google, Apple, Amazon, Microsoft, G-Research, Jane Street, Goldman Sachs and CERN. Yann LeCun and Bernhard Schoelkopf will sit on the scientific advisory board, lending the lab the kind of academic gravity that SAP has historically lacked in the AI conversation.

The substantive technical claim worth interrogating is that TabPFN-2.6 reportedly matches the accuracy of a four-hour automated machine learning pipeline instantly in a single model. If validated at enterprise scale, this collapses a meaningful piece of the data science stack. It implies that customers running churn prediction, supplier risk scoring or receivables forecasting could move from custom model training to in-context learning, where the model adapts on the fly to new data records without retraining. The downstream effect on enterprise data science teams, on third-party AutoML vendors, and on the consulting partners that build bespoke predictive pipelines could be significant.

The competitive risk for SAP is that hyperscalers can replicate the category. Google, Microsoft and Amazon all have the capital, the research depth and the data infrastructure to build or buy comparable tabular models. SAP’s structural advantage is access to the world’s ERP-grade tabular data through its installed base and the contextual metadata that surrounds it. The harder question is whether SAP can keep Prior Labs’ open-source momentum intact while monetizing the model inside a proprietary stack. The press release commits to continuing the open-source strategy. Open-source commitments under enterprise ownership have a mixed track record across the industry, and Prior Labs’ developer community will be watching governance signals closely in the months after close.

What does the Dremio acquisition mean for SAP Business Data Cloud and the lakehouse competitive landscape?

Dremio gives SAP an Apache Iceberg-native lakehouse, federated query infrastructure, and stewardship of Apache Polaris and Apache Arrow. The strategic logic is that SAP Business Data Cloud cannot credibly serve agentic AI workloads if it remains a closed environment. Customers run their data across Snowflake, Databricks, AWS, Azure and on-premise systems, and asking them to move data into SAP-native formats has been a long-standing friction point. By acquiring an Iceberg-native platform, SAP is conceding that the open table format has won the lakehouse standards battle and is positioning itself as a participant in that ecosystem rather than a competitor to it.

The competitive implication is sharpest for Snowflake and Databricks. Both have built their AI strategies around being the data foundation for enterprise agents, with Snowflake’s Cortex stack and Databricks’ Mosaic AI integrations. SAP now offers a credible alternative for any enterprise where ERP data is the gravitational center, particularly in manufacturing, supply chain, and finance functions. The serverless and elastic economics of Dremio also matter. Enterprise analytics workloads have historically been priced on fixed capacity, and SAP’s pivot to consumption-based lakehouse economics aligns its data layer with how customers increasingly want to buy compute.

The integration risk is real. SAP HANA Cloud is an in-memory engine optimized for transactional and operational performance, while Dremio is built for federated analytics over object storage. Combining them into a coherent product surface without confusing customers about which engine to use for which workload will require careful product management. The history of large enterprise software acquisitions includes more cautionary tales than success stories on this dimension, and SAP’s own track record with prior acquisitions such as Qualtrics and Concur has been uneven on integration speed. Customers and partners will judge the deal on whether the unified data layer materializes as promised or whether it becomes another optional module in the SAP catalog.

How should investors read the timing of the deals against SAP’s share price weakness?

SAP shares were trading around $172 on May 4, 2026, with a 52-week range of $160.66 to $313.28 and a current market capitalization of roughly $211 billion. The stock is down approximately 45 percent from the July 2025 high and is sitting only modestly above its April 2026 low. Q1 2026 results beat consensus on the top and bottom line, but analysts flagged a deceleration into Q2 and modestly softer 2026 outlook commentary, which has weighed on sentiment. Average analyst price targets remain materially above current levels, with consensus targets in the $258 to $340 range depending on the panel, suggesting the sell side still views the current price as an entry rather than a top.

Two acquisitions on the same day, one of which carries an explicit €1 billion four-year capital commitment, will be read by some investors as a confidence signal and by others as a defensive move. The bullish reading is that SAP has the cash flow and balance sheet flexibility to invest counter-cyclically while peers consolidate, and that the AI strategy is finally crystallizing into a coherent stack. The bearish reading is that SAP is responding to a perception problem, that the market has been pricing in AI displacement risk to enterprise application software, and that management is buying narrative momentum rather than near-term earnings accretion. Both readings can be partly true. What will matter for the stock over the next two to four quarters is whether SAP can communicate concrete revenue hooks from the Prior Labs and Dremio capabilities, ideally tied to specific customer wins or upsell metrics inside the existing installed base.

