Wall Street’s technology sell-off has raised an uncomfortable question — Is the artificial intelligence (AI) boom approaching a financial reckoning?
For Malaysia, the consequences extend beyond share prices. The country is increasingly integrated into global AI supply chains through semiconductors, electronics and data-centre investments. If the boom falters, Malaysia could face weaker exports, delayed investments and underutilised infrastructure.
The central risk is not necessarily that AI will fail. It is that investors may have committed too much capital in anticipation of returns that could take years to materialise.
1. The Magnificent Seven and the Circular AI Economy
The Magnificent Seven—Nvidia, Microsoft, Apple, Alphabet, Amazon, Meta and Tesla—have different exposures to AI. Nvidia supplies advanced chips, while Microsoft, Alphabet, Amazon and Meta are investing heavily in computing infrastructure. Tesla’s ambitions include autonomous driving and robotics.
Around these giants, an interconnected investment ecosystem has developed. AI model developers such as Anthropic require enormous computing capacity. Cloud providers invest in infrastructure, data-centre operators purchase chips and servers, and technology companies finance or partner with businesses that may become their customers.
Oracle and CoreWeave illustrate the scale of this expansion. The concern is whether spending is supported by sustainable revenue from end-users or by successive rounds of financing and investment commitments within the same ecosystem.
This circular economy is not automatically unsound. However, if major customers reduce spending, the consequences could spread through chip suppliers, cloud providers, infrastructure operators and lenders. Investors are increasingly asking whether revenue growth can translate into sufficient cash flow and profits to justify the cost of expansion.
2. Wall Street’s Divided Verdict: Correction or Bear Market?
Market participants disagree over whether the sell-off is a temporary correction or the beginning of a more prolonged downturn.
JPMorgan strategist Nikolaos Panigirtzoglou has warned that a recovery may be slower than previous rebounds, citing weaker dip-buying, the end of corporate share-buyback support and leveraged exchange-traded fund exposure. Ray Dalio has compared the AI boom to a classic financial bubble, while Michael Burry has questioned the optimism surrounding AI valuations.
The Philadelphia Semiconductor Index (SOX) recorded eight consecutive losing sessions in October, its longest such streak since October 2008. By 16 October, it had fallen approximately 16% from its 6 October level. The Nasdaq subsequently suffered its sharpest daily decline in four months on 23 October amid another technology sell-off.
The 2008 comparison concerns the duration of the semiconductor index’s losing streak; it does not establish that another global financial crisis is imminent. Nevertheless, the market is beginning to test whether AI investments can generate returns commensurate with their valuations and financing costs.
3. Malaysia’s Export Boom: Opportunity and Concentration Risk
Malaysia has benefited substantially from global demand for AI-related hardware.
According to MATRADE, Malaysia’s exports of AI-enabling products rose 42.9% to RM319.05 billion during January–April 2026. Electrical and electronics (E&E) exports increased by RM71 billion, or 32.1%, over the same period. Total trade reached RM1.127 trillion, while exports amounted to RM609.31 billion.
The World Bank raised Malaysia’s 2026 economic growth forecast from 4.4% to 5.1%, while warning that a correction in global AI-related activity could threaten regional growth.
These figures demonstrate Malaysia’s growing importance in the AI supply chain, but they also expose a vulnerability. Exporting AI-enabling hardware does not necessarily mean that AI adoption has transformed productivity across the domestic economy.
If American technology companies postpone purchases of chips, servers and related components, Malaysian manufacturers could experience weaker orders. A correction in international financial markets could also affect capital flows and the ringgit, although the magnitude would depend on broader economic conditions.
Malaysia must therefore distinguish between benefiting from the global AI investment cycle and building a more productive domestic economy.
4. Data Centres: RM225.97 Billion, But What is the Return?
Malaysia’s data-centre expansion presents a similar challenge.
In a parliamentary reply reported by Bernama on 14 October 2026, the Digital Ministry disclosed that 27 projects worth an estimated RM225.97 billion had been approved under the Data Centre Task Force up to July. Sixteen were in Johor, ten in Selangor and one in Negeri Sembilan. Of these, one was operational and 26 were still under construction.
The ministry projected 4,053 jobs from these projects, including 3,523 high-skilled positions. Malaysians were expected to fill 89.21% of the high-skilled roles, while approved projects were subject to a minimum 30% local-content requirement.
These are useful commitments, but projected employment is not the same as employment already created. Nor does the RM225.97 billion investment estimate establish how much domestic income, technology transfer or productivity improvement Malaysia will ultimately capture.
The government should ensure that operating licences are issued only to data centres that can demonstrate the existence of clients. This would help reduce the risk of building expensive facilities without credible demand.
The need for scrutiny is reinforced by resource constraints. Reuters reported that data centres accounted for a record 9.3% of national electricity consumption during the second week of August 2026, compared with an average of 7% for the year. Their share of Peninsular Malaysia’s electricity demand could reach 31% by 2035.
Malaysia should not reject data-centre investment, but it must establish whether the expected economic benefits justify the demands on electricity, water, land and public infrastructure.
5. Global Warnings: Infrastructure is Not the Same as Development
The World Bank’s 2026 World Development Report, The Promise of AI, emphasises that developing economies should prioritise adopting and adapting AI to local needs rather than attempting to replicate the expensive frontier infrastructure of leading technology powers.
The International Energy Agency’s report Powering Artificial Intelligence examines the implications of data centres for electricity systems, affordability, emissions and water.
The lesson for Malaysia is straightforward. The economic value of a data centre depends on more than its construction cost or computing capacity. Its contribution must be assessed against resource consumption, employment, domestic supplier participation and wider productivity gains.
A facility serving foreign customers can generate investment and export-related income. However, if most equipment, expertise and profits originate abroad, the benefits retained locally may be considerably smaller than the headline investment figure suggests.
6. What Should Malaysia Do?
First, strengthen commercial screening. Require evidence of actual clients, credible financing and realistic demand before granting operating licences.
Second, ensure investors bear the incremental costs of grid connections, electricity infrastructure, water provision and environmental safeguards rather than transferring these costs to households and other businesses.
Third, enforce the 30% local-content requirement and measure its results. The government should publish data on Malaysian suppliers, jobs actually created, salaries, training, technology transfer and domestic procurement.
Fourth, accelerate AI adoption across manufacturing, agriculture, logistics, healthcare and public administration. Malaysia’s strategic objective should be to use AI to improve productivity, not merely to host infrastructure for foreign customers.
Finally, stress-test the industry against lower utilisation, project cancellations, refinancing difficulties and weaker semiconductor demand. The government should monitor export orders, actual electricity consumption, investment completion rates and the difference between projected and realised employment.
7. Does AI Still Have a Future?
Yes. A market correction would not erase AI’s practical applications in software, manufacturing, research and business operations. However, the industry’s structure, valuations and investment pace could change substantially.
Before the long-term benefits materialise, some companies may fail, projects may be cancelled, data centres may operate below capacity, and technology suppliers may face weaker demand. Consolidation could eventually leave stronger businesses with more affordable computing infrastructure and sustainable business models, but that outcome is not guaranteed.
Malaysia must prepare for this transition rather than assume that current investment momentum will continue indefinitely.
The ultimate test is not how much capital enters Malaysia during the AI boom. It is how much productive capacity, domestic income and technological capability the country retains when the boom eventually matures.
No comments:
Post a Comment