I asked T-SQL Tuesday #100 participants to take a speculative view on the world 100 months from now. 32 people weighed in—in some form or another—and produced a set of extremely interesting and diverse posts. Despite their far-ranging nature, the various posts converged on similar themes in many cases:
- What is the DBA career path for the next eight years?
- What will be the longer-term impact of the cloud and other infrastructure advancements?
- Where will Machine Learning and Artificial Intelligence technologies take us?
- Things will change, but how much will stay the same?
I thoroughly enjoyed reading each and every one of the posts, and I’d like to extend a sincere thank you to everyone who participated. And without further ado, I bring you the roundup. I’ve grouped the posts based on the general themes I noted above.
Your Future Career
Unless you’re planning to retire later this year or that lottery ticket finally hit, it’s important to keep yourself in the know about what to expect next from your career. Ours is an ever-changing field and T-SQL Tuesday #100 participants expect a fairly turbulent eight years. Don’t worry, though; it will be a great ride for those who manage to hold on.Warwick Rudd focused on a future he believes will result in a large-scale DBA skills shortage—but thanks to improved automation, only a shortage of advanced DBA skills. Warwick mentions that “by the time T-SQL Tuesday 200 arrives, the baseline for the DBA is going to require the knowledge and skill sets of someone in the position now with quite a bit of knowledge. Nearly all of those base level tasks that allow junior positions to be filled will be being completed by automation and as such the requirement for a more in-depth knowledge of the product stack will be required."
Brent Ozar’s post could fit into multiple categories—it’s full of various predictions—but my favorite of his insights had to do with career path. Brent sees “data safety” as a big standalone future business (one that, in my mind, can’t come soon enough). He sees DBAs as becoming “reliability engineers,” which is an interesting shift in mindset. And on the non-career front, he sees Microsoft allowing other vendors to sell SQL Server serverless offerings. This last one I think is absolutely insane but also incredibly intriguing. We shall see.
Glenda Gable took a science fiction approach. Her concern about the current state of office affairs is that sometimes people’s words are misconstrued. Communication is tough. She sees a future in which an AI device will help you understand people’s actual intentions, so as to avoid those strange and sometimes career-limiting interactions.
Hugo Kornelis makes a number of interesting predictions across numerous areas. Among other things he predicts over-application of machine learning, followed by an extreme pullback—something I agree is bound to happen. But his most interesting prediction is the future of “performance tuning” as a career. Imagine that your life and daily plans were subject to an optimizer. Imagine if that optimizer made and error and, e.g., sent you out to buy ice cream, leaving it sitting in the hot car for three hours while you run other errands. You’d need a Life Plan tuner. And that’s the career Hugo has in mind for himself. I’m intrigued!
Rob Farley is the Old Faithful of T-SQL Tuesday, having never yet missed a single month. In his T-SQL Tuesday #100 post he goes a bit introspective, asking what’s really important in your career. He comes to the conclusion that machines are going to greatly outpace humans in solving the tech problems—but even the best metal mind can’t tackle business issues. So that’s where we should focus if we want any chance of future career success.
Eugene Meidinger notices that developer ramp-up time is ever increasing; the more complex our technology solutions become, the longer it takes to learn how to use them. Meanwhile, developer job duration seems to continually drop. What happens when the two meet and what, especially, happens when developers start leaving jobs before they’re even properly trained up?
Hamish Watson is an incredibly optimistic sort of person. He sees a bright future in which "all Data Professionals will have an appreciation and understanding of how to automate the delivery of value to the end user." He also predicts that Microsoft will greatly improve SSDT, and that companies will actually want to invest in data cleanup initiatives. All of which sounds great to me.
My own post was, I will admit, inspired by a debate had over Twitter. But this is a topic I’ve been thinking about a lot. I’m not the first to say it, I certainly won’t be the last, and I’m probably not wholly correct. But I’m calling it and in 2026 the DBA is dead.
Cloud and Infrastructure Changes
The public cloud is really nothing new, but its rapid adoption represents a sea change in the way we see infrastructure and overall application architecture. Change is not going to stop or even slow down in the coming years; quite the opposite. So where will we be in 2026?
Shane O'Neill brings us a few journal entries from the world of T-SQL Tuesday #200. He predicts few remaining on-premise database instances and complete virtualization of just about everything—including the PASS Summit and a virtual round of karaoke. I hope the drinks are non-virtualized, at least?
Deborah Melkin talks about future trends in data volumes, data analysis, and data protection. She brings up the question of “where does the data live?” A decision path that was, only recently, pretty much a single node, has blossomed into a small tree. "Are we storing the databases locally or in the cloud? If you’re hosting locally, are hosting the databases directly on a server or are you hosting VMs or using containers? And which operating system do you want, Windows or Linux?" What sort of structure will this tree have in eight more years?
Steve Jones will, in 2026, be 60—and officially the Old Man of SQL Server (version 23). He predicts more cloud, broader processor and platform support, and a revamped and actually working version of replication. He also predicts that SQL Server will support multiple first-class query APIs for different languages and paradigms. Which seems to me a lot more probable than any new replication fixes!
Robert Davis sees himself as semi-retired by 2026. I wish he’d informed us of how he plans to accomplish that one; I don’t see any early exit in most of our futures. But I digress. He predicts a new feature, Scalability Groups, which will be an enhanced Availability Group concept that will effectively support writeable secondaries for full read and write scale out. Cool idea and totally foreseeable. Plus, merge replication will be officially deprecated. We can only hope.
Kendra Little’s future vision centers around a new core infrastructural concept, the BAGI. Big Awesome Giant Infrastructure. These are huge memory-based compute clusters. And because everything is in memory, you pay for that memory—by the gram of weight that your data consumes on the chips.
