News analysis: Why 2025 will be the era of AI

From AI breakthroughs to infrastructure investments, 2025 reshaped the world of technology. CGTN's Sally Ihan speaks with Constellation Research Principal Analyst Ray Wang about the year's biggest technology milestones and the competitive forces heading into 2026.

sauce: CGTN America

In 2025, AI will dominate and permeate every aspect of our lives. Generative AI has moved into the mainstream, increasing the ability of models to process text, images, audio, and video. We've improved our ability to analyze and create this multimodal content. We find that moving from generative AI to agenttic AI allows companies to automate work processes and accelerate decision-making speed. Advances in humanoid robots have raised hopes in the West that physical AI will be ready in time as populations age and birth rates decline. Cybersecurity has made breakthroughs in attacking and defending against AI-driven attacks.

Big tech companies have poured billions of dollars into infrastructure such as chips, power and data centers. The impact on venture capital, private equity, and startup funding has changed forever. In fact, the battle for data center supremacy has taken the market by surprise. Data centers have surpassed commercial real estate to become the largest investment area. Data centers drove more than $61 billion in transactions and investments in 2025. Amazon, Google, Microsoft, Meta, and others dominated the investment, with CoreWeave, Oracle, and other NeoClouds fueling the frenzy.

Major Chinese companies such as Alibaba, Tencent, ByteDance, DeepSeek, Shibaura AI, and Hyakuchuan have pledged to invest up to $70 billion in data center expansion by 2026. What we learned in 2025 is that natural resources and real assets are more valuable than ever, as power, water, real estate, and critical minerals are central to any AI project.

A backlash from investors questioning the return on investment from AI intensified in the fall of 2025 after nine months of turmoil. There were some reports of AI burnout and overload, but they were short-lived. Companies that actually succeed with AI show 10x to 100x returns, outperforming companies that wait too long. Those dabbling are following the AI ​​exponential A-class that has emerged to outperform competitors without AI capabilities. After all, AI is a long-term investment game where the winners take the market.

Constellation Research estimates that China's AI industry is worth $150 billion, as it is embedded in smart cities, logistics, healthcare, pharmaceuticals, agriculture, robotics, consumer chatbots, and more. China's growth in integrating AI into the real economy sector has been significant from every perspective. From AI models to semiconductors and robotics, more than $125 billion of state funding has been poured into companies like DeepSeek, Huawei, Alibaba, Tencent, and Zhipu AI. Open source LLM has reached its golden moment, and Alibaba Tongyi and Qianwen have taken off.

AI robots like Agibot have taken hold as the race for humanoid robots reveals not only their advances but also their major limitations. Huawei has launched the Ascend 910 chip. Meanwhile, DeepSeek R1 and Wudao 3.0 have moved forward with the development of open source LLM. Overall, the competition between China and the US was very competitive, with the US having a slight lead.

Looking ahead to 2026, AI agents and autonomous workflows are expected to grow rapidly. Leaders are expected to focus on several areas such as chip advantage. Expect faster chips, higher power efficiency, and expanded TPUs for inference. Faster chips, advances in TPUs. In terms of model adoption, the battle between Chinese open source and Western models and their ecosystems will be decided by developers.

In the area of ​​data dominance, the battle for datasets will intensify as companies compete for data sources that feed the voracious appetite of models and the constant updating of signals. The emergence of the Data Inc company and its business model will materialize in the second half of 20206.

Real-world ROI permeates every results-based conversation. Winners achieve exponential efficiency and limitless possibilities.

The race continues. This battle is about mass adoption and investment. China leads in power efficiency and open source LLM The US has the model, the chips, the market ecosystem The real question is who has the better business model, which systems will go bankrupt first and which countries will benefit the most from AI investments. This will be the focus of attention in 2026.