Tan Su Shan on DBS AI Financing and Risk Repricing
DBS CEO Tan Su Shan says the bank has originated about 25 billion in AI-related financing year to date, but has kept only a fraction. She says tighter market…
DBS CEO Tan Su Shan says the market is repricing risks around artificial intelligence investment, a shift that could bring more discipline to financing. DBS has originated about 25 billion in AI-related deals year to date, though it has retained only a fraction.
The lending matters because AI projects draw financing across the hardware supply chain, from semiconductor production to data centers. Tan said DBS arranges and syndicates many of those loans rather than holding the full exposure itself.
“We are often the lead manager for a lot of the syndicated loans,” Tan said in an interview with Haidi Stroud-Watts at the DBS Asia Future Forward Conference in Singapore. She added that the bank originates the deals and syndicates them out.
DBS AI financing spans the hardware chain
Tan described demand for memory chips and upstream suppliers, with financing extending through connectors and cooling systems to data centers. She said the activity is mostly focused on hardware, while also reaching traders.
That breadth gives lenders several points to assess when considering exposure. Tan said DBS examines whether arrangements are triple net, who will take out the financing and whether a backstop exists. She cited a large semiconductor company backing its chips as one possible form of support.
“It really is heavily structured right now,” she said.
For the bank’s customers and investors, the scrutiny affects how large AI projects may be funded and what protections accompany loans. If financing terms reflect the risk more closely, investors will also be looking for evidence that substantial spending across the AI capital expenditure ecosystem can produce returns capable of repaying debt, bonds or equity holders.
Market terms are shifting
Tan said rates had risen over the previous two weeks and that bond prices for some hyperscalers had reacted. She described the market response as a prompt for investors to seek higher returns and examine financing structures more closely.
“The market has spoken and it’s not a bad thing,” Tan said. In her view, tighter scrutiny can test whether the large financial commitments flowing into AI investment are financially sustainable.
Some technology companies that once relied on cash flow or equity issuance are also tapping bond markets, the interviewer noted. Tan said data center deals in the United States are being packaged into asset-backed securities as they mature, securitized and sold in portions.
She also pointed to hyperscalers accessing bond and equity markets, and said some large language model companies are tapping the IPO market. Tan warned that weaker players could face credit rating downgrades, without naming any companies.
The changing mix of financing puts the focus on whether project structures and expected returns can support the obligations attached to them. DBS’s approach, as Tan described it, is to originate deals, distribute much of the exposure and assess the protections built into each loan.
“The market wants to ensure that these big financial commitments are actually financially sustainable,” Tan said.
Source: finance.yahoo.com
