As the next year barrels forward, technology companies are combing through their data to get preview of the year ahead.
In a virtual media briefing last Dec. 4, Dell Technologies released its five AI-centered trends for 2025 that it believes will influence the way people work in the Asia Pacific and Japan (APJ) region and other parts of the world.
The first three trends the American technology company identified revolve around how enterprises will be applying this technology.
Dell predicts that more organizations will be integrating AI into their actual operations, the rise of agentic, autonomous AI that can work with a measure of independence, and the fusion of AI with other technologies.
The next two trends focus on how AI is impacting both governments and the workforce. Dell is watching how government approaches to AI are evolving and how AI workforce disruption is unfolding.
AI goes from proof of concept to projects with impact
In their 2024 predictions released last year, Dell already forecasted that companies would start transitioning from experimenting with AI to implementing AI technology in their operations.
As the year played out though, John Roese, Dell Technologies global chief technology officer & chief AI officer, stated that many enterprises were reluctant to invest in AI when there was no evidence of return.
That being said, Dell revealed that early adopters — including Dell themselves as well as with their biggest customers — are benefiting from a successful AI implementation, reaping savings, revenue, margin improvement, and significant outcome changes.
Moreover, Dell pointed out that the standardization of tools and the design of methodologies for a successful AI integration are minimizing the uncertainties surrounding the adoption of this technology.
Due to the learnings from these early adopters and the clearer playbook on how to adopt AI, the technology company expects to see more companies progress in their AI journeys in 2024.
Speaking for the region, Dell Technologies president of APJ & Greater China Peter Marrs, said that APJ organizations are at the forefront of this trend and as they are continuing to invest in their AI advantage.
“We couldn’t be more excited about the path ahead. We’re at the start, but we are seeing great deployments across APJ,” Marrs said.
Marrs spotlighted that AI investments in Asia Pacific are projected to reach $110 billion by 2028, showing a compound annual growth rate of 24% from 2023 to 2028.
In Southeast Asia alone, $30 billion was invested into AI infrastructure such as the AI-ready data centers in Singapore and Malaysia.
“We’re very confident that while not every enterprise will finish their AI journey next year, many many more around the world will make significant progress and we will have far larger amounts of AI projects in production than we had this year,” Roese emphasized.
The rise of autonomous AI agents
The next trend discusses the emergence of a new type of AI. While the word “agent” is being overused in the technology industry, Dell is closely watching the rise of what it terms “agentic AI” or large language models (LLMs) with a degree of autonomy that can be ordered to do tasks over time and will not require human intervention.
These agents will be specialized, have long-term memories that will aid them in understanding context, and can interact with the physical world.
Roese asserted that agentic AI are a leap forward because today’s LLMs are reactive. They need to be prompted before they carry out a task like formulate a poem or write code.
On the other hand, agentic LLMs will be given a direction and be left alone to carry it out. For instance, they can be tasked with monitoring a machine to ensure it runs properly and taking corrective action once it malfunctions, all without a person interceding.
Roese forecast that there will eventually be programming agents, editorial agents, agents that understand radiology and agents that will control machines. Due to their degree of specialization, these agents will be deployed in teams of LLMs called ensembles or even teams of mixed LLMs and human workers to carry out their tasks.
He added that the ensembles of agentic LLMs will probably be drivers of AI PCs and edge AI since Roese posits that it will make sense to run parts of the ensemble locally and others on a centralized system.
“I will make a prediction for you. Next year, the word of the year in AI will be agentic, because this is a very big deal, this is a very significant advancement in AI technology, and it will be a catalyst for significant forward progress architecturally in how we do AI,” Roese declared.
Blurring the line between AI and other technologies
For the last trend dealing directly with the applications of AI, Roese reported that AI is becoming closely intertwined with other technologies in the broader technology ecosystem, spanning from quantum computing, edge computing, and cybersecurity.
For example, Dell speculates that quantum computing and AI is developing a deep, synergistic relationship. AI is already advancing progress in quantum computing and in the future, they believe quantum computing will provide a computing architecture that will boost AI.
Dell also extrapolates that edge computing’s dominant use case will be the running of AI workloads. Roese stated that computer vision, analytics, sensing and other common uses for edge devices will be supported by AI workloads.
Finally, regarding cybersecurity, the newness of AI entails that it has yet to develop security best practices. Dell predicts that companies’ AI integrations will be built on a foundation of zero trust, the current cybersecurity principle and framework that is founded on constant verification to minimize intrusions.
Government’s developing approach to AI
For the next trends, Dell moved on to discussing people’s responses to AI’s disruption. They started with the government’s approach to AI and specifically focused on how governments are facilitating successful AI adoption in their countries.
Dell identified three patterns of facilitation. The first was governments with industry. Here, governments act as strategic organizers, ensuring that the important organizations learn and adopt AI technology successfully to ensure that their country’s industrial base does not fall behind in the global AI race. Singapore was the primary example where this pattern manifested.
The next one was governments for industry. Here, governments’ main contribution to AI adoption is in directly supplying industries such as oil, gas, and manufacturing with AI-ready infrastructure, such as data centers.
The last they labeled “government for government”. Here, governments are building data centers and AI models for their own use. Roese commented that this is a rare pattern and was seen mostly in the Middle East.
Notably, however, Dell did not touch on developing government regulations meant to promote this technology’s transparency, accountability, and sustainability as well as curb the misuse of this powerful tool.
In the Asia Pacific region alone, a majority of nations are at least in the process of designing guiding principles regarding the use of AI, if not specific measures applicable to AI that put this technology under technology-agnostic regulations.
Preparing for AI workforce disruption
To round up their AI trends for 2025, Dell revealed that this coming year is when there will be more data around the actual workforce disruption that AI is causing.
While Roese admitted that “very basic jobs, that AI can do easily repeatably, will effectively disappear”, he maintained that AI is also creating many new roles in a variety of industries. He named software composters and AI interpreters as examples.
He also pointed out that the investment in AI infrastructure will increase demand for electricians, construction workers, other manual labor roles, and even specialized jobs like thermal plumbing.
Marrs also added that there were several industries in APJ specifically that are quickly adopting AI for a variety of use cases.
Financial services and insurance sector are already using it for fraud detection and customer service, while healthcare is adopting it to facilitate diagnostics and streamline patient care documentation. Manufacturing is also applying it for smart manufacturing and digital twins, and governments are developing local language LLMs and utilizing AI to improve citizen service.
Marrs, however, acknowledged there is AI talent and resource gaps in APJ that may prevent individuals in the region from capitalizing on the new opportunities created by AI. He stressed the urgent need for AI skills development in the region.
“Looking ahead,” Marrs said, “strong human-to-AI and human-to-human collaboration will be key to driving a shared digital future.”
As he concluded the briefing, Roese said: “The only successful position a company can take is to participate, to be part of [the development of AI]. There is no option to sit on the sidelines. Because as we can see now, 2025 will have more invention and more activity in AI than 2024 did.”
