The Magnificent Seven: Old wineskin for new wine
Investors must look beyond these stocks to find companies positioned to lead in the race for agentic AI
WHEN ChatGPT burst onto the scene in November 2022, it redefined our expectations for what artificial intelligence could achieve. For the next two years, the so-called “Magnificent Seven” tech giants (Alphabet, Amazon, Meta, Apple, Microsoft, Nvidia and Tesla) became synonymous with the AI trade, dominating headlines and portfolios alike.
But beneath the surface, the AI landscape has been quietly shifting. The next breakthrough – agentic AI – will require investors to rethink the playbook.
The next “ChatGPT moment” is near
The first phase of AI was defined by conversation. Large language models (LLMs) such as ChatGPT dazzled the world with their ability to generate human-like responses. But these systems were passive, waiting for prompts and then replying.
Now, we are on the cusp of a new era: agency. Agentic AI systems won’t just answer questions; they will act, plan, reason and execute complex tasks with minimal human intervention.
This is more than an incremental improvement, it’s a paradigm shift. Agentic AI blurs the line between tool and collaborator, empowering machines to anticipate needs, coordinate workflows and adapt to changing environments. The implications for productivity, security and creativity are profound.
Already, agentic AI workflows are driving a surge in computational demand, with estimates suggesting a five to 15 times increase in token consumption compared to traditional LLMs. The coming year is set to be pivotal, as leading players such as OpenAI and Anthropic pivot towards enterprise deployment, seeking to secure predictable income streams and accelerate adoption.
If the first “ChatGPT moment” was obvious, a chatbot on our phones, the next may be quieter but no less transformative. The arrival of agentic AI will not be marked by a single product launch or viral demo, but by a gradual, profound shift: systems that move from simply providing answers to autonomously delivering outcomes.
We may only recognise this new era in hindsight, as AI begins to anticipate our needs, complete tasks and reshape the way we work and live.
Infrastructure race for agents
However, building agentic AI isn’t just about smarter algorithms; it’s about infrastructure. The bottleneck is shifting from raw compute to the orchestration and memory layers that allow AI systems to operate continuously and autonomously. Data centres are evolving from collections of chips to fully integrated systems, where central processing units, graphics processing units and memory work in concert to support persistent, context-aware agents.
The numbers are staggering. AI spending is forecast to reach US$1.3 trillion by 2030, growing at a 25 per cent compound annual growth rate (CAGR), while the broader AI total addressable market is expected to hit US$3.1 trillion (30 per cent CAGR).
Global data centre capital expenditure is projected to total US$7 trillion between 2025 and 2030. Yet, even as investment surges, bottlenecks remain. Memory chip shortages are expected to persist well into 2027.
Asia is key to agentic AI revolution
Asia stands at the epicentre of the agentic AI revolution, its influence both indispensable and transformative as the world races towards more autonomous, persistent and context-aware systems.
The region’s technological prowess is unmistakable: Taiwan anchors the global supply chain with its cutting-edge semiconductor manufacturing, providing the essential hardware that powers advanced AI.
South Korea, meanwhile, has cemented its dominance in memory technology, evidenced by a remarkable 180 per cent year-on-year surge in semiconductor exports, which has propelled the Kospi index to record highs.
Japan’s expertise in precision components ensures that as AI systems become more complex, reliability is never compromised.
China’s vast scale is reshaping the global digital landscape, with its total server capital expenditure projected to grow at a 31 per cent annual rate through 2027, and AI server investments accelerating even faster at 47 per cent.
Not to be overlooked, South-east Asia is rapidly establishing itself as a critical hub for data centres, while India’s enormous, digitally savvy population is driving a new wave of applications and demand.
The numbers tell a compelling story: AI revenues in Asia are set to soar 15-fold, from US$28 billion in 2022 to an astonishing US$420 billion by 2027. Infrastructure spending alone is expected to rise at a blistering 50 per cent annual rate, reaching US$195 billion within the same period.
Looking beyond the familiar
The Magnificent Seven remain important AI companies. Alongside Micron and Broadcom, they are on pace to drive two-thirds of S&P 500 earnings growth in 2026, reflecting blockbuster capex plans by hyperscalers. Yet, these tech giants, once viewed as the main AI trade, are increasingly an old wineskin for new wine in today’s AI-driven market.
As the AI landscape shifts from passive models to autonomous, orchestrated systems, investors must look beyond the familiar Mag Seven label and assess a broader range of companies positioned to lead in the race for agentic AI.
Relying on yesterday’s winners risks missing tomorrow’s disruptors and the new sources of value emerging across the global technology stack. Diversified exposure across the AI value chain and geographies is now essential, as leadership in hardware and infrastructure rotates and new sources of value emerge.
The writer is managing director and chief investment officer for Southern Asia and Australia, UBS Global Wealth Management
