Taiwanese listed companies reported average revenue growth of 8% to 10% month-on-month in August, against a typical seasonal increase of 1% to 2%.
Equity investors spent the days before this recording trying to price a slowdown in artificial-intelligence spending, after a leading AI company published a blog post arguing for a more measured pace of frontier model development. Yap Fook Hien's answer is that the one market in the world where companies report revenue every month is not showing a slowdown at all.
"Taiwan's supply chain data has historically acted as a leading indicator for broader global end demand, and it gives us the assurance that global corporate fundamentals are intact."
Yap is a Senior Investment Strategist at Standard Chartered Bank, and what he gives here is the bank's own house view rather than a personal one — he speaks in the first person plural throughout, and ends with a specific idea the bank is putting to clients.
The full episode is covered here so you can skip it.
Here are the 5 arguments that matter.
👤 Speaker: Yap Fook Hien, Senior Investment Strategist at Standard Chartered Bank
📰 Published: 15 September 2026 on YouTube (Standard Chartered Wealth Insights)
🔴 YouTube | 🟢 Spotify | 🟣 Apple Podcasts | ⏱️ 4 min
Key Takeaways
Taiwan's August revenues were five to ten times the normal seasonal increase
8% to 10% month-on-month across many supply chains, against a typical 1% to 2%
Pacing the frontier is a change in where AI money is spent, not a cut in how much
More on safety and alignment, less on training runs, with inference demand still growing
A slower release schedule does not slow adoption, which is what inference demand tracks
Nothing in the regulatory response points to tighter restrictions, because both governments are treating it as a race
Standard Chartered's live idea is global semiconductors at mid-teens price-earnings valuations
1. Don't Overreact
Yap opened by naming what the week's price action was about, and then gave the bank's conclusion before the reasoning.
The trigger was a blog post, not a data point. Over the previous few days, he said, equity investors had been trying to understand the impact of a potential slowdown in AI spending, following a blog post from a leading AI company about pacing the frontier and advocating a measured approach to advancing frontier models. He does not name the company.
The house view is unchanged and it is a diversification call rather than a directional one. "From our perspective, investing in global tech equities continues to be attractive, and investors should diversify across both big tech and semiconductors, and not overreact to the headlines."
He then set out the structure of his answer explicitly: three ways of addressing it, taken in order.
2. Pacing Is Not Halting
The first of the three goes to what the proposal actually says, and Yap's reading is that the word "slowdown" is doing more work in the market's interpretation than in the document.
"Firstly, the proposal about pacing the frontier explicitly states that this does not mean halting model training or technical progress, but ensuring companies take adequate time to align and safeguard their models."
What follows from that, on his account, is a change in the mix of AI spending rather than a reduction in it: more of the money going to safety and alignment work, and potentially less going to training the models themselves. He put both halves as possibilities rather than as forecasts.
The demand that matters for the hardware is not affected. "The inference demand from AI models should continue to grow as adoption and use cases for AI continue to increase, even if potentially we may not get new model releases as frequently." The point is that inference tracks adoption, and adoption does not stop when the release schedule stretches.
3. No Tighter Rules Yet
The second argument is about the policy response, which he says has not gone the way a safety debate might suggest.
On the evidence so far, he said, it does not appear that there will be tighter restrictions. His explanation is that this is being treated as a race between countries rather than as a question of industry standards.
He quoted the US president directly on that point: "We're leading China in AI. We're the most sophisticated country in the world, and frankly, I want to keep it that way, because whoever wins AI wins."
The other side of the race said the same thing from the opposite direction. "Meanwhile, China's foreign ministry has dismissed the calls for a slowdown in AI development as fear-mongering."
4. Taiwan's 8-10% August
The third argument is the one carrying the actual evidence, and it comes from company disclosures rather than from commentary.
The corporate news flow since the weekend runs the other way from the slowdown story. "Updated comments from big tech companies since the weekend point to intact or accelerating AI spending plans."
Taiwan is his data source, and he explained why before giving the number: it is one of the few markets globally where listed companies report monthly revenues, a practice he said enhances transparency.
The number itself is a large deviation from the seasonal pattern. "Across many supply chains, Taiwanese companies reported average revenue growth of 8% to 10% month-on-month in August, significantly above the typical seasonal increase of 1% to 2%."
What makes that worth acting on is what it has led in the past. "Taiwan's supply chain data has historically acted as a leading indicator for broader global end demand, and it gives us the assurance that global corporate fundamentals are intact."
5. The Semiconductor Idea
He closed on positioning, and on a specific trade the bank is putting in front of clients rather than a general recommendation.
"In summary, we believe investors should not panic but stay the course to diversify their global tech investments across both big tech and semiconductors."
The idea is valuation-led. "We have an opportunistic idea now on global semiconductors, which we believe is an attractive investment as the industry is trading at mid-teens price earnings valuations, which offers an attractive risk reward in our mind."
He marked it as opportunistic rather than strategic, which is the bank's own language for a shorter-horizon idea sitting alongside the long-term allocation.
Bonus Insights
He never names the company behind the blog post, referring to it only as a leading AI company, even while quoting its argument closely enough to describe what it explicitly states.
The two political quotes are placed as a matched pair — Washington saying it wants to stay ahead, Beijing calling the slowdown argument fear-mongering — and are used to support a single conclusion about regulation rather than as commentary on either government.
The episode is dated on air: Yap opens by saying it is Tuesday, 15 September, which is how these daily items carry their own timestamp.
His case rests on a sequencing claim that he does not dwell on: that the monthly Taiwanese revenue figures are early enough in the chain to say something about global demand before the companies at the end of it report.
Yap's bottom line is that the pacing proposal changes where AI money is spent rather than how much of it there is, that the policy response so far points to a race rather than to restrictions, and that Taiwan's August revenue data says global demand is intact — so the position to hold is a diversified one across big technology and semiconductors, with an opportunistic idea in semiconductors at a mid-teens multiple.
Books & Resources Mentioned
The blog post on pacing the frontier (The proposal the whole episode responds to; Yap uses its title but does not name the company that published it)
Taiwan's monthly revenue disclosures (One of the few markets globally where listed companies report revenue every month, which is why he uses it as the leading indicator)
Watch the full episode:
If this was worth your time, send it to someone closer to the industry than you are.
Get the latest market chatter as it happens: