On 6 October, the World Bank raised its forecast for Malaysia’s economy to 5.1% growth this year and 4.7% in 2027. A day earlier, the ASEAN+3 Macroeconomic Research Office (AMRO) landed on the same 5.1% for 2026.
Two institutions, one number, one reason: artificial intelligence. And, alongside the upgrade, a warning from the World Bank that deserved more than the last paragraph.
1. The headline
The World Bank’s lead economist for Malaysia, Apurva Sanghi, said the new figure is 0.7 percentage points above the bank’s April forecast, and ahead of the 4.5% it expects for the wider East Asia and Pacific region.
AMRO moved from 4.9% to 5.1% for this year, and from 4.7% to 4.8% for 2027.
By any regional standard, that’s a good year.
2. The second sentence
Here is the part worth reading twice. According to Sanghi, more than 70% of Malaysia’s export growth in early 2026 came from demand for AI-related products.
Then he said this: “growth excluding AI-related goods has been weak, raising concerns about the economy’s dependence on the AI sector.”
That is an unusually blunt line to deliver at a briefing about an upgrade. A forecast raised on the back of one product category can be cut on the back of the same category.
3. How a reversal would reach us
Sanghi described two channels. The first is financial. If AI-related assets are repriced, global investors turn cautious and money leaves emerging markets, Malaysia included.
The second is trade. In the bank’s earlier analysis, a one percentage point slowdown in US growth lowers Malaysia’s growth by 0.8 percentage points.
Notice that neither channel needs anything to go wrong inside Malaysia. The risk arrives in the same containers as the boom.
4. Why it was always going to look like this
None of this is surprising if you look at where Malaysia sits: inside the electronics supply chain that AI data centres depend on. That’s exactly why the push to design its own chips by 2030 matters. Assembling someone else’s boom is not the same as owning a piece of it.
Sanghi also said Malaysia “still lags in equipping its population with strong foundational skills”. That line connects directly to graduate underemployment, where more than a third of employed graduates already work below their qualifications. An AI economy that can’t use the people it trains is a narrow one.
5. What AMRO wants done with the money
AMRO’s chief economist, Dong He, was more direct about the policy ask. What matters, he said, is “for these gains to be broadly propagated towards the economy.” He also encouraged the government to use the upcycle to review subsidy schemes and move to a more targeted structure.
Translation: while the money is coming in, fix the things that are politically hard to fix when it isn’t.
6. What to watch
Budget 2027, tabled on 9 October. The question is not whether it mentions AI. It will. The question is whether it treats the AI windfall as permanent revenue or as a window.
Spending it as though it will last is the easy option. Using it to widen the base, so that growth outside AI stops being described as “weak”, is the harder and better one.
The forecast is good news. The second sentence is the story.
Source: World Bank lifts Malaysia’s 2026 growth forecast to 5.1pc on AI investment wave, Malay Mail, 6 October 2026.
Source: AMRO upgrades Malaysia’s 2026, 2027 growth forecasts as AI investment boosts economy, The Star, 5 October 2026.