IBM: AI shifts from hype to profit engine, bubble risks loom

If it is said that 2025 was the year of artificial intelligence, or AI, that would not be an exaggeration, because it was the year AI became involved in human life to an extent never seen before in almost every aspect, from daily life to the global economy and geopolitics. Meanwhile, 2026 will be the year AI is tested as “the real thing”, which must deliver business, economic and policy outcomes in concrete form.

2026 may be the beginning of shedding the old image of AI as a pilot project or a tool to raise productivity and reduce costs for repetitive tasks, and entering an era in which AI becomes an important mechanism for creating new business models, creating new revenue streams, and determining competitive advantage in the digital era.

This is consistent with a study finding that 95% of senior executives worldwide see that within this year AI will be able to generate revenue back to replace the money invested (self-fund), at least in part.

The report APAC AI Outlook 2026: Transferable Value Across Industries, co-developed by IBM’s Institute for Business Value (IBV), reflects AI trends through a business lens in five main industries of the region, covering finance, telecommunications, manufacturing, energy and the public sector, as well as five AI drivers that look beyond Generative AI to Agentic AI.

Joohee McClelland, Managing Partner of IBM Consulting Asia Pacific, said organisations are moving beyond the experimental stage to building AI with good governance, that can explain where it comes from and why, and that aligns with strategic priorities at both national and organisational levels, from using sovereign models to quantum-powered AI. Under an advanced and trustworthy AI ecosystem, these are paving the way for organisations to grow strategically and sustainably.

AI transforms five Asian industry groups

The APAC AI Outlook 2026 report revealed important trends in using AI to drive five main industries in the region this year in a significant way.

First, the “finance and banking” group looks at using AI to develop new business models such as “trust-as-a-service”, using strengths in risk management and proactive governance to open new revenue sources and create competitive differentiation; or “super apps” driven by AI that provide real-time, highly personalised advice through smooth, natural, human-like conversations. Meanwhile, AI agents working with legacy systems can raise operational efficiency without having to rip out and rebuild the system.

Next, the “manufacturing” group will move from fragmented systems to intelligent supply chains and predictive automation manufacturing, where AI can forecast machine problems, adjust operations and raise production efficiency in real time, while moving towards mass personalisation at scale through advanced AI and digital twin technology that develops products quickly and cost-effectively.

For the “telecommunications” group, providers will look at bringing in Agentic AI to manage networks, operational support systems and customer experience to solve problems faster. Meanwhile, AI ecosystems will turn data into real-time insights, and digital services through cross-industry collaboration platforms will lead to new lumps of revenue.

Next, the “energy” group will see AI help accelerate the energy transition, making systems smarter, more resilient and more efficient, improving grid stability and integrating renewables. “Green AI” will appear, in which AI helps reduce its own energy use and carbon through optimising hardware, workloads and cooling systems. Self-adapting infrastructure will cover everything from chips and operating systems to carbon-emissions tracking, allowing AI to create value without adding unnecessary energy burdens.

Finally, in “public services”, AI will help raise productivity and agility, allowing people to access services in finance, agriculture, public health and education at appropriate prices and aligned with local context. In 2026, government agencies will place importance on responsible, safe, ethical and transparent AI through strong governance frameworks to reduce risk and build trust from society.

Five big AI trends move business

IBM also summarised five AI technology trends that will drive organisations across the five key business sectors:

Internet of AI: Generative AI and Agentic AI will become core organisational infrastructure. Foundation and specialised models will work together to improve efficiency, governance and innovation at scale. At the same time, AI will move from centralised data centres to distributed systems that consider privacy. AI will learn through smartphones, IoT and edge devices using persistent memory, on-device processing and privacy-first design.

AI as a growth accelerator: Efficiency is only the starting point. The true return on investment (ROI) of AI will happen when AI creates competitive differentiation, transforms business models, and creates new products, services or revenue sources—turning AI from a cost centre into an investment that generates sustainable returns.

ROI from trustworthy AI: Organisations that invest highly in AI ethics tend to generate profits and ROI from AI better on an ongoing basis.

Sovereign AI: Sovereign AI will proceed alongside global AI models. Local models and sovereign infrastructure will grow to meet data sovereignty, language diversity, cultural context and national security needs, reflecting a shift in the structure of the global AI landscape.

Quantum & AI: As quantum computing approaches “Quantum Advantage”, the technology will have potential to accelerate AI model training through optimisation, sampling and simulation of complex systems. Conversely, AI will improve quantum workflows, including algorithms, error correction and resource allocation.

Summary for Asian and Thai business

Joohee said the most exciting thing in Asia Pacific is that AI innovation is no longer limited within a single industry.

“We are beginning to see progress in telecommunications that becomes inspiration for manufacturing, and prototypes from manufacturing are being extended into energy. The learned value and cross-industry transfer itself accelerates AI readiness across the region, helping organisations leapfrog beyond old development cycles. It reflects that organisational leaders are scaling AI use with purpose and focusing on measurable growth.”

