The global equity firm has cautioned crowded AI trades mask rising risk, and has argued that overlooked companies with resilient cash flows could drive 2026 rotation.
The investment landscape heading into 2026 is marked by contradiction, according to leading global equity manager Ninety One, with weak consumer data sitting alongside markets bidding up highly cyclical AI infrastructure suppliers as defensive assets, while some of the most reliable recurring-revenue businesses in corporate history trade at a discount.
Head of quality Clyde Rossouw said the greatest risk lies in the crowded trade.
“The conviction that the AI infrastructure wave will never crest. When the inevitable air pocket in capex arrives, the market will abruptly remember the cyclical nature of these businesses and high starting multiples will offer little protection.”
Opportunity, Rossouw argued, lies in mispriced resilience rather than momentum.
“True defence is not found in chasing momentum into cyclical extremes but in identifying where resilient cash flows are being offered at a discount because they lack a positive AI narrative.
“When the market re-focuses on fundamentals, we would expect some of the perceived ‘AI losers’ to come to the fore.”
After a year in which AI-linked mega-cap stocks were treated as “defensive” safe havens, the global equity manager has warned that markets may be dangerously mispricing risk heading into 2026, setting the stage for a sharp rotation away from crowded AI trades and toward overlooked companies with resilient cash flows and strong balance sheets.
Ninety One said market concentration, stretched valuations and heavy reliance on a narrow group of AI infrastructure spenders have left investors exposed.
“History shows that when cyclicals are priced as defensives and defensives as cyclicals, a great rotation tends not to be too far away,” Rossouw stated.
“True defence is not found in chasing momentum but in identifying companies with resilient cash flows, strong balance sheets and the ability to compound through different economic and market environments.”
Rossouw pointed to 2025 as a clear preview of how quickly market leadership can change, as political uncertainty, trade shocks and, later, valuation concerns around AI triggered abrupt rotations, during which stocks with resilient earnings and strong balance sheets held up far better than momentum-driven names.
These episodes, he suggested, have underscored the growing importance of fundamental quality as markets enter a more volatile and less forgiving phase in 2026.
“These kinds of dislocations are precisely where a fundamental quality approach can show its defensive characteristics,” Rossouw said. “Our purist approach is built around companies that can weather earnings shocks.”
While AI’s popularity is undeniably real, with the underlying technology potentially proving more transformative than the early internet years, Ninety One said the current level of spending may have pulled forward years of future earnings and demand for infrastructure suppliers.
“There will inevitably be digestion periods, inventory corrections, and a temporary halt or slowdown in the pace of the buildout. Businesses that rely on massive, coordinated capex plans from a handful of the world’s largest companies are, by definition, cyclical,” Rossouw said.
The market largely shrugged off this cyclicality in 2025, pushing valuations of AI-exposed companies to extremes. Investors should be mindful of this concentration risk, according to the firm.
Rossouw warned: “History shows that capex cycles can reverse quickly, and a level of caution is warranted. If sentiment changes, the fallout could be significant. This is a moment for valuation discipline and fundamental analysis, not blind faith in a single narrative or factor exposure.”
High-quality businesses in software, IT services and information services were caught in the cross-currents of 2025, Rossouw said, punished by fears that AI would erode pricing power or displace their products.
Many now trade at or below market-level valuations despite recurring revenues, high switching costs and asset-light business models.
“The irony is stark. Only a few years ago, these companies were expected to be the primary monetisers of AI. Fundamentals haven’t changed; sentiment has.
“For long-term investors, this creates the potential for attractive compounding from businesses that can grow earnings steadily, without relying on aggressive assumptions about the pace of AI adoption,” Rossouw said.
Ninety One also stated that the market may be underestimating AI’s potential impact on pharmaceutical innovation.
If AI improves R&D productivity and accelerates drug discovery, major pharmaceutical companies could generate far better outcomes for every dollar spent, the firm suggested.
Yet this opportunity is not reflected in the valuation of many stocks in the sector, which have fallen out of favour amid the view that they are not technologically driven.
Rossouw said traditional defensives were caught between a rock and a hard place in 2025 in a momentum-driven market, lacking the excitement of AI exposure while remaining exposed to squeezed consumers.
“But in a more uncertain macro environment with a less dominant AI-theme, their defensive attributes should come back into favour.”
