By Krystal Hu

Jan 14 (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.)

Saks, the high-end department store, has just filed for bankruptcy in one of the largest retail collapses since COVID. As fewer shoppers head to stores for advice from sales associates, many are turning to a new kind of helper when they purchase: AI.

The past holiday season was a case in point. As consumers spent a record $257.8 billion online, AI-driven traffic skyrocketed 693%, according to Adobe data. These AI-guided shoppers mean business and were 31% more likely to click “buy” than those from other sources, nearly double last year's rate.

Consumers also appear increasingly comfortable with AI guidance. Adobe found they spent 45% more time on sites and viewed 13% more pages per visit when arriving via AI-powered recommendations — signs of higher intent and trust. That trust is pushing Walmart WMT.O to partner with ChatGPT and Google’s Gemini to let AI users discover and buy directly from its site.

From changing how consumers shop to reshaping how enterprises run, the Reuters tech team is setting our sights on the biggest themes and predictions in the new year. In a world moving this fast, the only safe bet is change. Scroll down to see our predictions.

OUR LATEST REPORTING IN TECH & AI:

• Exclusive-China's customs agents told Nvidia's H200 chips are not permitted

• Microsoft rolls out initiative to limit data-center power costs, water use impact

• Britain seeks 'reset' in copyright battle between AI and creators

• Advocacy groups call on Apple and Google to remove X over sexualized deepfakes

• Belgian cybersecurity startup Aikido hits unicorn status with new funding round

• Honeywell plans to carve out quantum computing unit Quantinuum via IPO

AI THEMES TO WATCH IN 2026

As market sentiment swings between AI optimism and caution, here are our three predictions that will shape the next phase of the AI boom this year:

1. Big funding, but not yet full IPO boom

Investors are expecting to pour more capital into the hottest private names dominating the AI race, while preparing for a longer timeline before they can cash out through blockbuster IPOs.

The largest foundation model labs continue to attract funding on a scale typically associated with scaled public companies. In just the past few weeks, Elon Musk’s xAI raised $20 billion, Anthropic is in talks to close a round of up to $10 billion, and OpenAI received $22 billion from SoftBank.

Last year’s funding already showed extreme concentration, with a handful of labs absorbing the bulk of AI funding. In fact, nearly two-thirds of global venture capital in 2025 flowed into AI, according to PitchBook. Investors I speak with expect that trend to continue — with new capital competing for access to fewer companies and ever-larger checks.

“The consensus is that capital is effectively unlimited for the best private AI companies right now — they can raise money on their own terms,” said Larry Aschebrook, managing partner at G Squared, who sees little urgency for these companies to tap public markets this year. Public investors want predictable cash flows, while frontier AI labs are still burning capital to build scale — a mismatch likely to keep them private for longer.

2. The next bottleneck in data centers is moving beyond chips

The industry has spent the past two years fixated on GPUs. That fever is expanding. Our reporting shows the next constraints are physical: not just high-bandwidth memory and power grid but the basics of having enough skilled electricians and construction crews. Even when chips are available, getting them online has become the real challenge. The AI race is becoming more about who can secure land, electricity, labor and long-term infrastructure — and finance it creatively enough to keep projects moving.

There’s also growing political risk. With U.S. midterm elections in November, potential backlash and policy shifts are becoming harder to ignore. While the current administration has largely encouraged data-center buildouts and new power generation to keep the U.S. competitive in AI, local communities are pushing back after seeing rising utility bills and heavy use of land and water. Companies are already responding. Microsoft MSFT.O, for example, said this week it would adopt a new rate structure designed to prevent data-center power costs from being passed on to consumers — a sign that public scrutiny could play a bigger role in how the AI infrastructure boom unfolds.

3. The trillion-dollar question: Will businesses keep paying?

Enterprise AI spending will ultimately determine whether today’s trillions in market cap and billions in startup valuations are sustainable. Tech giants like Amazon AMZN.O and model makers such as OpenAI and Anthropic are all chasing a meaningful share of corporate AI budgets. Our reporting shows that while AI pilots are everywhere, returns remain uneven, and executives are often reluctant to talk publicly about what hasn’t worked. If companies double down on spending based on the promise of agents and automation, the boom is justified. If they don’t, the gap between capital spending and revenue will become increasingly difficult to ignore.

This question also sits at the heart of AI’s circular financing model. In OpenAI’s case, some of its largest investors are also its biggest customers. Cloud providers make that loop work by persuading more enterprises to spend on AI models and infrastructure, so as long as fresh demand keeps coming in, the balance holds. But with even today’s best models still falling short of being fully “plug-and-play” for businesses, we’ll closely watch this year whether AI becomes good enough — consistently enough — for companies to keep paying.

CHART OF THE WEEK:

Source: LinkedInLabor Market Report

As fears about AI replacing jobs are often met with executives insisting that AI also creates new ones, it’s worth asking: what kinds of jobs is this AI boom actually creating? LinkedIn’s latest data offers a useful reality check. Between 2023 and 2025, AI generated roughly 1.3 million new jobs. By sheer volume, the biggest driver has been data labelers, largely a lower-paying, part-time gig — a reminder of how much human labor still goes into training and refining AI systems. Companies are also formalizing AI leadership, creating nearly 300,000 “Head of AI” roles.

The fastest growth, though, is happening where technology meets real-world work. Demand for AI engineers — and especially forward-deployed engineers — has surged as companies struggle to make AI useful inside their organizations. LinkedIn data shows forward-deployed engineers' roles have grown more than 40-fold since 2023, even if they remain a smaller slice of overall hiring. These jobs exist because AI doesn’t “just work” out of the box. Someone still has to sit with the business, adapt models to messy systems, and make sure the technology delivers value.

(Reporting by Krystal Hu; Editing by Lisa Shumaker)

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