If 2024 was the year of “awakening” and 2025 was the year of “accelerating build”, 2026 will be the year the winners are those who truly turn AI into profit and sustainability.

From Gen AI to the Agentic AI era

One of the most closely watched trends is the shift from Generative AI, which “helps think, helps create”, to Agentic AI that can plan, decide and act by itself.

From now on, AI will not stop at answering questions or creating content, but will begin to take on the role of process manager, working in place of humans in tasks that are more complex and continuous.

Salesforce said Thailand’s AI market is growing rapidly. Industry analysis forecasts average annual growth as high as 32.33%, with the total market likely to reach around 233,000 million baht by 2031.

The technological revolution with Agentic AI opens an important growth opportunity for Thai organisations, by strengthening workforce capability through AI agents that increase work efficiency, upgrade business operations, and improve customer experience.

AI as a necessary cost of competition

However, 2025 showed clearly that having AI is not a strategic advantage, but has become a “necessary cost” of survival. Organisations that do not invest in AI will begin to lose out in speed, cost and decision-making ability. Organisations that invest slowly or without a system will face growing problems in keeping up with competitors, as will public sector organisations.

A Gartner survey found that within 2026, 80% of chief information officers in government organisations will raise budgets for GenAI by 38% compared with the previous year, indicating pressure to expand GenAI use to deliver faster and more efficient public services.

This picture is not only at organisational level, but extends to the national level. Countries with ready digital infrastructure, talent and governance frameworks will attract more investment and innovation. Countries that still see AI only as a supporting tool risk being left behind in the global economic chain.

Thai tech guru: AI battlefield in 2026 still fierce

Dr Thanachart Numnonda, founder and director of the IMC Institute, analysed that looking at the overall AI market from the transition between 2025 and 2026, interesting changes are appearing. Even though Google has risen again in technology, in brand and consumer perception, the word “ChatGPT” remains a household common noun that people use to refer to AI, with more than 700 million users per week and consumer chatbot market share as high as 75%, giving OpenAI a huge business advantage.

But this throne is not permanently stable, because a competitor like Google is breathing down its neck, seen from the number of Gemini app users surging to 650 million from 450 million.

How does this competition affect Thais? Even though official Thai-language test results are not yet disclosed, it can be forecast that Gemini 3.0 will likely be one of the best-performing models in Thai, because Google has a huge Thai language data trove from Search and YouTube. The model’s ability to understand video and images will help it understand cultural context and dialect better.

Watch the Thai-language model Typhoon

At the same time, Thai models are starting to create their own space, such as Typhoon, developed further for specialised work that foreign models still do not do as well, especially work involving complex Thai documents such as tables and government forms, or understanding local dialect.

Therefore, this is not an era of sticking to only one AI anymore. Smart businesses and developers will use a method of combining multiple models.

Finally, what will happen next is a capital war and frenzied rapid development. It is forecast that major technology companies will collectively spend more than US$400 billion in 2026 to compete for No.1. Product or model launch cycles that used to be six to seven months will be shortened to only a few months, or perhaps only a few weeks.

The direction of technology will move towards AI agents—intelligent assistants that can truly think and act in place of humans, not just answer questions—such as automatic code writing, ticket booking, or handling complex documents.

Economic opportunity or a major bubble

Bloomberg reported that concerns over an AI bubble have continued from 2025 into 2026, because investors have poured in money at levels never seen before, but in reality almost no one knows for sure whether these funds will deliver returns as expected.

Technology companies spending hundreds of billions of dollars on advanced chips and data centres are doing so not only to support rapidly growing chatbot use such as ChatGPT, Gemini and Claude, but also to prepare for a deeper structural shift: the movement of economic activity from humans to machines. Total costs could eventually rise to the trillions of dollars, funded by venture capital, debt, and more recently, novel “loop” financing structures, raising concern among Wall Street investors.

China in the chip and AI battlefield

Another major turning point that continues to shape this year is China becoming a key variable in the global AI battlefield, since launching the Chinese chatbot DeepSeek under restrictions on advanced chips from US sanctions. China chose a different path, focusing on AI that emphasises “efficiency and low cost” rather than chasing Western giant models.

DeepSeek is seen as a “warning signal” to the global AI industry, after China’s model can compete in many areas at significantly lower cost. Meanwhile, US and allied restrictions on exporting advanced chips and chip technologies have pushed China to accelerate development of its own AI supply chain, especially advanced chip development.

At the same time, the Chinese government has incorporated AI as a core pillar of its economic and technology development plans, pushing real-world deployment in industry, smart cities and the public sector. Analysts see AI competition between China and the United States shifting from “who has the bigger model” to “who can deploy AI more widely and more cost-effectively.”