REPORTED: remarks made at a DBS conference in Singapore on 12 October 2026, as reported by CNBC and Reuters. TSN has not seen a transcript or video.
The heads of two of the world’s biggest chip companies told a conference in Singapore on Monday that the thing most likely to hold back the AI boom may not be chips at all, but money.
“[What] I think might be the biggest constraint is capital,” Broadcom chief executive Hock Tan said at the DBS Asia Future Forward Conference, according to CNBC [1]. CNBC reported that he said financing could become a bigger obstacle than shortages of semiconductors, memory chips or electricity [1].
Intel chief executive Lip-Bu Tan (no relation) made a similar point. “If there’s any challenge in the front, it’s capital and the cost of capital, and then the other part is in the longer run, people want to see the ROI,” he said, CNBC reported [1]. ROI means return on investment: whether the huge sums spent on AI eventually earn money back.
Why capital matters
Building AI needs data centres, advanced chips and power supplies, and companies are borrowing heavily to pay for them. When interest rates and bond yields rise, that borrowing gets more expensive, and projects that may take years to pay off look less attractive. That is the worry both men were pointing at.
Shortages that come and go
Hock Tan said physical bottlenecks keep appearing but get solved. “[In the] last two years we have seen a series of bottlenecks that keep coming out, but then we overcome it,” he said. “Memory shortage today might not be that way a year from now” [1].
According to CNBC, he suggested shortages of wafers (the silicon discs chips are made from) could ease by 2028 or 2029, and that power availability could also improve [1].
Demand still ahead of supply
Neither executive predicted a slump. “There is more demand for compute infrastructure than we can supply,” Hock Tan said [1]. CNBC reported that he added this imbalance could persist through 2029 and potentially into 2030 [1].
CNBC also reported that Lip-Bu Tan said the industry was unlikely to see a significant slowdown over the next two to three years [1]. Beyond that, CNBC reported, both said growth depends on how quickly businesses adopt AI and whether it delivers real economic benefits [1].
Chinese AI models: cheap and “good enough”
Reuters reported separately from the same DBS panel that Lip-Bu Tan said the token costs of Chinese frontier AI models have been lower and adoption has been “very fast” [2]. Tokens are the small chunks of text AI models process; companies pay per token, so lower token costs mean cheaper AI.
According to Reuters, he said the performance of Chinese models such as DeepSeek, Kimi and GLM is “good enough” [2]. “I’m a strong believer of open source, that’s the best way to drive efficiency,” he said, Reuters reported, adding that markets will decide which models suit which uses [2]. Open source here means models whose workings are published for anyone to use and adapt, as many Chinese models are.
Reuters also reported his forecast that quantum computing could become mainstream in the next three to five years [2]. That is his prediction, not an Intel product timeline.
What is not known
- No transcript or video of the panel has been checked; all quotes come via CNBC and Reuters.
- Neither executive gave figures for how much financing is needed or where it might fall short.
Sources
- CNBC (Lee Ying Shan), “Intel, Broadcom CEOs warn capital crunch could become AI’s biggest bottleneck”, 12 October 2026, 09:27 BST. https://www.cnbc.com/2026/10/12/intel-broadcom-ceos-warn-capital-crunch-could-become-ai-bottleneck.html
- Reuters (Jun Yuan Yong), “Intel CEO says Chinese frontier AI model adoption has been ‘very fast'”, 12 October 2026, via Global Banking & Finance Review. https://www.globalbankingandfinance.com/intel-ceo-chinese-frontier-ai-model-adoption-fast/
Related stories
- Broadcom Reportedly Arranging $50bn+ for OpenAI’s Custom Chip, as Oracle Seeks Its Own Chip Financing
- Volta Launches a $5.012bn Six-Year Loan, PitchBook Reports; It Is Launch-Stage Pricing, Not a Closed Deal
- Bain’s Bridge Data Centres Is Reportedly Trying to More Than Double a Loan to About US$6.6bn; a Separate Thailand Loan Has Stalled
