2025: What’s Next in Platform Engineering – Predict 2025
The world of platform engineering is evolving rapidly, and staying ahead of the curve is essential for anyone looking to ensure they don’t get left behind. Join Luca Galante, a leading voice in platform engineering and the host of PlatformCon, as he unpacks his top predictions for the future of the field in 2025. Drawn from Luca’s extensive conversations with hundreds of engineering teams and his work in the world’s largest platform engineering community, this session will help you prepare for what’s coming, avoid common pitfalls, and realign your platform strategy for maximum impact in what may be the most important year for platform engineering.
Luca will dive into:
– The growing “Backstage backlash” to the once mighty Developer Portal
– How the pendulum of engineering history is swinging back toward standardized practices, what that means for you, and why “silos” might not be such an enemy after all.
– The Pareto principle of platform initiatives: Why “good enough” can be better than aiming for universal developer happiness.
There is no better way to future-proof your organization and prepare yourself for platform engineering in 2025.
Transcript
On. All right. Welcome everybody to my predictions for platform engineering in 2025.
I'm Luca from the platform engineering community. I actually see a lot of platform engineering initiatives, um, throughout the year. And so I've put together three predictions that are really rooted into trends and things that I've seen this year in 24, um, and that I think are very interestingly developing into 2025.
And so that's kind of what I, what I want to talk to today. Um, really just like cover this, the three things going into next year. Um, and because they are sort of, you know, rooted into 2024, I wanted to, before we get to the predictions, just quickly kind of take a look at 2024, what happened?
Um, you know, we released this data apart from engineering report, uh, just a month ago. Um, that has a lot of interesting data and we're gonna look at some of that data as well. Um, but in general, um, you know, we've seen the community really exploding in terms of size.
You know, this thing keeps compounding at its same rate platform calling. This year we had over 800 talk submissions. Um, we had, you know, the, the YouTube subscribers, uh, for the platform engineering YouTube are keep, you know, doubling or tripling year over year.
Um, the Slack now has 25,000 members in it. So there's a lot of stuff that's happening. Um, and the community and, and I think Platform Con, because it's been around, you know, every year since 22, is actually a really good also meter of that.
We had a hundred thousand, over a hundred thousand views just in the first couple of weeks, um, since of Platform Con. And, and that's, you know, forecast to keep growing next year at Platform Con 25 as well. Um, so again, this really speaks to the fact that I think platform engineering is really going mainstream at this point.
These numbers are really quite impressive. Um, and you know, we, we, we, we speak about this in the community, we speak about this in the platform engineering show with Alan on the podcast that we recently launched about, uh, you know, how platform engineering really is sort of the new DevOps and is kinda taking over the software engineering, DevOps industry, cloud native industry, and so on. Um, and I think, you know, this growth, this compounding is a, is a good proof of that.
But you don't have to just take my wording, kind of the community numbers for it. I think it's also interesting to see how, you know, very established analysts like Garner are also sort of, you know, talking about platform engineering in a very concrete way. Um, they named it a top strategic technology trend, both in 23 and this year in 24, they forecast that by 2026, 80% of enterprises will have some sort of platform engineer initiative either rolled out or rolling out.
Um, and I think what's really interesting to see is, you know, I've obviously was talking very active to a lot of the Gartner analysts all this years. Um, and I was very, um, happy when they finally put platform engineering on the hype cycle that the general software engineering hype cycle back in 22. You can see up here on the left, and you can see how, you know, platform engineering was kind, they, it was out on the hot cycle, but it was, it was still sort of like at the beginning of that hype curve that they, that they designed.
Um, in 23 last year, you can see the, the, the red arrow, uh, moves up. So it was kind of like peaking. And now what's interesting is this year they actually release an own hype cycle for platform engineering specifically.
And you can see stuff, for example, internal developer platforms, uh, which is one of kinda the key concepts that we push in the community. org, which is one of the, so like key, um, websites that, that people have been, um, looking at for the last really, like five plus years at this point. And you can see that on the, on the on on hype cycle.
Very interesting. Also, what's very interesting is that, um, a subcategory, effectively a platform engineer, which is infrastructure platform engineering, and we're gonna talk a little bit more about that later in the predictions, um, actually has been named seven times of seven different hype cycles, whether it was the hype cycle for S-R-E-I-N-O teams, uh, it, and so on, right? So very, very interesting, I think, um, to see just how this, the, the, the space keeps compounding, keeps maturing.
