Tech CEOs at Davos Focus on AI Returns and Risks

Tech CEOs spent much of their time at the World Economic Forum in Davos, Switzerland, this week talking about artificial intelligence. The event drew business heads from Silicon Valley and beyond to the snowy Swiss town. They moved from broad excitement about AI to hard questions on costs, profits, and changes to work. This shift came as companies face pressure to show results from big AI spending.

Background

The World Economic Forum brings together company leaders, government officials, and others each year in late January. This year's meeting took place amid global talks on trade and politics. But inside meeting rooms and over coffee, AI stood out. Last year, many spoke about huge investments needed to build AI tools. Now, in 2026, the chat turned to payoffs. CEOs said early tests of AI, like giving workers chat tools, did not deliver big gains. Workers used them for small tasks, but profits stayed flat. Leaders now push for bigger changes led from the top. They met in panels, breakfasts, and side talks to swap notes on making AI work across whole companies. Names like OpenAI, Microsoft, and Anthropic came up often as examples of firms chasing enterprise deals over consumer apps.

Key Details

Talks boiled down to four main points. First, CEOs want clear returns from AI money spent. They call this ROI, or return on investment. One leader said 2026 is the year companies must prove AI boosts output and cuts costs. Firms like Cisco and IBM put up signs in Davos claiming they cracked the code for AI profits. But many admitted scaling AI across a business is tough. It needs new setups for data and worker training.

Shift to Top-Down AI Plans

Early on, companies let staff pick AI tools on their own. That led to scattered use with little impact. Now, bosses take charge. They pick key areas like supply chains or customer service for AI overhauls. Jim Hagemann Snabe, head of Siemens, told a group of executives that CEOs must act like dictators to push these plans. He said front-line workers did not spot the best spots for AI, so leaders must step in.

"CEOs need to be dictators in identifying where their businesses would deploy AI and pushing those initiatives forward." – Jim Hagemann Snabe, Siemens Chairman

Second, a race is on for software that runs AI agents in offices. These are smart programs that handle tasks like humans. Companies like Workday, Microsoft, and Snowflake say they hold key data on workers and money. They aim to manage AI agents just like they do people—handling access, tasks, and checks. Workday's CEO sees his firm as the front door to work for both humans and AI. Microsoft's tools link data from different spots without big moves.

Third, job losses loomed large. Some CEOs warned AI could replace many roles soon. Anthropic's Dario Amodei said artificial general intelligence—AI that matches human thinking on most tasks—sits two to three years away. Humanoid robots could follow in seven to eight years. He put jobless rates at up to 20% if not handled right. Others split on speed but agreed firms must speak plainly. One FTSE 100 chair compared Amodei's warnings to more hopeful views from Google execs who see new jobs ahead.

Fourth, risks beyond jobs drew note. Leaders flagged deepfakes, false info, bias in AI, and worker displacement. Few worried about sci-fi takeovers, but real issues like bad data use stayed front of mind. Amodei also hit out at easing chip sales to China, calling it a security risk like selling nuclear gear to bad actors.

What This Means

These talks signal AI's next phase for businesses. Companies face choices on spending. Bottom-up tools fade as top plans take over. This could speed profits but demands big shifts in how firms run. Winners in the agent race, like Microsoft or Snowflake, may control enterprise AI flows. That gives them power over data and work.

Job shifts bring wider effects. If AI cuts roles fast, governments may need new plans like training or aid. CEOs called for joint action from business and leaders. Amodei and Google's Demis Hassabis spoke of slowing AI growth if all nations joined in. They said solo pauses would not work without global deals.

Regulation ideas varied. Snabe pushed rules tying AI to human values, not case-by-case limits. He targeted ad-based AI models, saying they chase user time like social media harms. On chips, Amodei urged tight U.S. controls to keep AI leads safe.

Optimism lingered. Enterprise focus offers steady cash over consumer ups and downs. Amodei noted his firm's business clients bring direct value. Panels on scaling AI heard from oil, consulting, and payment bosses on hurdles like staff buy-in and tech fits. Solutions include clear goals, pilot tests in one area, then company-wide rollouts.

Davos chats show tech heads balancing push and caution. They eye AI as a base tech like the internet but know paths to value run steep. Firms that nail orchestration and processes stand to gain most. Others risk sunk costs if plans stall. Policy makers watch close as AI touches trade, security, and daily jobs.