ML, AI, and Automation of Everything
Machine Learning is changing the world. But is it doing so for the better? With all of that power comes a lot of responsibility. And what will it bring to the table when leveraged within the database frameworks of today?
Bert Wagner imagines next-level, ML-driven automatic plan tuning capabilities. This future only makes sense. Among other predictions, Bert sees the "ability to choose an appropriate execution plan based on historical time-of-day server usage."
Arthur Daniels sees a similar future, but perhaps one that’s not quite as positive. What do you do when the AI-based query store starts auto-forcing the wrong plans? You’re forced to dig in to the open-sourced query optimizer and figure out which special trace flag is needed to shut the thing down.
Doug Purnell’s future vision is a bit more positive. He foresees PowerBI-based AI-generated notifications for "transactions, indexing, backups, DR/HA, and security tuning." All of which, coupled with PowerShell, will allow for seamless and totally online production migrations. Sounds pretty cool!
Eric Blinn wonders who will actually be running that PowerShell code. Will it be you, or your virtual DBA, Victor? It’s unfortunate if you don’t want to work with Victor; you’ve been financially forced to do so by Microsoft. Victor solves problems faster and better than you ever could. Can this possibly end well?
David Fowler takes us even further along the same path—and back, to the heady days the ‘80s. Skynet has become self-aware. And SQL Server is hosted therein. That terminator looks awfully familiar; does anyone know where I can find John Connor?
Randolph West’s 2026 world is one of drones everywhere, SQL Server running in Raspberry Pi containers, and special glasses that allow you to blink to accept calls. The elevator in your office building will use a predictive model to figure out which floor to drop you on without your having to push a button. But what happens when the batteries die on your fancy phone-enabled glasses?
Ginger Grant is concerned about the accuracy of Machine Learning. She predicts more automation, more cloud, more managed services…but an overall contraction when ML models start to get too many things wrong. She also issues a call to action: If you want to get involved in this world, learn Python. (And possibly don't bother with R at all.)
Riley Major provides a general description of ML and discusses his projections of SQL Server including ML and AI as first-class citizens, right there in the SELECT statement.
The More Things Change…
IT is trend-driven and amazingly cyclical. Eight years is a long time, but is it really that long a time? Maybe things won’t be so different after all.
Koen Verbeek considers BI in 2026. The key here is evolution, not revolution. He foresees ML-driven ETL and the need for ever-greater data protection. But decent data models will still be needed, self-service will still be a pipe dream, and at the end of the day there will still be data warehouses. Koen notes that "it’s totally cool what your machine learning model can do and how it will predict everything our customers will need, but the CFO [will still need] that financial report." And business users will still want everything in Excel. Undoubtedly.
Matthew McGiffen notices that a lot of companies are always several versions behind whatever is current. So his prediction for 2026? “DBAs will finally have the go-ahead from their companies to upgrade to the versions of SQL that are current now and start using this stuff!"
Jeff Mlakar makes various predictions about how little things will change. Security will still be awful, semicolons still won’t be required in T-SQL, and all of the currently deprecated features will still be there and still marked for deprecation. Sounds about right!
James McGillivray believes that data professionals will still be in high demand in 2026. None the less, they won’t be able to fix much. Most organizations still won’t focus on data quality, and master data will still not be in great shape. We’ll have a lot more tools at our disposal, but the core aspects of the job really won’t change.
Kenneth Fisher’s 2026 involves SQL Server releases once every two weeks. This is necessary, apparently, to keep up with the speed of business. Gone are DevOps—replaced, per Kenneth, by BizOps! What about our beloved Profiler? In 2026 it’s still right there when you need it.
Nate Johnson sees relational databases as still alive, albeit mostly in the cloud. Data quality will still be a huge issue and just about everything will be columnstore-based. Oh, and sp_whoisactive will be built in to SQL Server. Fingers crossed, Nate!
Kevin Hill reveals a nice surprise about auto shrink. Finally.
Other Topics and Creativity
The future is a wide and varied place, and naturally not everything fit neatly into the four major buckets. In reality, it’s almost certain that nothing will.
Kennie Pontoppidan imagines massive change in the next eight years. He sees an evolution of databases well beyond what was imagined back in the System R days and foresees quantum databases, built-in data classification, multi-language support, and most importantly direct data distribution lines that involve virtually all devices. Kennie believes that "the concept of [the] database will be diluted; we will be talking [instead] about data swarms."
Amy Herold, on the other hand, predicts a future of strange and horrible decisions made by Microsoft. The company will, in Amy’s mind, replace Entity Framework with NHibernate, get rid of Agent in favor of ActiveBatch, and force us all to write our queries in some as-yet-unimagined JSON-based format. Shudder.
Todd Kleinhans has been reading up on DNA-based storage. It’s super-high density, but read speeds are abysmal. No worries, DBA friends! For in 2026 we’ll see the advent of the non-clustered DNA index…
Taiob Ali is concerned about the future implications of “digital dementia,” a condition that involves "deterioration of cognitive abilities due to overuse of technology." He suggests reducing screen time, making an effort to be social (IRL!), and actually using your brain on occasion.
Bob Pusateri was that kid in elementary school who refused to color inside the lines. And maybe, on rare occasion, ate a glue stick. He decided to forego the future altogether and instead wrote about last month’s topic. He certainly does have a unique hobby!
That’s that for T-SQL Tuesday #100. Thanks for reading and thanks again to the participants for the great posts and the insight. It was a great pleasure to host.
If you would like to host a future T-SQL Tuesday, please contact me via the form on this site. I’d love to get you onto the calendar. Let’s help T-SQL Tuesday actually make it another 100 months and see how many of these predictions come true.