The future is already here – it’s just not evenly distributed.” – William Gibson, Author
With 2025 barely here, William Gibson’s quote feels more accurate than ever. Things appear to be in flux everywhere.
Historian Niall Ferguson recently diagnosed an ongoing cultural „vibe shift“ while, at the same time, we are seeing accelerated progress in Technology and Biotechnology, as exemplified by recent breakthroughs in Quantum Computing and protein structure prediction.
(And let’s not even start with politics.)
A lot is going on. For this reason, I won’t point you to individual reads today, but I will share what I make from looking into my crystal ball.
On a Macro Level, two developments stand out for me
AI
Hardly unexpected, these two letters will continue to dominate anything with a software component (and then some).
While I am not sure the general progress in foundation models (LLMs) will continue at the same rate as in ’24, the application of AI will start to bear fruit in domains beyond the language-oriented professions the current chatbots are largely catering to.
In a recent „State of Product Management“ report covering a cross-section of predominantly US industries, roughly 66% of polled product managers confirmed that they had started integrating some form of AI into their products.
I expect to see significant progress in agentic AI and „Service as Software,“ as NYU Professor Scott Galloway labels it. I would not be surprised to see the first truly useful large action models (LAMs) possibly combined with robots applying them in 2025.
2025 will also be the year that will see serious challenges to OpenAI’s dominance. Amazon’s new family of Nova models and Google’s recent updates to Gemini are impressive. Sadly, beyond Mistral’s, no European models seem to be able to keep up.
At the same time, smaller yet powerful open-source models like Meta’s Llama 3.x even run locally (e.g., on the MacBook, I am writing on) and can be fine-tuned or augmented for specific use cases without creating data protection and security issues in their application.
Energy (and Data Centers)
The other macro topic I foresee dominating the discussions in 2025 will be energy, with grid stability increasingly becoming a focus. In some parts of the US, grid connection queues for renewable energy projects have grown to more than five years due to the need to individually study the impact of adding more solar or wind projects to the already challenged grid. Similarly, the CEO of German network backbone provider 50Hertz suggests stretching out the buildup of solar capacity to avoid overloading the infrastructure. PV companies Enpal and 1komma5grad even warn of brownouts and blackouts caused by PV power oversupply as early as this Easter.
At the same time, power requirements for data centers alone are forecast to double (US) and triple (Europe) by 2030, driven by AI and continuing digitalization. Fears of rolling blackouts led Ireland’s grid operator to halt the construction of new data centers near Dublin until 2028. According to official records, data centers consumed 21% of the nation’s electricity last year.
Combine ideologically misguided policy-making with global and regional instability, and a perfect storm seems on the horizon (figuratively speaking; still, climate change certainly contributes to the melee). It will require a joint effort from innovative technologists and rationally thinking politicians to safeguard our economies.
Closer to our offices, I see things shifting too:
Agile Doctrine Continues to Succumb to Agile Pragmatism
In a recent post about the decline of the Scrum Master’s role, the renowned German agilist and trainer Stefan Wolpers writes, „If you hang out in the “Agile” bubble on LinkedIn, the dice have already been cast: Scrum is out…“. And indeed it would appear so. Fewer participants in agile trainings, fewer Scrum Master and Agile Coach roles in job portals. I believe there are two factors at play: Agile is finally normality. Organizations have developed some agile derivative based on the key agile concepts that delivers in their context. In addition, increasing outcome orientation shifts the focus from the (Scrum-) process to delivering its intended results.
The Product Operating Model is the new Kid on the Block
Increasingly, there is also an impression that Agile’s processes are overly bureaucratic and fail to deliver adequate customer focus and business agility. Truth be told, none of the tech giants follow a mandatory agile process, let alone apply a compulsory agile scaling framework. The Product Operating Model is Big Tech’s approach to developing successful products. This is not getting lost on an increasing number of executives. With rising economic pressures forcing digital and IT organizations to de-scale, I expect many organizations to switch to the Product Operating Model in 2025. After all, even McKinsey began touting its benefits about a year ago.
Successful CIOs will be well-versed in Business Model Development
Data, machine learning, and AI increasingly impact companies‘ business models. CIOs who can actively improve and shape business models in cooperation with their business counterparts will make a critical difference in the success of their companies.
Cyberattacks Will Reach New Levels
Sadly, I expect AI and increasing geopolitical tensions to lead to a new level of cyber threats in 2025, increasingly targeting critical infrastructure. Ransomware- and identity-related attacks are expected to increase in particular. If attacks continue to grow similar to 2024, more than 90% of companies in Germany will experience some form of cyberattack. Based on recent rates, the overall damage caused – estimated already at approximately a quarter trillion Euros in Germany in ‘24 – may grow even faster.
Structured Hybrid Becomes the New Normal
Structured hybrid work with 3-2 (3 office days – 2 WFH days) will become the „new normal“ for most enterprises. As scientific evidence, such as that provided by Stanford economist Nick Bloom, increasingly demonstrates that productivity does not decline under this scheme, along with additional benefits like reduced attrition rates tipping the scales in its favor, most larger enterprises are likely to adopt it formally.
N.B. High-profile exceptions will apply in specific dynamic tech environments that rely heavily on „swarming“ as a problem-solving approach (Tesla!).
These predictions are my personal views. If you find them underwhelming or too obvious, new glasses may help me get better next year. I have been eyeing Rayban’s MetaAI-augmented glasses for a while, but Halliday’s recent CES announcement really got me excited.
Have a great start to 2025!
PS: Check out our new „Lab-Project“: „Deep Dives with Jane and Austin“, a growing series of quirky shortcasts created from my research materials by using a stack of AI-tools.