The applications that matter most to your business rarely stand still. They earn customer loyalty, open new revenue opportunities, and help people make better decisions. When they succeed, the demands on them grow: more customers, more transactions, more data, and new expectations for AI-powered experiences built on that data.
Too often, technology leaders are told that meeting those demands means starting over. Getting cloud-native scale, better economics, or AI capabilities seems to require rewriting proven applications, retraining experienced teams, or moving to an unfamiliar database. For organizations whose most important applications run on SQL, that trade-off puts years of business logic, operational knowledge, and trusted data at risk.
Our view is simple: modernizing shouldn’t mean abandoning what works. Azure SQL Database Hyperscale brings cloud-native scale, competitive economics, and AI to the SQL applications you already run, and to the ones you build next.
Built on SQL, engineered for cloud scale
Hyperscale runs the SQL engine your teams already know, rebuilt on a distributed architecture in the cloud. It doesn’t require you to procure or maintain SQL software; it’s a cloud-native service, provisioned in minutes and updated continuously.
You can preserve valuable SQL investments while removing constraints on an existing application or start a new application with room to grow. Its architecture separates compute from storage and allows read workloads to scale independently. That flexibility also creates economic value: you can align resources more closely with demand rather than treating every increase in data, transactions, or reads as the same infrastructure problem.
AI makes that flexibility even more important. Applications increasingly need to connect a customer’s intent with current products, orders, inventory, or account information. Hyperscale brings relational data and vector search together, so teams can build intelligence around trusted operational context. Modernization, scale, AI, and cost discipline are not separate conversations. They’re connected decisions about how much of your team’s energy can go toward the next business opportunity.
Keep what works, make room for what comes next
Modernization should begin with the value an application already delivers. Years of business logic, operational knowledge, and SQL expertise are assets, not obstacles. Hyperscale shares a common SQL Server codebase and supports familiar T-SQL, stored procedures, tools, and application frameworks. That continuity can reduce unnecessary rewriting and reskilling.
But databases rarely move alone. They sit inside applications with their own infrastructure, dependencies, and release cycles. Azure Catalyst is becoming the one-stop shop for SQL Server migrations at enterprise scale. It brings application-aware discovery, assessment, sequencing, and migration execution together in one unified experience. Powered by application designs, it helps customers plan and carry out migrations across infrastructure, data, and application tiers in coordinated waves, with greater speed and confidence.
For developers, the same thinking extends into the tools they already use. Modernize CLI brings an agentic experience that assesses an application and its database dependencies as a single workload. It pairs application-code modernization opportunities with SQL Server migration readiness, compatibility, and target recommendations in one consolidated report, giving developers a clear path to plan and carry out their modernization.
For a new application, the opportunity is to avoid making tomorrow’s scale a prerequisite for getting started today. Teams can begin small and adjust resources as adoption grows. Hyperscale’s independently sized read replicas can support read-heavy experiences without placing all the work on the primary transactional compute. The business gains more options for adding capabilities as demand changes, rather than committing early to a single forecast of success.
SAP Commerce Cloud illustrates why that matters for an established application. A fixed, bundled resource model constrained scaling during traffic spikes. SAP migrated 6,000 production databases to Hyperscale in one year, with extensive testing, before Black Friday 2025.
$17.8 billion worth of gross merchandise went through the system with 100% uptime. The peak number of orders was 648,000 orders per hour, which is almost 11,000 orders per minute.
Conner Helton, Senior Manager of Product Management for Commerce Cloud, SAP
The reported peak-shopping period shows what that capacity enabled. This is modernization that helps a successful commerce platform keep growing, not a departure from the SQL investments behind it.
That same SQL foundation also creates a path to what comes next: adding AI experiences grounded in the operational data your applications already trust.
Put intelligence where the business context is
A useful AI experience needs more than a plausible answer. It needs context that reflects the business now. SQL Database Hyperscale brings relational, JSON, graph, and vector capabilities together, enabling developers to combine semantic similarity with full-text search, relational filters, and joins. Teams can build more relevant AI experiences using the operational data their applications already depend on, with familiar SQL security, permissions, and governance. With the Microsoft Ignite 2026 announcement of the new Azure SQL Database container, now in public preview, developers can start locally on the same engine that runs in the cloud, provisioning and managing it directly from Visual Studio Code.
Intelligence also depends on reaching all the data that holds that context, even when it lives in more than one database. Cross-database querying for Hyperscale, now in limited preview, lets developers use familiar three-part names to query across databases in the same Hyperscale elastic pool from a single connection, bringing related data together without first consolidating it. Teams can keep databases separate by application, business domain, or tenant, while each database continues to check its own permissions.
