Artificial Intelligencer-How AI and politics dominated Davos

By Krystal Hu and Jeffrey Dastin

Jan 21 (Reuters) - (Artificial Intelligencer is published every Wednesday. Think your friend or colleague should know about us? Forward this newsletter to them. They can also subscribe here or email me to share any thoughts.)

How real is AI demand?

It’s the question hanging over markets — and one that C.C. Wei, CEO of TSMC 2330.TW, has been asking himself too.

What’s on the line for Wei, the 73-year-old steward of the world’s most important chip manufacturer, is more than $52 billion in capital spending this year to expand capacity, jumping as much as 40% from 2025.

“If we didn’t do it carefully, that would be a big disaster to TSMC for sure,” Wei said.

Over the past four months, Wei has gone beyond the usual customer checks, speaking not only with TSMC’s direct clients but with their customers’ customers, to verify whether AI demand is real and durable enough to justify that level of investment.

His conclusion: it is.

AI, he said, is already helping companies grow in ways that show up in real financial returns. Even TSMC itself is seeing productivity gains from using AI software internally, something Wei expects to lift margins.

“I look at AI, it looks like it’s going to be endless — for many years to come,” he said, effectively betting the future of the 38-year-old company on AI’s ability to reshape the famously cyclical semiconductor industry.

This week, we’ll dig deeper into how AI conversations are unfolding in Davos, and what corporate leaders are planning headcount as AI moves from promise to deployment. Scroll on.

OUR LATEST REPORTING IN TECH & AI:

• OpenAI says Stargate project will not hike consumer electricity bill

• Young workers most worried about AI affecting jobs, Randstad survey shows

• Trump's greenlight for Nvidia AI chips to China draws fire from lawmakers, former officials

• OpenAI to test ads in ChatGPT in bid to boost revenue

• Global sports face challenges from 'AI slop' misinformation

DISPATCH FROM DAVOS: AI EVERYWHERE, POLITICS LOOMING LARGER

There are essentially two topics dominating conversations in Davos this week: AI and President Donald Trump.

The tech presence at the World Economic Forum's annual meeting has always been sizable, but this year it feels bigger, louder, and more embedded in the Swiss alpine town.

Anthropic, a little-known AI lab just a few years ago, for the first time set up its own office on the main drag — a clear sign of how aggressively it’s courting enterprise customers. Google, meanwhile, hosted its press event for the fourth straight year, drawing what appeared to be its largest crowd yet.

Demis Hassabis, co-founder and CEO of Google DeepMind, said the path to human-level artificial general intelligence is becoming clearer — but still has “missing ingredients.” He put AGI at five to ten years away, longer than timelines floated by peers at Anthropic and OpenAI, where executives have suggested it could arrive as early as 2026 or 2027.

Even Davos’ side shows carried an AI twist. The town’s favorite piano player — a fixture during the week — was drawing interest for events hosted by Cloudflare NET.N, a company that is building a licensed AI data marketplace aimed at curbing unapproved “AI scraping.”

Politics, though, was never far from the surface. Executives buzzed all week about Trump’s speech on Wednesday, speculating who might join him for dinner afterward. His renewed focus on Greenland, and the risk of a fresh trade clash between the U.S. and Europe, dominated hallway conversations. In the audience for Trump’s address were two rare Davos sightings seated side by side: Nvidia NVDA.O CEO Jensen Huang and Apple AAPL.O CEO Tim Cook.

Trump insisted he would not seize Greenland by force, but left little doubt about the pressure he intends to apply if Europe blocks his ambitions. For Europeans, it underscored a reality they’ve been grappling with since last year’s WEF: the need to reduce dependence on the U.S. and accelerate investment in homegrown tech, from AI to defense.

Still, AI was everywhere — in conversations with Microsoft MSFT.O CEO Satya Nadella, OpenAI’s policy head Chris Lehane, and in Trump’s own remarks.

Anthropic CEO Dario Amodei was blunt as usual, warning of potential job losses as AI capabilities accelerate. He also criticized U.S. policy for allowing advanced American chips to flow to China.

Optimism remained, but so did reality checks. Meta’s META.O CTO teased that a new model from the company’s “superintelligence” team has landed internally — expectations for the effort, fueled by an expensive talent war, will be closely scrutinized after the recent Llama setback.

Some earlier, ambitious promises are already slipping. Isomorphic Labs, the AI drug-discovery startup spun out of Alphabet, now is hoping for its first clinical trials by the end of 2026, Hassabis said, later than the end-of-2025 timeline he shared at the same conference last year.

If Davos is any guide, AI’s future still looks enormous — but could be far less linear than the confident forecasts shared on these same stages 12 months ago.

CHART OF THE WEEK:

A survey from Wing Venture, which collected responses from 181 C-level tech leaders across enterprises, offers a candid look at how large companies are planning their headcount as AI moves from pilots to production.

About 66% of large enterprises (10K+ employees) expect headcount reductions of 10%–25% over the next three years as they deploy AI agents. Smaller companies anticipate cuts as well, but the expectation is notably more pronounced among large firms, which have more scale and more processes to automate. Wing Venture’s partner Peter Wagner says the survey does not point to across-the-board layoffs. Instead, it highlights targeted reductions in specific functions where AI-driven efficiency gains are most immediate.

(Reporting by Krystal Hu and Jeffrey Dastin; Editing by Lisa Shumaker)

((krystal.hu@thomsonreuters.com, +1 917-691-1815))