UOB sees niche in financing AI growth

[SINGAPORE] As the demand to finance infrastructure for artificial intelligence grows in South-east Asia, UOB is looking to develop in this area, given its experience of financing the sector.

The local lender was recently part of a syndicate for a US$3.1 billion graphics processing unit (GPU) credit facility for ZanKore, along with Citigroup, ING, Natixis CIB and Qatar National Bank.

ZanKore is an AI infrastructure platform launched by Indosat Ooredoo Hutchinson – a major Indonesian mobile telecommunications and digital services provider – with Nvidia, Ooredoo and Nokia as partners.

UOB was a senior mandated lead arranger, underwriter and book runner for the deal, which it said was so far one of the largest in South-east Asia.

Frederick Chin, head of wholesale banking and markets at UOB, said: “We stepped up to advise the structure and connect some of the dots.”

As ZanKore is committed to doubling its data centre capacity to 200 megawatts (MW) by the first half of 2027, it is looking to replicate this syndication and secure more financing for the build-up.

The platform ultimately is looking to build its capacity to 1 gigawatt (GW), which would require funding of about US$70 billion, said Vikram Sinha, CEO of Indosat, ZanKore’s majority shareholder.

“In about six months, we are looking at another US$6 billion, and we are starting a new syndicate in which UOB will be our lead adviser, because it understands this sector and region very well,” he added.

This is not the first GPU financing transaction UOB has been in. A smaller bilateral deal was done earlier, said Chin, but he did not disclose the dollar amount for that deal.

“The intent is to get to market and educate other players to come and participate, understand the structure and risk. I think we are quite confident that this deal will set the benchmark and start creating liquidity for more deals to come,” he said.

Banks’ edge in financing GPUs in S-E Asia

Banks have an edge in South-east Asia in financing GPUs, as private credit is “non-existent” in the region, he said. In the US, private credit players have been the main financiers of the AI build-out.

The local lender has been financing parts of the AI ecosystem, mostly in the area of infrastructure, covering power assets, data centres and now, GPUs.

This is referred to as the third of five layers of the AI universe, a concept initiated by Nvidia CEO Jensen Huang. The first layer is power, the second is chips, the third is infrastructure, the fourth is models, and the fifth, applications.

Chin estimates that the current power constraints amid demand for GPUs and servers in South-east Asia could result in US$20 billion worth of financing demand.

“That’s the space that we are focusing on,” he added.

UOB is looking to deepen its expertise in the fourth and fifth layers of the AI ecosystem before it starts underwriting that sector.

The bank is tapping into its technology, media and telecommunications sector specialist group, as well as its engagement with tech companies and clients involved in the field, and is “truly plugged in”, said Chin.

Currently, UOB’s exposure to AI financing accounts for under 5 per cent of its wholesale banking loan book, which stood at S$258 billion as at Jun 30, 2026.

Upcoming catalysts such as the Asean Digital Economy Framework Agreement (DEFA) that could stir up more financing demand in South-east Asia.

The agreement is estimated to double the potential of Asean’s digital economy, from US$1 trillion to US$2 trillion by 2030.

Chin said: “With that kind of number, you can expect demand for AI’s services to only increase.”