Legacy databases as cash engines in a cloud-first world
Legacy database monetization describes how long‑established database platforms keep generating recurring license and support revenue for vendors, even as enterprises shift spending toward cloud, analytics, and AI‑optimized data services that threaten to replace or bypass those older systems. Microsoft SQL Server is a prime example of this tension. On one hand, SQL Server barely featured at Microsoft’s AI‑heavy Build conference, and leadership changes signal that attention has moved toward Azure Data, Fabric, and open source services. On the other hand, analysts say SQL Server still brings in billions in on‑premises database revenue and continues to grow. That creates a powerful financial motive to preserve compatibility, extend support, and encourage slow, managed cloud migration rather than a rapid cut‑over to cloud‑native alternatives that might disrupt this reliable income.
SQL Server’s revenue gravity and the Azure on-ramp
SQL Server revenue strategy hinges on two linked goals: protect on‑premises income and use it as a feeder into Azure services. According to Gartner’s Adam Ronthal, Microsoft makes around USD 15 billion (approx. RM69 billion) from the on‑prem DBMS market, “largely from SQL Server,” with about 33 percent share and high single‑digit growth. Walking away from that revenue is not an option, even if strategic focus is shifting toward Azure and Fabric. Instead, Microsoft offers flexible licensing that lets customers bring SQL Server to clouds, including AWS RDS, without paying twice. It also builds Azure SQL and Fabric’s SQL engine on the same core technology, promising an easier upgrade path. The result is classic cloud migration economics: keep the legacy base happy, then gently nudge them toward higher‑margin managed services when workloads or new AI projects demand it.
Cloud migration economics and the risk of being too legacy to love
Even as SQL Server keeps printing money, signs of strategic sidelining are hard to ignore. A former SQL Server boss has left, Azure Data leadership now stretches across Fabric and multiple open source services, and SQL Server 2025’s new vector search barely rated a mention at Build. That worries long‑time observers who fear the product could “languish” while attention flows to Postgres‑based HorizonDB and other cloud‑first offerings. The cloud market’s center of gravity has moved toward PostgreSQL APIs and distributed, managed services. For Microsoft, that means SQL Server must coexist with, not crowd out, Postgres and NoSQL options inside Azure and Fabric. The financial calculus is clear: preserve legacy database monetization for as long as possible, but invest marginal engineering effort where future AI‑driven workloads and cloud‑native data architectures are likely to land.
Databricks and the new wave of enterprise database competition
Enterprise database competition is no longer limited to Oracle versus SQL Server. Cloud‑era players like Databricks are building AI‑centric data platforms that aim to displace specialized engines such as ClickHouse, Splunk, and even classic data warehouse workloads running on SQL Server or Oracle. Databricks’ open‑source DBRX model, for example, required about USD 20 million (approx. RM92 billion) in cluster spending, underscoring how capital‑intensive modern AI infrastructure has become. While only part of that amount funded the final training run, the experiment shows how much vendors are willing to spend to build proprietary AI and analytics stacks that lock in customers higher up the value chain. These newcomers focus less on traditional relational features and more on unified analytics, vector search, and tight AI integration that can make legacy systems look slow to evolve by comparison.
Balancing cash cows with AI bets in enterprise data strategy
The SQL Server story shows how hard it is for incumbents to balance old and new. On‑prem SQL Server remains a cash cow, but the fastest growth is in cloud platforms and AI‑ready data services such as Fabric, distributed PostgreSQL, and managed lakehouse offerings that compete for the same budgets. Startups and cloud‑born vendors, including Databricks, intensify the pressure by targeting lucrative slices of logging, analytics, and transactional workloads once owned by ClickHouse, Splunk, Oracle, or Microsoft. For enterprises, this mix changes cloud migration economics. Rather than a single cut over to Azure SQL or an AI platform, they often keep legacy databases for core systems while shifting new AI and analytics projects to modern engines. Vendors are betting they can keep extracting legacy database monetization while steering that new spend into their own cloud and AI ecosystems.