The deal terms also matter for capital discipline. Prior Labs’ headline number is a four-year investment commitment rather than an upfront purchase price, which suggests SAP has structured the financial exposure as a research and development line item rather than a single goodwill event. Dremio’s undisclosed terms make the total deal value harder to assess, though Dremio’s most recent private valuations placed it in the low billions of dollars range. Investors should watch for clarification in SAP’s next quarterly disclosure, particularly on goodwill recognition, integration costs and any guidance adjustment to operating margin trajectory.

What are the regulatory, integration and execution risks SAP must navigate before close?

Both transactions are subject to regulatory approval. Prior Labs is a German entity acquired by another German entity, which simplifies one dimension of antitrust review, but the European Commission has been actively scrutinizing AI-related transactions, particularly those that consolidate frontier research capability inside large incumbents. The Commission’s Article 22 referral mechanism and its evolving stance on AI competition mean Prior Labs’ approval timeline cannot be taken for granted, even if the deal’s relative size is modest. The Dremio transaction adds a transatlantic dimension, with US Department of Justice and Federal Trade Commission review likely, alongside potential interest from the Committee on Foreign Investment in the United States given Dremio’s customer base in regulated industries.

Integration risk is the second major exposure. Prior Labs’ commitment to operate as an independent unit inside SAP is a deliberate structural choice, and one that will be tested when productization pressure rises. AI labs typically thrive on research velocity that does not align neatly with quarterly enterprise software release cycles, and the tension between independence and integration will require active management. The press release frames Prior Labs as an autonomous lab with a direct path to productization through SAP AI Core and Joule, which is a careful balance to articulate but a harder one to maintain in practice over multiple years.

Execution risk also extends to talent retention. Frontier AI researchers have alternative employment options at compensation levels that few enterprise software companies have historically matched. SAP will need to design retention packages, research autonomy commitments and publication freedoms that keep the Prior Labs team intact through the integration period. Loss of two or three senior researchers in the first 18 months would materially weaken the strategic value of the acquisition, and this is a risk that boards and investors should track closely.

What are the key takeaways from SAP’s acquisitions of Prior Labs and Dremio for the company, its competitors, and the enterprise AI industry

• SAP is committing to build rather than rent enterprise AI capability, with a more than €1 billion four-year investment in Prior Labs that signals a frontier AI lab ambition unusual for an application software incumbent.
• The dual announcement of Prior Labs and Dremio on the same day is a coordinated model and data layer play, not two independent transactions, and should be evaluated as a single strategic stack.
• Tabular foundation models are positioned as the category SAP intends to own globally, an attempt to define a new playing field rather than compete head-on with hyperscalers on large language models.
• Dremio gives SAP Business Data Cloud an Apache Iceberg-native foundation and concedes that open table formats have won the lakehouse standards battle, a meaningful positioning shift for SAP.
• Snowflake and Databricks face a credible new competitive vector for enterprise AI workloads where ERP data is the gravitational center, particularly in manufacturing, supply chain and finance.
• The timing of the deals against SAP shares trading roughly 45 percent below the July 2025 high will be read by some investors as confidence and by others as a defensive narrative correction.
• Capital discipline will be tested. Prior Labs’ €1 billion commitment is structured as a four-year investment, which softens the immediate balance sheet impact but raises ongoing operating expense visibility.
• Regulatory approval is the near-term gating risk, with European Commission AI scrutiny and US transatlantic review both likely to extend the closing timeline beyond the indicated Q2 or Q3 2026 windows.
• Integration risk concentrates around Prior Labs’ independence promise and Dremio’s coexistence with SAP HANA Cloud, two integration designs that have failed at other enterprise acquirers in the past.
• Talent retention at Prior Labs is a critical execution variable, and the loss of senior researchers in the first 18 months would materially weaken the strategic value of the acquisition.