AWS S3 Tables Can Now Move Your Hive Tables to Iceberg Without a Custom Pipeline
AWS S3 Tables can now migrate Apache Hive tables to managed Apache Iceberg tables without custom pipelines, with incremental syncs keeping production writing until a short cutover.

AWS S3 Tables Now Migrates Hive to Iceberg for You
The AWS weekly roundup of November 9 included a quiet update that matters more than it looks: Amazon S3 Tables can now migrate Apache Hive tables to managed Apache Iceberg tables without custom pipelines. You pay only for the data copied, and incremental syncs keep production writing until a short cutover.
If you run a data lake, that sentence is worth pausing on. Migrating a table format has traditionally meant building and babysitting a pipeline, scheduling downtime, and hoping nothing breaks halfway through. AWS is now offering to do the tedious part for you, inside the service, with the production table staying writable until the final switch.
Why the Hive to Iceberg Move Matters
Apache Iceberg has become the default table format for modern data lakes, and Hive's older format is increasingly a legacy burden. Iceberg brings features data teams actually want: schema evolution that does not break queries, time travel to earlier versions of a table, and cleaner handling of concurrent writes. Hive tables, by contrast, are the thing you inherited and have been meaning to replace.
The migration path is the practical part. S3 Tables handles the copy, keeps the source and target in sync incrementally, and lets you cut over when you are ready rather than during a maintenance window you had to negotiate with everyone. Paying only for the data copied, rather than for a migration service or a standing pipeline, keeps the cost predictable.
This is not a flashy launch. There is no new model, no price cut, no keynote moment. It is the kind of unglamorous infrastructure improvement that data engineers notice and everyone else ignores, which is exactly why it is worth covering. The cloud gets better in increments like this, not only in big announcements.
The Hyperscalers Are Growing Fast, and Spending Faster
The S3 Tables update arrived in the same week analysts were digesting the hyperscalers' third-quarter results, and the numbers are extraordinary. AWS revenue grew 39.3% to $46 billion, its fastest growth in nearly five years, with a contracted backlog of $578 billion and an annualized run rate of $184 billion. Microsoft disclosed Azure revenue in dollars for the first time: $33.2 billion, up 48%. Google Cloud more than doubled to $30.7 billion, up 103%, and reported a backlog of $561 billion.
Combined, the three providers now carry roughly $1.8 trillion in contracted backlog, and their capital expenditure plans for 2026 approach $600 billion at the midpoints. Amazon expects about $220 billion in capex for 2026, mostly on AI infrastructure. Alphabet spent $51.4 billion on capex in the quarter, up from $24 billion a year earlier, and kept its 2026 guidance at $195 to $205 billion while saying spending will increase significantly in 2027.
The growth is real, but so is the cost. Free cash flow is being squeezed at all three companies, and Alphabet had its worst day on record after its results despite Google Cloud doubling. Investors are asking a reasonable question: if the backlog is $1.8 trillion, why is the cash not showing up yet? The answer is that building data centers and buying chips costs money before it earns money, and the payback period is measured in years.
What Else Changed in the Cloud This Week
Microsoft launched a new round of Azure promotions in the Cloud Solution Provider program on November 2. The headline offer is Azure Database Modernization: up to 30% off eligible pay-as-you-go database usage for organizations moving on-premises SQL Server workloads to Azure. If you have been postponing a SQL Server migration, that discount is worth a real look, though the fine print matters and the eligible usage is narrower than the headline suggests.
On November 9, OpenAI listed ChatGPT Enterprise on AWS Marketplace, and eligible purchases can count toward existing AWS spending commitments, depending on each customer's contract. The same day, Anthropic released Claude Fable 5.2 on AWS, Google Cloud, and Microsoft Azure, with list prices 20% lower than the previous version. AI models are now bought through your cloud bill, and that changes how teams budget for them.
The Bottom Line on S3 Tables and Iceberg
If you have Hive tables in S3, the new migration path is worth testing on a non-critical table this week. The cost is limited to the data copied, the production table stays writable, and the cutover is short. If it works as advertised, it removes one of the last excuses for staying on Hive.
The bigger picture is that the cloud market is in a strange moment: record growth, record backlog, record spending, and record investor anxiety all at once. The providers are building for a future they are confident about, and charging for it now. For data teams, the practical takeaway is simpler: the tools keep getting better in small, useful ways, and the migration that used to take a quarter can now take an afternoon. For more on the cloud landscape, read our AWS European Sovereign Cloud story, our ChatGPT Enterprise on AWS Marketplace coverage, and our look at Cloudflare's Deno acquisition.