Um, and, and that's kind of one of the things that I, uh, that's really, I think that, that the common thread of these predictions that I want to go into, which is really the space is maturing, is waking up to a lot of best practice and blueprint that we can either, uh, sort of borrow from outer, you know, adjacent verticals and industries, um, or really finally understand, okay, hey, this is actually what best practice looks like for platform engineering. Um, and so without further ado, let's jump into the, uh, predictions. This is the first one, backstage backlash, uh, which sounds very, uh, very interesting.
Uh, but what it means, um, effectively is that, and you know, this was very obvious to me. Um, for example, at CubeCon this year, if you look back at Q Con last year in 23, um, uh, the North America one especially, so like, literally like, you know, basically, uh, you know, 12, 13 months, um, ago, uh, portals were really the big hype, right? Like everybody at Cube Con's kind like, oh, apart from engineering and internal developer portals, portals, portals, portals, um, I think, um, CubeCon this year, you could see the tune has really started to change.
Where, again, as this phase is maturing, a lot of people are realizing, Hey, actually, you know, I spent all this time in installing backstage, which can take up to like, you know, 12, 18 months in a lot of cases for large enterprises. And then I actually have no adoption. I have no usage, right?
Um, and, uh, and this is something that, you know, not only we're seeing, you know, in, in our community, but really talking to people like Garner talking to people like ThoughtWorks that implement a lot of these, uh, portals, um, you know, across many, many enterprise accounts. Um, we are, we're getting the same sort of message, right? And I mentioned earlier the, the sort of the data platform engineering report, you know, it's backstage still has by far the, the biggest market share.
I would even say this is actually our own data, and it's still very, very conservative from, you know, uh, other data points. When I talk to like, you know, this Garner and other partners of the community, actually, the figure is over 80, 90% in some cases in terms of backstage market penetration. But while there is a lot of, um, sort of penetration by the tool, um, that actually doesn't, is not reflected in terms of usage, right?
Um, and why is that? Well, if you look again at, um, sort of the, the, you know, this definition by Garner, uh, internal development portals are, you can really think of it as the, the sort of like the front end of your platform, right? It's something that developers but also executives, for example, use to discover and access the underlying internal developer platform capabilities, right?
So this is really, really important to understand that effectively, you know, building an a, building a platform is like building an application. You have a backend, you have a front end, and you know, ideally, um, you, you, you really need both, right? And I think the mistake that a lot of, uh, platform engineering teams have been making, and again, the thing that a lot of teams, a lot of organizations are not waking up to is the city of like, Hey, my portal is my platform.
It's not, your portal is just a UI basically interface to your platform. Um, but if you just do a portal, um, you're not actually gonna address the underlying sort of challenges. Um, you're not gonna enable developers sales service.
Ultimately, you're not gonna achieve what platform engineering is really designed to help you achieve as an organization, right? Um, and, and this is the thing is what a lot of teams, a lot of organizations have been realizing now where it's like, Hey, I spent all this time building, you know, installing backstage thinking that that's my sort of like NBO sort of solution to platform engineering. Um, well, actually now I have, you know, very low adoption.
I don't, um, I don't, I don't, I just don't see this, um, this thing really making a big difference for me, right? Um, and again, um, this is something that, um, I think a lot of people are now realizing, but, um, you know, experts, uh, so that, you know, being there done that a few years ago had already been calling this out for, for a while, right? org, uh, which is our, our community side, um, I think like four or five years ago.
Uh, this is from Aaron Erickson who, um, uh, uh, guess wrote it, uh, for the, for the blog, um, and where he drew, and Erickson, by the way, is the guy that, uh, built the, uh, the sort of the main enterprise grade platform at Salesforce. So he definitely knows one or two things about adoption, um, Ander platform engineering, and he draws this very helpful nudge of thing where he basically compares, um, you know, platform engineering and building an, an internal developer platform or IDP for short to building a house, right? Where it's like, Hey, you know, if you were to build a house, you would obviously start from definition and then, you know, add your walls, add a roof, add the windows, add the doors at the end.
You wouldn't start from a door, a door first, and then kinda like build everything around it, right? They'll be silly. And that's a little bit, actually, you know what a lot of people's, a lot of people have been doing, right?
A lot of platform teams, you know, maybe new platform teams, young platform teams that had a recent mandate to do platform engineering. Were trying to get and understand why this happens, right? They were trying to get usually a, a quick win under their belt, right?
