DBS says the artificial intelligence trade is entering a more discriminating phase, as investors shift from broad enthusiasm for AI infrastructure toward companies that can translate adoption into hard financial outcomes. The bank's view reflects a market that is beginning to separate the builders of the AI stack from the businesses using AI to improve efficiency, lift margins and generate recurring growth.
Selective AI Phase
DBS argues that the next leg of the AI rally is likely to be led less by companies making the largest capital commitments and more by those able to extract operating leverage from existing systems. In practical terms, that means investors are increasingly looking for evidence that AI is not just a strategic narrative, but a driver of productivity, cost discipline and stronger earnings quality.
The shift matters because the first wave of AI enthusiasm was dominated by hardware suppliers, cloud providers and infrastructure-linked names that benefited from the rush to build capacity. That trade remains relevant, but DBS suggests it is no longer sufficient on its own. As the market matures, valuation support is likely to depend more heavily on whether AI spending converts into revenue acceleration or margin improvement rather than simply higher capex.
This is a familiar pattern in technology cycles: the market initially rewards the enablers of a new platform, then gradually moves toward the adopters that can monetise it most effectively. DBS appears to believe AI is now approaching that second stage. For investors, that implies a more selective approach, with greater scrutiny of balance sheets, return on investment and the pace at which AI tools are embedded into core business processes.
Lower Capex, Higher Return
DBS is favouring lower-capex AI adopters, a stance that reflects a preference for businesses that can participate in the AI economy without taking on the full cost burden of building large-scale infrastructure. These companies may be better positioned to preserve free cash flow while still benefiting from AI-driven efficiency gains.
That preference is especially relevant in a market environment where capital intensity is under renewed scrutiny. Firms that must spend heavily on data centres, chips, power and networking to stay competitive may still enjoy strong demand, but their earnings conversion can be slower and more volatile. By contrast, companies that integrate AI into software, services, workflows or customer engagement can often scale faster with less upfront investment.
DBS's framing suggests investors should pay close attention to the quality of AI exposure, not just the size of it. A company that uses AI to automate repetitive tasks, improve forecasting, sharpen pricing or enhance customer retention may deliver a more durable earnings uplift than one that is simply spending aggressively to keep pace with the industry.
At the same time, DBS is not abandoning the infrastructure theme altogether. The bank continues to maintain exposure to infrastructure enablers, signalling that the physical and digital backbone of AI remains a critical part of the investment case. That includes the ecosystem supporting compute, storage, connectivity and power demand, all of which remain essential as AI workloads expand.
Market Tests Ahead
The broader implication is that AI investors may now face a more demanding market test. The easy phase of buying anything linked to AI infrastructure may be giving way to a period in which earnings delivery matters more than thematic exposure. That could widen the performance gap between companies that can prove AI-led operating gains and those whose business models depend on continued capital spending by others.
For macro investors, the DBS view also speaks to a wider debate about the productivity effects of AI across the economy. If adoption begins to show up in stronger margins and better revenue growth, the technology could become a meaningful support for corporate profitability even in a slower growth environment. If not, the trade may remain concentrated in a narrow set of infrastructure beneficiaries.
In that sense, DBS is signalling a more nuanced AI market: still constructive, but less indiscriminate. The bank's preference for lower-capex adopters suggests the next winners may be those that use AI to do more with less, rather than those that spend the most to build the future from scratch.