Taiwan's overall PMIs look fine. The details are shakier, but so far, it looks like the weakness is coming from softer demand for consumer tech as semi prices boost the cost of electronics devices. The data don't (yet) show obvious deterioration in the data centre component of the AI boom.

The unusually long section of analytical boxes in the BOJ's main outlook report shows that the bank has a lot on its mind. That inevitably brings nuance. However, the overall tone is clear: nominal growth remains strong, and financial conditions loose. There's no clear reason not to hike further

China's nominal cycle looks to be slowing again. The rest of the region remains strong, but confidence that this strength can be retained is fraying, and will likely remain under pressure, especially if concerns over the data centre build-out and energy prices worsen again after the mid-terms.

Headline official PMIs continue to closely hug the 50 threshold. I am not sure that tells us much about the economy. But two components of the PMI are useful. The ongoing slump in construction indicates no property revival. Price indicators show PPI still strong, but warn of some weakening in CPI.

Rather than slowing as the government had forecast just a couple of months ago, GDP growth re-accelerated in Q3 to 14.4% YoY. That was because of domestic demand rather than net exports, and while that was mostly capex, consumption is growing more than 5%, and the positive output gap widened.

The BOJ argues today not just that "underlying inflation" has reached 2%, but that it is expected to stay there. However, it seems clear that stability won't be achieved without some help, as its analysis also says that all the risks to prices are to the upside.

Earnings releases and September macro data show the cycle has strong momentum. Growth is dominated by semis, and so dependent on the global cycle. However, the doubling in tax revenue in September shows one way in which the benefits of the semi cycle can start to broaden out.

With both growth and inflation strengthening, it will be tough for the BOJ not to hike faster – whatever Takaichi thinks. However, back-to-back or 50bp moves are unlikely. Moreover, the peaking of property prices and sluggishness in some wage indicators point to MT headwinds.

September money and credit data for September are like other indicators for Taiwan: quite strong, but not high enough on their own to think that they suddenly break the CBC's apparently strong desire not to change policy.

Today's Q3 GDP data continue the same story: real growth is solid, but more importantly, the huge windfall income gain from the surge in semiconductor prices remains a major tailwind for growth in 2027. The question remains how much of that income windfall will be spent.

The sustainability of the AI cycle is critical for the region. If it ends, rates will likely be lower, but currencies not necessarily weaker. Otherwise, my themes are: China – tougher 2027; Japan – flatter curve; Korea – macro spillover; Taiwan – market spillover.

Like so much else in Taiwan, labour market dynamics will be determined by the fortunes of the AI cycle. If that turns down, it will be hugely important. A sharp downturn hasn't, however, been the CBC's expectation, but existing wage trends seem inconsistent with the bank's view that inflation slows.