And so it's like, Hey, you know, let's just put a platform on top of our existing setup, and off you go, we're doing platform engineering. Um, and the problem with that is that, again, if you think about building a platform as effectively building any other application, you have a backend, you front end, you know, you, we've won a few things in the last couple of decades of building these applications, of building, for example, microservice applications, right? So you a, you have a single responsibility principle, um, which is really the idea that like, one of these modules should only do one thing, um, and that you don't wanna misuse a module to do something that it's not supposed to, right?
And so, an example, a perfect example is this, is, well, if you start from the front end, right? You then maybe, you know, have a pretty dashboard that you can show executives after like a few weeks. Um, but actually that actually doesn't have a lot of substance underneath it, right?
Um, and then what happens is when you know, maybe they're even impressed, then you, then you need to actually build this, you need to start building the real platform. And then what people do, what platform teams do is they start shoehorning business logic into the front end of the platform, which again, if there's anything that we've learned from application architectures, you really want your business logic in the backend because you don't wanna expose business logic. You don't wanna expose, you know, potential vulnerabilities.
But more importantly, the certainly the front, uh, kind of modules are not built to handle that type of logic, right? Like the reality is that, you know, backstage, for example, or native portal is really just like a UI on top of that, that, you know, visualizes things that gives you access, again, to the underlying platform capabilities. It's, it doesn't provide you with the actual capabilities underneath it, right?
It doesn't let you orchestrate, um, sort of the, uh, you know, logic underneath it doesn't let you orchestrate your workflows, your infrastructure and so on, which is ultimately really where the pain from a developer experience is, and from an infrastructure operations team as well, and where a lot of the value is that can be unlocked with your platform engineering initiative. So that's, I think, really, really important. And so, again, I I, I've seen a lot of an increasing amount of people, especially in the second half of 2024, really come, you know, come to realize like, Hey, you know, portals are, are okay, they're cool, but I really need some sort of like API orchestrating stuff underneath them.
That's really where a lot of the value lies. And, and so I predict that, that, you know, more and more enterprises will figure that out going into 2025. And this will actually become a very established, you know, hopefully best practice of like, Hey, you need to start from the backend and then add your front in later.
Um, because your front end can just be interchangeable, right? You can change your doors, you can repaint your doors, um, you know, it's much harder to kind of like, uh, rip your, your, your foundation off and, and rebuild it every time, right? So, so I think that's prediction number one.
Number two, um, developers will lose direct access to infrastructure. What do I mean by that? I think a lot of teams, um, and here again, you know, not to pick too much on backstage, but this can really be done through any interface, can also be just a CLI or, uh, you know, a a code-based interface and a P based interface to your platform.
It doesn't matter. The point is there are, you know, cloud consoles, there are, um, sort of, um, you know, but mostly, let's be honest, mostly it's either through a portal, through product console that developers can play this fun, you know, what I call infrastructure tes of just like, you know, choosing different sort of like pieces of infrastructure. Uh, in this case it's an EKS class, right?
But you can have posters, you can have DNS, you can have all sorts of, um, you know, fun, uh, stuff that you need because you wanna, you know, test the code, change, deploy something and so on, and keep adding them into, um, you know, into the, the void, right? Into, into the, the sort of like the, the enterprise ether of, um, uh, of your infrastructural setup, right? And the problem with that is like, while developers are very happy, you know, they just like click around and they create all these things and they're, you know, everything is provided, you know, this creates, um, a nightmare for operations colleagues to maintain, right?
Um, it's really, really hard. Um, a a and this doesn't really create, you know, costs in terms of, um, you know, obviously you end up, and I think we're all very familiar with this, um, story of a lot of sort of, um, you know, idle resources that are being created. They're not really being used.
They're not really being optimized, but more importantly, I think it creates a lot of overhead, uh, maintenance, um, and scalability issues on infrastructure and operation side of things, right? And, and it creates a lot of friction between developers, operations, and so on, right? Um, and so I think a lot of teams are now waking up to that and to understand that, like, hey, your platform can really be thought of as a, if I, like, if it's a vending machine layer, right?
For, uh, your developers where developers can, you know, select the infrastructure, but following a very precise predefined set of, uh, templates and, uh, sort of like guardrails that the platform engineering team designs as a product for them, right? So it's very, very important, I think, to understand that like, hey, platform engineering is this multiplayer game where yes, developers is your end user, but you need to make everyone happy, right? You need to, to make your developer happy and, and, and let themselves serve things, but not at the cost of sort of like infrastructure and operations teams sort of mental health, right?
