Market View: The Trillion-Dollar AI Spending Boom: Who Wins, Who Gets Disrupted?

SPEAKER_1 (0:01)

Best of Live Well with Michelle.

Explore what it means to live better with Live Well with Michelle. From mindset and personal growth, to money and everyday living, through thoughtful conversations and practical insights, feel balanced and in control. Live Well with Michelle, every day.

Michelle Martin (0:21)

Alrighty, let's get through markets, shall we? First up, I want to say thank you to the wonderful people at YCH, who had me at Supply Chain City yesterday for the YCH Group Supply Chain Summit 2026 I had an absolutely fabulous time as your host and moderator. There is something about the YCH spirit, I have to say, people so intelligent, so warm, as we discussed smart supply chains it's like being in Silicon Valley of Singapore, you know, that's what I felt there. Wonderful! It was absolutely the highlight of my day being at Supply Chain City. I had never even been in the building. This is a nerve center here in Singapore. They're a made in Singapore company, homegrown. Yeah, more than 70 years old. Incredible. It was such a privilege and honor to meet everybody at Supply Chain City and everybody at YCH. Thank you again for having me. And it was, you know, I have to say, what a day of discussion. Deep, thoughtful, great perspectives. I really enjoyed myself. All right. On to Market View. And today I'm following the money in AI because the numbers are becoming extraordinary.

Let's start with, what should we start with? Alibaba. I want to start with Alibaba. Because it's unveiled a sweeping strategy encompassing AI models, chips, data centers. Its next-generation QWEN models could scale between 5 and 10 trillion parameters. Imagine that. Compared with 2.4 trillion that its current flagship model can handle. Alibaba has also unveiled its next-generation Chenwu V900 AI chip, which it says delivers three times the performance of its predecessor. And then there is the physical infrastructure. Alibaba Cloud wants its global data center capacity to exceed 20 gigawatts by 2032 So is Alibaba looking like China's answer to the vertically integrated AI giants in the US? Does owning more of the AI stack create a competitive advantage, or does this mean Alibaba has committed itself to an enormously expensive arms race? I think a lot of people are going to be looking closely at Alibaba's new strategy. We're already seeing their shares respond to that. Shares jumped around 3%, at least in Hong Kong yesterday, after the Alibaba, the tech giant, unveiled this new AI chip and plans to sharply expand its data center footprint. Alibaba stepping up its bed on infrastructure, powering the AI boom at the Cloud Units' annual flagship Apsara conference in Hangzhou. They introduced the next generation AI chip. It's called Chenwu V900. That's the one I've been talking about. The company also has plans to expand their Alibaba Cloud Global Data Center capacity to more than 20 gigawatts by 2032 It's all part of this roadmap, spanning Chips Cloud and AI models.

Next, let's look at AI spending, shall we? Jamie Dimon says that hyperscaler spending could hit one trillion next year. What does this mean? What does this mean? That the arms race is becoming breathtakingly expensive, right? You have a Jamie Morgan CEO saying, spending across this hyperscaler ecosystem has already risen from about 300 billion to 700 billion. It could reach one trillion next year. Now, let's think about that number one trillion. A trillion going into chips, data centers, power, networking, and the infrastructure needed to build AI. At what point does the question for investors become, who is spending the most? I think we've already been there, actually. Maybe the question will shift to, who can actually earn an adequate return on all this money? I mean, we're talking trillions. You know how many zeros that is? Could this really become a trillion-dollar annual investment cycle? Then where are the picks and shovels winners? You know?

All right.

From the hyperscalers spending, I want to learn who could benefit from increased spending here in Singapore. There's a Singapore angle to all this AI spend, which by the way, I covered this morning, as I looked at all this AI spend and what it means for bond market. That's a very good show. You should listen to it for sure if you're interested in bonds. I had a fabulous time with my guest, Tu Ha Chow, Head of Fixed Income Asia at Robeco.

All right. Back to what I'm talking about, the AI spend and the Singapore angle. UOB, a local bank, sees a niche financing the growth of AI infrastructure across Southeast Asia. The bank was recently part of a syndicate providing a 3.1 billion GPU credit facility to the AI infrastructure platform Zancor alongside Citi ING, Texas and Qatar National Bank. So could the AI infrastructure become a new meaningful lending business for Singapore banks? There's the risk question as well. Financing GPUs, financing data centers, not the same thing as financing a factory or an office tower. The technology can become obsolete extremely quickly, which is what I put to my panelists yesterday. So if you're lending billions against AI infrastructure, how then do you price technological obsolescence?