AI was the biggest discussion @Davos

Many believe the big themes of Davos last week were around Donald Trump, his ambitions to take over Greenland and Canada, and Mark Carney’s speech.

Stop “living within a lie”, Mark Carney #Davos

They were the headline news but, underneath those headlines, the main theme of Davos was AI. AI featured in almost every session and here are the bullet points that were most relevant to me.

There is clearly an:

Economic shift: AI is moving from experimentation to core economic engine, reshaping how productivity is delivered and where value is created — similar in scale to past transformational technologies like the internet or cloud.

Workforce evolution: Many current jobs may change or disappear, but new roles and sectors will emerge — those who adapt will thrive.

Competitive battle: The winners won’t be those with the smartest models — but those who use AI effectively, sustainably, and in ways that rewire real world outcomes.

Satya Nadella, CEO of Microsoft, made the point that AI isn’t just better tools but a platform shift like the PC, the internet, and the cloud. It digitialises reasoning itself, lowering costs and changing how work is done at a fundamental level. This results in massive implications for organisational structures as it flattens hierarchies, and shifts competitive advantage from who knows more to who can execute fastest.

Alex Karp, CEO and co-founder of Palantir Technologies, believe that AI doesn’t fix broken systems but that it exposes them. Organizations that don’t work well in reality (fragmented data, unclear processes, lack of accountability) will struggle because AI highlights their inefficiencies instead of masking them.

Dario Amodei of Anthropic made clear that the loop is the story. He’s not saying “models will get better”, but that we are close to a phase where AI writes most software end-to-end and starts accelerating AI research itself. If that feedback loop closes, timelines compress hard. The constraint won’t be ideas. It’ll be chips, training time, and deployment speed.

Demis Hassabis, Google DeepMind, pushed back on the lazy definition of AGI (Artificial General Intelligence) stating that the missing ingredient isn’t more compute power but creating  systems that can form hypotheses, ask the right questions, and deal with messy experimentation.

Larry Fink of BlackRock feel that diffusion will decide who wins. Fink’s investor frame was blunt: it won’t be the creators of the tech that win; it’ll be the economies and companies that adopt fastest. If AI doesn’t spread into healthcare, education, public services, and the real economy, we’re not in a revolution — we’re in a supply-side bubble.

The weird and wonderful alien called Elon Musk argues that the real bottleneck for AI is energy and infrastructure, not algorithms. He predicts that AI will become very cheap and ubiquitous as costs fall and, in the longer-term, AI-driven robots will handle most production, although energy constraints (grid capacity) is the limiting factor.

Key Takeaways

• AI is infrastructure: Reasoning becomes cheap and pervasive — not just an add-on tool.
• Organisational readiness matters: AI success depends more on structure and execution than model sophistication.
• Energy & execution over invention: The most valuable opportunities lie in scaling AI effectively, not just building models.
• Smaller teams can compete: With AI, fewer people can produce more output — but those who remain matter more.
• Diffusion > discovery: The biggest winners will be those who use AI to change real outcomes, not just talk about it.

The overall conclusion is that AI is no longer a tool or feature. It is becoming infrastructure that underpins entire workflows, companies, and economies. Reasoning itself is being digitised, lowering the cost of intelligent decision-making and reshaping how work and value creation happen.  This means:

• Productivity shifts: Work that once required many people can now be handled by AI-augmented teams.
• Organisations flatten: Hierarchies may shrink because AI reduces the need for mid-level coordination and information gatekeeping.
• Competitive advantage shifts: Not to those who have the fanciest models — but to those who embed AI into real execution and outcomes.

• Job disruption & transformation: AI could affect large percentages of roles—especially entry-level work—potentially displacing some jobs while creating others. Youth and less experienced workers may be disproportionately affected.
• New jobs will also emerge: Some leaders argue AI will create new employment opportunities even as it automates others.
• Skills & retraining gap: Workforce readiness and training become crucial to capturing economic benefits rather than exacerbating inequality.
• Operational capacity matters more than ideology: Organisations that actually work in messy real-world conditions perform better than those that are only good on paper.
• Shift in investment focus: Investors will prioritise teams and organisations with solid execution and structural readiness over those with just technical hype.
• Energy demand spikes: As AI scales, electricity and cooling capacity will become strategic economic constraints.
• New industries emerge: Robotics, energy solutions, and hardware infrastructure become fundamental sectors in the AI economy.

It is fascinating to see how quickly the world is changing in the Great Divergence of the Intelligence Revolution.