Where hey, um, you know, and, and, and also I think it's very, very important, this is one thing that I'm seeing top performing organizations, top performing platform engineering teams realize is like, Hey, I cannot only over-optimize on the developer. I need to get the infrastructure and operations team on board with the platform too. And how do you do that?
Well is by selling to them effectively, right? This idea, they're like, Hey, you're, you know, you can look at this platform engineer initiative as a kind of like a, a layer that not only shields developers from the complexity of the underly underlying infrastructure, but also the shields you from, you know, this, this basically infrastructure te that a lot of developers like to play, uh, a little bit too often, right? Um, and, and so, um, I think it's, I I'm seeing this, um, you know, good platform teams realize, Hey, this is how I get everyone on board.
And, and really, you know, having, using this mental model of a golden path and of, of this kind of this guardrails for developers, um, the really, really essential, um, and I think if you zoom out, um, you really realize that like, you know, um, there is this, this, this, this word that is almost like a bad word in our industry, and it shouldn't be, which is silos, right? Um, I think, you know, we had a very interesting conversation, uh, Casper, um, from, uh, from human tech. Casper from Gumma did, had a really interesting conversation with Kelsey Hightower, apart from Khi, um, where basically Kelsey at some point said, Hey, you know, the reality is that silos are a good thing, right?
And you should have seen the q and a later because a lot of people got really mad. Um, and I think is is for not for no good reason, right? Um, the reality is that if you look back 20, 30 years ago, we had, um, we were, we were too silent, right?
You had this, you know, application developer says admin throwing, um, kind of like code over the fence, and like, really a way too silent. But I think, um, and this is, uh, something again that, that we were talking about the other day with Alan on the podcast. Uh, you know, Devox was really this like swing of the pendulum all the way on the other side where it's almost this kind of like anarchy of like, Hey, let's shift everything left in the developer now.
So do we align all the incentives? But that is kind of missing the point, which is really what other, every other industry in the, you know, in the last 200 years plus since the industrial revolution have, has, has been trending towards, which is, um, progressive specialization com, com, com commoditization, right? And really industrialization.
And I think, um, you know, the, the pendulum is now swinging back into the middle, which is really the sweet spot where we wanna be. We don't wanna have, you know, silos that really never talk to, to each other and, and, you know, inefficiently don't communicate. But we also, I don't wanna, um, you know, kind of remove this idea of like, people need to specialize, right?
Um, and, and so I, I think that silos are coming back in a good way with an a PIA platform layer in between them that lets you communicate the right way. Um, and, um, and I think this is really what a lot of people, a lot of engineer organizations around the world are waking up to. Um, and I think it's gonna make a huge difference going into 2025.
So that's prediction number two. Last but not least, apart from engineering initiatives need to be Pareto optimal. Um, and so obviously I think a lot of people are familiar with this Pareto principle d or, or kind of like 80 20 rule, right?
Which is really you wanna optimize for the, uh, majority of use cases, right? But I think, um, the important thing to, to, to realize here is, is I was saying, um, powerful engineering is, you know, very kinda like complex multi multiplayer game, um, with where you have a lot of different stakeholders, right? You have like application developers, you have infrastructure and operations team, we talked about them, but you also have security teams, you have architects, you have executives, right?
And the point is that you kind of need to get everybody on board with this thing because with your platform initiative, I mean, because, you know, um, it's gonna affect every one of them. Um, the platform initiative kind of touches all the different parts of your engineering organizations, so you need to get everybody on board. And the problem is that it's very easy to kind of lose momentum, right?
Because you're talking to person A, B, C, by the per, by the time you get to person Z, um, you know, it's been six months person a completely forgot about you, and you kind of lost momentum. And this is where, you know, blueprints and, and you know, frameworks that we talk a lot about in the community, like mini revival platform, MVP framework, um, and others can really be helpful in making sure that you, you know, not lose the momentum and you keep it. But I think it's also very, very important that you realize, hey, while the platform should, uh, you know, really make, have like a 10 x uh, improvement, provide a 10 x improvement to, for example, you know, usually developers or the main users, it is very, very important that it doesn't, um, sort of detract from anyone's, right?
So I think the point here is really is like, Hey, you wanna make some people really happy, but you definitely don't wanna make anyone unhappy, right? Um, and I think, um, you know, this is an, an essential part of driving, uh, apart from engineering industries forward. Um, again, you know, you can think of, um, I think this, this, this sort of a multiplayer game, uh, analogy of you have developers, you have infrastructure and operations team, we looked at them, you know, you kind of want to drive automation and standardization as kind of the key value drivers for both of those stakeholder groups.