Consider a commerce assistant finding a replacement for an unavailable product. Vector search in Hyperscale identifies relevant alternatives from the product catalog. Then, in the same query, cross-database querying joins those results to live stock levels in a separate inventory database, so relational filters narrow the list to products that are in stock and meet the shopper’s needs. The same approach supports recommendations, retrieval-augmented generation, and AI agents, all grounded in operational context, without requiring a separate vector store or a pipeline to copy data between databases.
For an existing application, these capabilities open a practical path to adding intelligence without first replacing its data foundation. For a new application, they let teams design transactions and AI-enabled experiences together from the start. Hyperscale provides the cloud-native architecture to support those applications as data volumes, read demand, and business importance grow. The result is a connected path from the SQL investments you trust today to the experiences you want to build next.
Ready to start? The Azure SQL Developer Hub gives developers a fast start, with ready-to-use prompts for scenarios like searching your content by meaning with vector search, and agent skills that teach GitHub Copilot and other AI coding assistants how Azure SQL works. Start free in the cloud, build the retrieval layer for a RAG app on your own data, and keep the same code as you scale.
Treat economics and resilience as design choices
Growth is more valuable when the operating model can grow with it. Hyperscale offers provisioned compute for planned capacity and serverless compute with autoscaling and usage-based billing for changing demand. Separately, elastic pools let databases with uneven demand share compute, which can help SaaS providers and teams managing database estates use resources more efficiently. These are choices to evaluate against workload behavior, not a promise that one configuration always costs less.
Enterprise SQL trust and competitive cloud economics need not be a trade-off. A recent Microsoft-commissioned study with Principled Technologies found Azure SQL Database Hyperscale delivered up to 52% higher performance and 51% better price-performance than Amazon Aurora PostgreSQL.1 As modern applications add AI-powered search, recommendations, and agents, they place new demands on the data foundation. For both existing and new applications, the goal is to support that intelligence alongside core transactions while keeping performance and cost in balance.
As an application becomes more critical, recovery and protection belong in the same design discussion. Hyperscale provides snapshot-based backups and fast restores, together with configurable high availability, zone redundancy, and geographic options, to help teams plan for disruption. Identity, SQL permissions, row-level controls, encryption, and network protections help protect access to business data. The right choices depend on the application’s recovery objectives, risk profile, and configuration.
We’ve moved banks from fear of downtime to a scalable, resilient, and highly available platform. With Azure SQL Database Hyperscale, the focus has shifted from outages to digital adoption, innovation, and proactive engagement. Availability and resilience have become part of banks’ value propositions, not just an operational requirement.
Nanda Badrappan, Deputy Chief Technology Officer, Temenos
A foundation that keeps moving forward
Your applications will keep evolving. Your data foundation should, too. Our recent announcements at Microsoft Ignite 2026 reflect our continued investment in Azure SQL Database, helping you build anywhere, deploy with confidence, and operate with AI.
As you look at the applications that will define your next advantage, consider where today’s constraints are limiting tomorrow’s possibilities. Azure SQL Database Hyperscale gives you room to modernize what matters and build what comes next.
SQL’s history is a foundation to build on, not a boundary. The technology your teams trust today can carry your applications and your ambitions forward.
See Hyperscale in action
Watch Bob Ward on Microsoft Mechanics for a closer look at Azure SQL Database Hyperscale.
[1] Price-performance claims based on data from a study commissioned by Microsoft and conducted by Principled Technologies in September 2026. The study compared performance and price performance between a 16 vCore and 32 vCore Azure SQL Database using premium-series hardware on the Hyperscale service tier and the db.r7i.4xlarge and db.r7i.8xlarge offerings for Amazon Aurora PostgreSQL I/O-Optimized (“Amazon Aurora”). Benchmark data is taken from a Principled Technologies report which used the HammerDB TPROC-C benchmark. The TPROC-C workload is derived from the TPC-C Benchmark and results were obtained with the HammerDB TPROC-C workload. The HammerDB TPROC-C workload is derived from the TPC-C benchmark and is not comparable to published TPC-C Benchmark results, as this implementation does not comply with all requirements of the TPC Benchmark. Price-performance is calculated by Principled Technologies as the cost of running the cloud platform continuously divided by new orders per minute throughput, based upon the standard. Prices are based on publicly available US pricing in East US 2 for Azure SQL Database and US East for Amazon Aurora as of September 2026. Performance and price-performance results are based upon the configurations detailed in the Principled Technologies report. Actual results and prices may vary based on configuration and region.