Thursday 19th February 2026
AI’s winners and losers: why advisers should look beyond the mega-caps
While AI is a core theme driving equity performance, not all tech companies will rise with this tide. Examine how the disruption of AI is creating a world of winners and losers.
Global equity markets have delivered solid returns over the past year, buoyed by artificial intelligence enthusiasm. But beneath the surface, the AI trade is creating sharp divides. Not just between sectors, but between business models.
For advisers constructing portfolios in 2026, the key question is no longer whether AI matters. It is who ultimately benefits and who gets disrupted.
The concentration risk beneath the rally
Over the 12 months to 31 January 2026, global equities rose by 8.5 per cent, supported by favourable economic conditions and continued investor enthusiasm for AI-linked growth. Yet market breadth has remained narrow.
In the US, the S&P 500 Equal Weight Index underperformed its market-cap weighted counterpart by nearly five percentage points. The ten largest US companies now account for roughly 40 per cent of total market capitalisation.
“This concentration of returns has largely been a function of the unrelenting AI investment cycle. A relatively small number of mega-cap stocks have posted spectacular gains as their AI cloud businesses accelerate and drive revenues.”
Nick Markiewicz, Ellerston Capital
For advisers, this presents an increasingly familiar dilemma: should they remain exposed to dominant AI infrastructure beneficiaries; or position for a potential broadening of returns?
The AI “winners”
Nick Markiewicz, portfolio manager of the Ellerston Global Mid Small Cap Fund, observes that a widening bifurcation has emerged between AI “winners” and “losers.”
On the winning side are companies with proprietary AI intellectual property or those leveraged to surging AI infrastructure spending. US hyperscalers have sharply lifted 2026 capital expenditure forecasts, with annual capex now expected to exceed US$600 billion ($845 billion), up from US$450 billion ($630 billion) only months earlier.
“These companies still have innovative financing options available should they wish to push investment even harder through securitisation structures and bond issuance. Alphabet recently raised US$20 billion in new bonds, attracting more than US$100 billion in orders” observes Markiewicz.
A large share of this capital is flowing into graphics processing units (GPUs) and data centre capacity.
“These companies all have one thing in common, a ready access to power,” Markiewicz says. “That remains in acute shortage and difficult to bring online quickly.”
For advisers, this reinforces the importance of understanding not just AI exposure, but infrastructure constraints, particularly around energy and data centre scalability.
The AI “losers”
While infrastructure beneficiaries have thrived, parts of the software ecosystem are under pressure.
Recent updates from Anthropic and OpenAI have significantly improved agentic coding capabilities, enabling autonomous AI agents to build software and connect directly with enterprise systems. The implications for traditional SaaS models have been immediate.
The Morgan Stanley US SaaS Index has fallen nearly 30% since December.
“While we are yet to form a confident view of the lasting impacts of agentic coding, the sell-off is a reminder of how quickly AI is developing. The market is aggressively re-pricing perceived AI losers in a shoot first, ask questions later manner,” Markiewicz says. (Unlike generative AI, which primarily creates content, agentic AI acts as an agent to achieve specific objectives by navigating complex environments, using tools and adapting to feedback.)
For advisers, this highlights a new risk: valuation assumptions built on durable software moats may no longer be as robust as previously assumed.
A potential valuation inversion?
The disruption may go further than software price/earnings (P/E) multiples.
“We are also questioning whether markets may begin placing greater valuation premiums on businesses with difficult-to-replace tangible assets, which is a potential inversion of the past two decades, during which intangible-heavy firms have commanded expanding multiples,” says Markiewicz.
If this thesis plays out, advisers may need to reconsider long-standing growth allocations, particularly those skewed toward asset-light, intangible-heavy models.
The Rotation Beyond AI
Importantly, the market narrative is not solely about AI.
Since the start of 2026, the Russell 2000 (which tracks the US market’s ‘small-caps’) has begun to outperform both the S&P 500 and the so-called “Magnificent Seven,” potentially signalling a shift in leadership.
“While still early, we believe this reflects a combination of improving cyclical expectations and growing investor caution around the declining quality of mega-cap cash flows as investment spending accelerates, or the beginning a rotation into businesses with less perceived AI revenue risk,” Markiewicz notes.
For advisers, this may represent a classic broadening phase; one where smaller and mid-cap exposures begin to participate more meaningfully in returns.
The macro backdrop: Uncertainty remains
Despite the potential rotation, risks remain elevated.
“Despite strong trailing returns, the outlook for global equities through the remainder of FY26 remains highly uncertain. Trump’s policy direction has been described as erratic, equity markets are trading at record highs and investors remain unforgiving of earnings disappointments,” says Markiewicz. “Offsetting these risks, inflation remains muted, industrial activity appears to be improving, and further rate cuts could provide support for short-cycle sectors.”
For advisers and their clients, the message is clear: AI remains a powerful force reshaping markets. But the easy narrative of “own the mega-caps” is increasingly incomplete.
Laurence Parker-Brown is head of content at The Inside Network and managing editor of The Inside Adviser and Investor Strategy News.