And ultimately that affects, you know, uh, basically it lets you move faster without breaking things, right? That's, uh, one way I think about it, which is really, you know, you cut your time to market, your innovation cycle goes up, um, uh, but you do that in a way where you're still like reducing costs, you're being more compliant, you're being more secure and so on, right? Um, and, and, and so it's very, very important that you have clear ways of kind of, you know, get everyone on board with you on this journey.
Um, and a big part of that is, yes, following clear frameworks and blueprints, but also really, um, you know, making sure that, um, everyone is, has at least a sort of like net neutral to net positive, um, kind of outcome from the platform engine initiative. And obviously in some cases, like developers, you're gonna have a te you, you need to provide, so like AAN X, right? But it's really making it, it's really important, I think, to make sure that, um, nobody is, um, sort of, you know, against your platform initiative for, for no reason, right?
Because the, the reality is a lot of the platform engineer initiatives kind of like end up dying and losing steam, um, because they don't get the, the right, the right support internally. And it's really a tragedy of the commons. That's really what happens, right?
Where, you know, ultimately platform engineering initiatives are a net positive from an organizational perspective always. But it might be that from the selfish perspective on individual stakeholder or stakeholder group, you know, they're kind of like, well, this is actually a bit annoying, it doesn't really do too much for me and so on, and they're gonna block that, right? And so it's very, very important that we, you know, have a, a, a, a good strategic approach and tactical approach to get everybody on board.
And I think following, making, you know, following the parade of principle A 20 rule, I think is a very good, um, sort of way of thinking about that. Um, and by the way, if you are, um, and so that was my third, um, and, and kind of last prediction. I think a lot of people are also waking up to this, again, as the space is maturing.
Um, I think, uh, organizations are starting to do platform engineering on a whole other level, which is super exciting for me to see. Uh, and, and, um, something that we talk a lot about in the community. Um, if you're not familiar with the platform engineering community, it's really the largest out there.
We have, give or take probably 200,000 practitioners, um, across the different community destinations. We have Pop from Weekly, which is a newsletter that goes out to over a hundred thousand people every week. We have the Slack, which is really the beating heart of the community.
You can go in there and like jump on conversations right away. There's already 25,000 people in there. We have 35, uh, meetup groups all over the world that meet every week.
Um, you know, there's Platform Con that I talked about, of course. Um, and we have a lot of, um, sort of courses and trainings and certifications that we've now rolled out where we do speak about a lot of these frameworks that you see here, where we provide you with the blueprints, um, and sort of like the best practice that we see from top performing platform engineering teams, top performing engineering organizations, um, and kind of making sure that they trickle down to mainstream, uh, as efficiently as possible. Um, and so come join the community, come join the courses, come join the events as well.
We're gonna host, um, not one, but two live days this year. One in London on the 25th of June, um, and one in York on the 26th of June. This slide is wrong.
Um, but this is where, um, I'm really excited because actually the trainings that I was mentioning earlier, we're gonna bring them to in in Live, in person, um, uh, in, in, in both locations. We're gonna have great speakers like Kelsey that you just looked at earlier on the slide at Gregor Hope and Nikki Watt a lot more. And so, um, I think I'm personally very, very excited.
I see every time, you know, I go to Platform Con, I go to CubeCon. I just see this crazy, you know, compounding every six to 12 months of the platform engineering space. The first CubeCon I've been to, um, like four or five years ago, you know, I had to explain to everybody, okay, what is platform engineering?
Who's a platform engineer? Um, then I think a couple of years ago, uh, it really exploded. Everybody was talking about it.
Um, and now the space is really maturing. They're, you know, it's not just about the hype any longer, but they're really figuring out, okay, what can it do for me? And how is real platform engineering at Enterprise Scale being done?
How do we follow best practice? And I see, you know, more and more that happening in the community. I see people really waking up to what works and what doesn't.
Uh, and then sort of like come and talk in the community at the, I have about, um, you know, the best practice that they're that're figuring out. So I'm very, very positive, um, of, of all the changes, all the improvements that I keep seeing in the, in the community, in the space. And I'm very excited about, um, what 2025 has in store for us.
So, um, looking forward to next year. Um, and those are my three predictions. Um, and with that, thank you.
Happy New Year. See you on the other side.



