Davos 2026: An AI roundup

Davos 2026 marked a transition in how AI is discussed by those deploying it at scale: from inevitability to accountability.

Image: World Economic Forum

Every year, the richest and most powerful people in the world convene in a small Swiss town called Davos. Whether you agree with the elites or not, there’s no discounting the impact they have on the world and the importance of this gathering.

This year, AI was one of the main topics on the agenda. According to reports, Davos was abuzz with people trying to pitch their AI solutions to movers and shakers.

What’s interesting to me, though, was the considerable maturation of the AI conversations happening among the most established.

My Davos invite must have been lost in the mail, but according to reports, we’ve moved on from debating whether AI is a thing. Now humanity has to figure out how we’ll use it.

Davos 2026 was the transition from AI hype to enterprise-scale deployment, with leaders debating governance, economic impacts, and timelines for advanced capabilities.

---

The productivity paradox and the “Vanguard 12%”

While AI was prominent at Davos for years, the 2026 summit exposed a deepening “AI divide.” According to the PwC 2026 Global CEO Survey, 56% of CEOs report that significant AI investments have yielded no increased revenue and no measurable cost savings.

That being said, an elite group of firms, dubbed the “Vanguard 12%,” has achieved the elusive “double win” of growth and efficiency. Microsoft’s Satya Nadella warned that if real-world outcomes don’t soon match the rising investment, the sector risks being viewed as an economic bubble.

---

“White-collar globalization”

BlackRock Chairman Larry Fink offered a sobering framing of the transition, describing it as “white-collar globalization.” Just as the hollowing out of blue-collar manufacturing defined the 1990s, Fink suggests that AI is now performing a similar “stress test” on middle and upper management.

When software can synthesize strategic memos and quarterly reports with the fluency of a senior analyst, the traditional justifications for high executive salaries are being called into question. As the summit’s attendees noted, “intelligence” is fast becoming a commodity, so judgement is the remaining premium.

---

The apprenticeship crisis

As I wrote a few weeks ago, labour experts are warning of the erosion of the corporate ladder. PwC Global Chairman Mohamed Kandi highlighted an emerging “Apprenticeship Crisis,” noting that when firms automate the grunt work typically performed by junior staff, they inadvertently destroy the training grounds for future leaders. Without the experience of navigating the weeds of a profession, the next generation may lack the nuanced judgement required to audit the systems they oversee.

---

AI as Industrial Infrastructure

Nvidia CEO Jensen Huang consistently described AI as a “five-layer cake”—a stack comprising energy, chips, data centres, models, and applications. Huang argued that scaling AI is no longer a software problem, but a physical industrial buildout. He remained bullish on job creation, citing a “significant boom” in salaries for those building the infrastructure layer, from electricians to chip engineers. “Everybody should be able to make a great living,” Huang remarked. “You don’t need a Ph.D. in computer science for this.”

---

Displacement and the “Tsunami”

IMF Managing Director Kristalina Georgieva did not mince words, comparing the technology’s impact on the global labour market to a “tsunami”. The IMF estimates that 60% of jobs in advanced economies will be transformed or displaced, with JPMorgan’s Jamie Dimon stating unequivocally that the technology will lead to fewer jobs in certain sectors within the next five years.

---

The closing signal

The consensus at Davos 2026 was that we have reached a point of “sovereign AI.” This means tech leadership isn’t a corporate choice, but a requirement for national and individual survival. Yoshua Bengio warned that the danger lies in the “delusion” of intelligence without accountability.

For me, the message from the Alps is clear: The most valuable person in the room is no longer the one with the fastest answer, but the one who knows why the AI’s answer is wrong.

---

Your meeting cheat sheet

Davos jargon can be dense, but the concerns behind it are universal. If you want to translate these high-level signals into immediate professional credibility, try dropping these lines in your next sync:

1. When the team is discussing a new AI tool

• Don’t say: “This AI is really fast and impressive.”

• Say: “As we scale this, how are we defining the governance layer? If the system makes a ‘confident’ mistake, do we have a clear line of accountability for who audits the output?”

• Why it works: It shows you’re thinking about the “infrastructure” shift Jensen Huang mentioned, not just the tool itself.

2. When leadership is asking about ROI

• Don’t say: “I think it’s making us more productive.”

• Say: “The data from Davos 2026 shows a huge gap between the ‘Vanguard 12%’ and everyone else. To make sure we aren’t in the 56% ‘zero-ROI’ group, should we focus more on our foundational data layer before we launch more pilots?”

• Why it works: You’re using the PwC stats to back up a strategic pivot. It sounds expensive to not listen to you.

3. When discussing team roles or hiring

• Don’t say: “We need someone who knows how to prompt.”

• Say: “In a world of ‘white-collar globalization,’ we need people with the judgement to audit AI work. How are we protecting our ‘apprenticeship’ path so our juniors are learning the nuances of the business?”

• Why it works: It addresses Larry Fink and Mohamed Kande’s warnings directly. It positions you as a leader who cares about long-term talent, not just short-term automation.

4. The “mic drop” question to end the meeting

• Try this: “Where in our current workflow are we over-relying on system confidence and under-investing in human verification?”

---

AI in the news will return

If you found this week’s newsletter useful, share it with someone who wants to learn more about AI.