A few years ago, AI felt like a fantastic gimmick. Last week in Amsterdam, that illusion fell apart completely. I've been part of the circus that the IT sector sometimes is for many years now. Just when you think you've seen everything, reality catches up with you again. And that reality is, of course, AI.
Our generation had its 'man-on-the-moon moment' with the launch of ChatGPT. But honestly, back then, for many of us, AI was nothing more than a clever party trick.
Last week, I attended the Google Cloud AI event in Amsterdam (which had been given a completely new name for the occasion) together with clients and colleagues. It was a clear eye-opener: there was truly not a single session where AI did not play the leading role. The one-trick pony was learning new skills, and the practical transformation has truly begun.
These are my most important takeaways.
If you want to know how urgent the situation is, you only have to look at the graph on ZeroDayClock.com. Security is simply no longer something you can follow up on manually.
The new reality: Where there used to be months between the discovery of a vulnerability and the actual exploit, we are now talking about a matter of hours.
Without aggressive automation and AI-driven security, your organization will fall hopelessly behind. It's not a question of 'wanting to,' but of survival.
Google's answer to this challenge is their acquisition of Wiz, which adds very deep cloud security to the portfolio.
A striking trend: the number of AI agents developed by employees is growing into a serious KPI within organizations.
Google is capitalizing on this cleverly with Gemini Enterprise. This framework has the ambition and the features to become the central AI point of contact for many companies.
Thanks to seamless integrations with platforms like Microsoft 365, Jira, and Google Workspace, you can now give AI tasks that work across platforms on all your data sources simultaneously. Finally, those stubborn data silos are being broken down.
Google explained numerous different ways to build agents yourself, each tailored to a specific target audience (from low-code for the business to advanced for the data engineers).
In the presentations, Google heavily emphasized their unique position in the AI landscape. As one of the few industry players, they can offer a complete end-to-end story. And you noticed that in the coherence of the presentations.
Everything happens in-house:
The beautiful thing is that Google does not dictate this as a vendor lock-in. By open-sourcing their new protocols for processing payments in inter-agent communication, the platform remains flexible enough for integrations with other AI providers. For example, we have customers who combine Google Cloud Gemini and Claude to give their customers the best experience. Without hacks, without any friction.
It's an open secret that a giant like Anthropic trains its models on Google's hardware. During a deep-dive session on these TPUs, it became truly clear why.
The underlying architecture, with optical interconnects, self-healing networks, and massive scalability, made many a tech heart beat faster. The fact that this technology is many times more efficient than traditional Nvidia GPUs was, of course, emphasized by the Google speakers with a broad smile.
I'm not easily surprised by a marketing pitch anymore. But what I saw in Amsterdam was not future talk; it is the tech stack of today and tomorrow. AI has definitively emerged from the gimmick phase, and Google Cloud is firmly back in the race. As a Google partner, we are riding along on their new developments, ready to tackle the projects of tomorrow.