SEPTEMBER 5 — Blackstone president and chief operating officer Jon Gray has said something that Asean governments, corporations and especially small and medium enterprises should take seriously: the enormous sums being poured into artificial intelligence do not necessarily prove that AI is a bubble.
They may demonstrate precisely the opposite.
Gray’s argument is straightforward. In a conventional bubble, investment creates capacity far beyond genuine demand.
In AI, demand for computing power is still running ahead of available supply.
Data centres, electricity, advanced chips and computing capacity remain constraints.
Blackstone says demand for large language models across its companies increased sevenfold within six months while compute struggled to keep up.
That is not the normal anatomy of a bubble.
There will certainly be companies whose valuations cannot be justified.
Some AI ventures will fail. Some data centres may eventually prove badly located or excessively expensive.
Gray himself does not claim otherwise. Blackstone’s conviction is not that every AI company will succeed, but that demand for AI and its economic impact could exceed present expectations.
This distinction matters enormously for Asean.
The region should not confuse speculative excess in parts of the AI industry with the technological transformation itself.
The internet survived the dot-com crash. Railways survived railway speculation. Electricity survived the failure of individual electricity companies.
Technological revolutions can produce bubbles without themselves being bubbles.
AI increasingly belongs in this category.
Blackstone estimates that the five largest hyperscalers spent approximately US$415 billion on capital expenditure in 2025, against combined revenues of roughly US$1.7 trillion.
It expected their capital expenditure to rise another 45% in 2026. Importantly, much of this expenditure is being financed from corporate cash flows rather than the highly leveraged financing characteristic of some earlier infrastructure booms.
More importantly, AI is moving beyond Silicon Valley.
It is entering ordinary businesses.
That is where Asean should pay attention.
The decisive measure of AI’s future will not simply be whether OpenAI, Nvidia, Google, Microsoft, Meta or another technology giant produces the most powerful model.
It will be whether a restaurant in Kuala Lumpur, an exporter in Penang, a logistics company in Johor, a manufacturer in Vietnam, an online merchant in Indonesia or a tourism operator in Thailand discovers that AI can make everyday operations faster and cheaper.
That transition has begun.
SMEs do not need to construct billion-dollar large language models. They can rent intelligence through the cloud.
They can use AI to translate documents, answer customer enquiries, prepare invoices, analyse inventories, create advertising material, identify prospective customers, write software, forecast demand and manage logistics.
A company with twenty employees therefore does not have to become an AI company.
It merely has to become a company that knows how to use AI.
This is why the SME revolution may ultimately provide the strongest rebuttal to the bubble argument.
Blackstone’s own portfolio offers evidence of the broader transition. Its 2026 investment outlook reported that 77% of surveyed portfolio-company CEOs had increased spending on AI-related software in the third quarter, compared with fewer than half increasing spending on non-AI software.
Its mid-year update subsequently reported that AI adoption among Blackstone portfolio companies had reached roughly 50%, suggesting businesses are moving from experimentation towards actual deployment.
Asean should recognise the implications.
The region has hundreds of millions of increasingly digital consumers and an enormous ecosystem of micro, small and medium enterprises.
If AI reduces the cost of knowledge, translation, marketing, programming and administration, SMEs may be among its largest beneficiaries.
Malaysia’s strategy should therefore extend beyond attracting data centres.
Data centres are important. So are semiconductor investments, reliable electricity, cloud infrastructure and connectivity.
But Malaysia also needs to ensure that the intelligence generated by this infrastructure reaches factories, universities, government agencies and SMEs.
The next stage should consequently be mass AI literacy.
Workers should learn how to use AI tools. Universities and technical institutions should integrate them into teaching.
Companies should train employees rather than merely announcing AI strategies.
Government agencies should help SMEs experiment with affordable applications instead of assuming artificial intelligence belongs exclusively to technology corporations.
Asean must do the same collectively. Malaysia has started to encourage the use of AI more aggressively as marked by free courses by YTL, invariably, even logistic companies that are exposed to AWS, for that matter, Sky Lark.
The region should avoid being trapped in the argument over whether the present excitement surrounding AI resembles 1999.
That is increasingly the wrong question.
The better question is whether Asean will possess the infrastructure, energy, skills, companies and workforce necessary to participate in an economic transformation already underway.
Blackstone is right to distinguish AI infrastructure investment from a classic bubble.
Asean is also right to bet on AI.
The strongest evidence will not ultimately come from Wall Street valuations.
It will come when millions of Asean SMEs discover that AI is no longer something happening in Silicon Valley or Shenzhen.
It is sitting on their laptops and smartphones, helping them compete.
And once that happens at scale, calling AI merely a bubble will become increasingly difficult.
* Phar Kim Beng, PhD is the Professor of Asean Studies at International Islamic University of Malaysia and Director of Institute of International and Asean Studies (IINTAS).
** This is the personal opinion of the writers or publication and does not necessarily represent the views of Malay Mail.
