DBS CEO Tan Su Shan: Market Is Repricing the Risks of Financing the AI Buildout

On 12 October 2026, Tan Su Shan, CEO of DBS Group, said in an interview with Bloomberg that the market is repricing the risks around investments in AI infrastructure. In her assessment, this could lead to greater discipline in financing such projects. The conversation with Heidi Stroud-Watts took place at the DBS Asia Future Forward conference in Singapore. The published excerpt contains no specific amounts, deal names or rate thresholds. But the wording itself deserves attention, because it comes from the head of the largest bank in Southeast Asia by assets.

What exactly was said

The thesis is short: the market is reassessing risk. This is not a forecast of a crash, nor a call to halt construction. The banker describes a change of regime in which money for AI infrastructure stops being nearly automatic. The key word here is "discipline." It means tougher selection of borrowers, closer scrutiny of deal structures, and a harder look at where the revenue that will cover debt service is going to come from.

What the statement leaves out matters too. As far as the excerpt shows, Tan Su Shan does not name specific companies or segments, and does not give DBS's own lending volumes in AI. So her remarks are best read as an assessment of sentiment in the capital markets, not as an announcement of a change in DBS policy.

Why a lender's words carry more weight

A venture investor risks its own capital and can afford a long-horizon bet. A bank lends against a payment schedule. When it talks about reassessing risk, it is talking about borrowing costs, covenants and its willingness to join the next big deal.

The AI buildout is capital-intensive: data centers, power supply, cooling, accelerator purchases. A significant share of these projects is financed with debt, often on the strength of long-term contracts with tenants of the capacity. When lenders reassess risk, the first projects to feel it are those whose revenue is not yet backed by contracts and whose equipment becomes obsolete faster than the loan is paid down. Singapore remains one of the region's data center hubs, and DBS sits at the center of the local financial ecosystem, so for the bank this is not an abstract discussion.

Key parameters of the signal

• Who is speaking: Tan Su Shan, CEO of DBS Group, the largest bank in Southeast Asia by assets.
• Where and when: DBS Asia Future Forward conference, Singapore, 12 October 2026.
• Main thesis: the market is repricing the risks of investments tied to the buildout of AI capacity.
• Possible consequence: more disciplined financing.
• What the source does not contain: figures on loan volumes, rates, specific borrowers and timelines.

What companies and investors should do

For companies that build or lease AI capacity, the practical takeaway is simple: prepare your financial model for a more skeptical reading. A lender needs confirmed contracts for capacity utilization, a clear power plan and a realistic equipment lifespan. Projects built on demand expectations will have a harder time than projects with an anchor tenant.

Investors should watch for signs that discipline is turning into numbers: wider spreads on AI infrastructure debt, renegotiated terms in syndicated loans, and more frequent bank requests for collateral or guarantees. For now this is a signal, not a fact. But if other large lenders follow the DBS chief's words with similar assessments, the adjustment in the cost of capital for AI projects will become noticeable, and those most dependent on borrowed money will feel it first.