By Mark Sailes, Technical Presales Consultant at Leighton
AWS re:Invent 2025 was defined by scale, speed, and an ever-increasing number of options.
With a staggering 530 total product releases – 414 announced before the event and another 116 unveiled during the conference – it was impossible not to feel the sheer volume of opportunity available within AWS. Absorbing, understanding, and ultimately taking advantage of these innovations will be a major challenge, but also a defining one for organisations over the coming years.
For me personally, this year’s re:Invent was a first as I attended as a customer, rather than as part of AWS itself. That shift in perspective sharpened the focus on what these announcements mean in practice, how businesses could and would implement them, where value can realistically be unlocked, and what tradeoffs will need to be made.
One of the most memorable moments of the event was Werner Vogels’ keynote. Not only was it typically insightful and thought-provoking, but it also marked his final keynote before handing the delivery in future years over to the next generation of AWS leadership. Werner has long been a voice of pragmatism amid the hype, and that was evident again this year, particularly in his framing of AI’s impact on developers and engineers.
Unsurprisingly, AI dominated the agenda. AWS previewed agents targeting specific software roles such as DevOps,and Security. You can now automate penetration testing for web application. These previews are impressive, but they also highlighted an important nuance – context matters. What each individual system, application, or organisation requires from AI agents will vary significantly. At least for now, these tools are most effective when used alongside skilled people rather than as outright replacements.
This creates a strategic challenge for companies. Many are now facing the complex task of identifying where AI agents can be integrated effectively without compromising the quality of outputs. Today, most of these implementations are still governed centrally through policy guardrails, which helps manage risk but can also slow experimentation as companies find a way to tailor to their own organisational needs. Striking the right balance between innovation and control will be critical.
One particularly notable announcement was the introduction of Kiro, a new Integrated Development Environment (IDE) and the first AI coding tool built around spec-driven development. The implications here are significant. For companies that have invested in producing internal standards, ways of working, and delivery patterns, Kiro will help developers use Large Language Model (LLM) to produce solutions that more closely match expectations.
Developing the solution is just one part of the software development process. Software developers still need to consider everything else. Fundamentals are more important than ever. New pain points will arise as other parts of the process attempt to keep up. It doesn’t matter how fast you can add features if you still have manual testing, or a bi-weekly release cycle.
This tension was echoed strongly in Werner Vogels’ keynote who made it clear that history suggests that AI and the rise of AI agents is just the start of a new beginning. We have been through similar shifts before – from COBOL and drag-and-drop tooling to the move from on-premises infrastructure to the cloud. Each wave has simply driven an evolution of the role, not eliminated it.
What we are seeing now is another new beginning. Our industry evolves, and so do the tools we use. AI will automate certain tasks, replace others, and fundamentally transform how software is built and operated. But AI does not make developers obsolete, provided they evolve with it.
Finally, AWS also announced several changes aimed at cost optimisation. New database savings plans which can be applied to any database service. These can reduce your database costs by up to 35% when you commit to a consistent amount of usage over a one-year term. Savings plans allowing customers to switch between database types without losing their discounts. CloudFront pricing has also been diversified with the introduction of flat rates to help customers better manage cost fluctuations.
In summary, AWS re:Invent 2025 made one thing clear – we are entering a period of rapid transformation. The opportunity is enormous, but so is the responsibility to adopt these technologies thoughtfully. Those who succeed will be the ones who combine powerful new tools with skilled people, clear strategy, and a willingness to evolve.
leighton.com