The new enterprise AI stack: own the agents, own the data
Enterprise AI acquisitions and data lakehouse integration now describe a strategic shift in how large software vendors design their stacks: instead of relying on external AI APIs and scattered data warehouses, they are buying agentic AI platforms and open data lakehouses so customers can own specialized models and run them directly on governed, unified data across their operations, analytics, and security systems. That is the real story behind Datadog acquiring Adaptive ML and SAP completing its purchase of Dremio. Together, these moves show that integrated AI agents and data platforms are no longer nice-to-have extensions; they are becoming the heart of enterprise software strategy. If your stack still treats AI as a bolt‑on to legacy data silos, these deals are a warning sign: your vendors—and your competitors—are re-architecting around owned models and open, real-time data.
Datadog’s Adaptive ML bet: turning observability into agentic AI
Datadog’s acquisition of Adaptive ML is not about sprinkling generative features into dashboards; it is about turning observability into an agentic AI platform. Datadog openly argues that off‑the‑shelf models are easy to deploy but insufficient for complex, domain‑specific production challenges, and that sustained AI value depends on tuning, evaluating, and refining models against real‑time signals. Adaptive ML’s Adaptive Engine does exactly that: it fine‑tunes open models with reinforcement learning and synthetic data, evaluates them with custom AI judges and A/B tests, and closes the loop by feeding production signals back into training so organizations can build specialized models they own that improve with each deployment. When these capabilities are combined with Datadog’s observability and security data, the company expects AI agents to move beyond investigation and alerting toward systems that can learn, adapt, and act in support of autonomous operations. For customers, this means your monitoring stack is on its way to becoming a controlled environment for internal AI agents, not a passive source of metrics.
SAP + Dremio: data lakehouse integration for agentic AI at scale
SAP’s completed acquisition of Dremio is an equally clear signal that the battle for enterprise AI will be won in the data layer. Dremio brings an open, high‑performance data lakehouse platform into SAP’s Business Data Cloud, letting customers combine SAP and non‑SAP data to run analytical and AI workloads in real time with no data movement or conversion and with improved economics for enterprise analytics. SAP is explicit about the problem this tackles: despite rapid advances in AI, many projects stall because organizations cannot effectively access, govern, or contextualize the data needed for meaningful results. By adopting Dremio’s Apache Iceberg‑native architecture, SAP positions its Business Data Cloud as an open enterprise lakehouse that can support real‑time analytics and agentic AI across fragmented data estates. According to Philipp Herzig, CTO of SAP SE, “Enterprise AI doesn’t stall because the models aren’t good enough; it stalls because the data isn’t ready for AI agents.” In practice, this turns SAP from an application vendor into a single, open platform for AI‑ready intelligence.
Why enterprises want to own AI models—not rent them via APIs
Both acquisitions share a deeper motivation: big customers are tired of renting intelligence through opaque third‑party APIs. Datadog highlights that off‑the‑shelf models cannot handle complex production problems and that value comes from models tuned to an organization’s live signals and judged against its business outcomes. Adaptive Engine’s focus on fine‑tuning open models, generating synthetic data, and closing the training loop from production is about control—enterprises own the models and the improvement cycle, rather than being locked into a vendor’s black box. On the data side, SAP is investing in openness, governance, and performance so that agentic AI can scale without sacrificing transparency or compliance. As organizations expand agentic AI initiatives, Dremio gives SAP a stronger foundation for unlocking value from both SAP and non‑SAP data while maintaining the qualities increasingly required for enterprise AI success. If you are serious about AI, this is your new requirement list: model ownership, data explainability, and platform‑level integration.
What this consolidation means for your enterprise stack
The message from these enterprise AI acquisitions is blunt: integrated data platforms and AI agents are becoming table‑stakes for enterprise software vendors. SAP now has a more open, AI‑ready data platform that combines business context, real‑time analytics, and simplified data integration, directly targeting the barrier of fragmented, disconnected data. Datadog, meanwhile, is turning observability data into the training ground for specialized agents built and owned by customers, with plans to fold Adaptive ML’s capabilities into its platform over time and support its push toward autonomous operations. With the Dremio transaction complete, SAP has strengthened its ability to deliver a single open platform from raw data to governed intelligence. In short, the era of bolt‑on AI widgets is ending. If your stack still treats AI and data as separate projects, these moves should force a strategy rethink: future‑ready enterprises will demand vendors who bring agentic AI and unified data lakehouse integration as part of the core product, not as optional extras.
