The platform engineering space has seen incredible growth over the last few years, and we’re now at a point where things are really starting to mature. In my review of 2024 from a few weeks ago, I focussed mainly on community efforts and only gave a quick outlook on what we are planning to do in 2025. Let’s go deeper into what I think 2025 will look like for platform engineering as a whole.
In 2024, the platform engineering community exploded in size. PlatformCon received over 800 talk submissions, our YouTube views doubled year over year and the Slack community now boasts 24,000 members. PlatformCon itself had over 100,000 views in the first few weeks. It's clear that platform engineering is becoming the "new DevOps," taking over the software engineering, DevOps, and cloud-native world.
This isn't just community hype; established analysts like Gartner have also recognized platform engineering, naming it a top strategic technology trend in both 2023 and 2024. Remember Gartner’s forecast that 80% of enterprises will have some sort of platform engineering initiative by 2026? They've now even created a hype cycle specific to platform engineering, showcasing the maturity of the space, specifically highlighting concepts like Internal Developer Platforms (IDPs). Interestingly, infrastructure platform engineering itself appeared on seven different hype cycles.
But what about the big questions for the industry and where platform engineering is developing? The key theme for 2025 is that the space is maturing and incorporating best practices and blueprints from adjacent verticals and industries. As you’re probably tired of hearing me repeat at this point, “It’s the move from hype to reality”.
Prediction 1: Backstage backlash
While portals were all the rage at KubeCon last year, there's a growing realization that they are not the be-all and end-all of platform engineering. Many teams are discovering that they've spent considerable time (12-18 months) installing Backstage, but have little to no adoption. This is not just within our community but is reflected in conversations with analysts and consultancies like Thoughtworks, who see the same trend across many enterprises.
Backstage still has a significant market share, potentially 80-90% market penetration, but this isn’t reflected in usage. The problem is that many organizations mistake the portal as the entire platform, but Backstage is not your platform. Portals are the front end of the platform, a UI that allows developers (and executives) to discover and access the platform’s underlying capabilities. Building a platform is like building an application; it has a front end and a back end. Simply focusing on the portal (the front end) doesn’t address the underlying challenges, nor does it enable the developer self-service that platform engineering should facilitate.
Young or new platform teams often try to get a quick win by putting a portal on top of their existing setup. However, if you start with the front end, you may end up shoehorning business logic into it, which goes against the principles of application architecture. The portal is meant to provide access to the underlying platform; it doesn't provide the underlying capabilities itself.
Therefore, in 2025, more enterprises will realize that they need to start with the back end, focusing on APIs and orchestration, and then add a front end later. Front ends can be interchangeable; the foundation is much harder to replace.
Key Takeaway: Don't confuse the portal with the platform. Focus on building a solid back end first and then consider the user interface.
Prediction 2: Developers will lose direct access to infrastructure
Developers have enjoyed the freedom of clicking around in cloud consoles or using CLIs to build infrastructure, but it often creates a nightmare for operations teams.
While developers enjoy playing “infrastructure Tetris,” this leads to unoptimized resources, increased costs, and maintenance and scalability issues. It also creates friction between developers and operations.
Platform engineering should function as a vending machine layer, where developers select infrastructure based on pre-defined templates and guardrails designed by the platform team. This approach makes everyone happy, not just the developers.
Top-performing platform teams understand that they must get the infrastructure and operations teams on board. The platform should not only shield developers from the complexity of infrastructure but also shield operations from the unconstrained creation of resources. It's about offering a “golden path” for developers, using guardrails to help them create infrastructure without creating operational nightmares.
We're swinging away from the "anarchy" of "shift everything left," which was really the essence of early DevOps. The reality is that other industries over the last 200 years have trended toward progressive specialization and standardization, and platform engineering is no exception. “Silos are fantastic”, as Kelsey Hightower pointed out at last PlatformCon, as long as you have an API or platform layer to facilitate communication. The industry is waking up to the need for a balance between developer freedom and operational control.
Key takeaway: Create a self-service vending machine model for infrastructure, where developers can select from curated options that are standardized and optimized for the enterprise.
Prediction 3: Platform engineering initiatives must focus on Pareto efficiency
Platform engineering initiatives should prioritize Pareto efficiency, which means that improvements for some stakeholders should not negatively impact others. Pareto efficiency, originating from economics, describes a state where it is impossible to improve one stakeholder's outcome without worsening another's. In the context of platform engineering, this principle means a platform's design should ensure that new tools or processes benefit some users (like developers) without diminishing the experience or productivity of others (such as operations or security teams). If a solution appears beneficial for one group but shifts responsibilities or burdens onto another, the overall efficiency can remain the same or even diminish.
Platform engineering is a multiplayer game, involving developers, infrastructure and operations, security, architects, and executives. It's important to get everyone on board, but it is also very easy to lose momentum.
Clear frameworks and blueprints help maintain momentum. A platform should aim to provide a 10x improvement for key users like developers, but if this results in operations teams working extra hours to maintain new resources, the net gain is los.. Similarly, a highly standardized platform may reduce the workload for operations but introduce friction for developers who need deeper visibility into their tooling [your conversation history]. In both cases, any localized improvement is counterbalanced by a corresponding drawback elsewhere. The key to avoiding this pitfall is to engage with users early, evaluate potential ripple effects, and optimize across the entire application lifecycle rather than isolating improvements in just one domain.
Similarly, a highly standardized “black box” platform may reduce the workload for operations but introduce friction for developers who need deeper visibility into their tooling. In both cases, any localized improvement is counterbalanced by a corresponding drawback elsewhere. The key to avoiding this pitfall is engaging with users early, evaluating potential ripple effects, and optimizing across the entire application lifecycle rather than isolating improvements in just one domain.
Key takeaway: Platform engineering should benefit all stakeholders, but what matters most is Pareto Efficiency, so the net outcome is positive.
Final thoughts
The platform engineering space is maturing rapidly, and organizations are starting to approach it on a whole other level. I'm extremely excited to see the changes and improvements, with people waking up to what works and what doesn't.
If you are not already part of the community, we would love for you to join us. We have 200,000 practitioners across various destinations, including Platform Weekly, our newsletter that reaches 100,000 people weekly, a Slack community with 24,000 members, 35 meetup groups worldwide, and the world’s largest platform engineering event, PlatformCon, which covers a full week of virtual content alongside and an in-person program with Live Days in London, New York and soon Paris with speakers like Kelsey Hightower, Gregor Hope, and Nikki Watt.
Not to mention all the incredible courses, trainings, and certifications to help provide the blueprints and best practices.
I’m looking forward to seeing what 2025 has in store and I'm excited to see you in the community!
Happy New Year,
Luca
.webp)
