Three developments offer different perspectives on the AI boom: Malaysia’s investigation into a shipment of advanced chips, widespread generative-AI use in Taiwan, and a sharp technology-stock sell-off as investors weigh AI spending against returns. Together, they show the growing reach of AI and the pressure to manage its risks and justify its costs.
- Malaysia seized 30 servers, though their connection to new chip-export rules remains unconfirmed.
- A Taiwan survey found that 50.5% of adults had used generative AI in the previous six months.
- Alphabet’s shares plunged as investors scrutinized spending, despite strong revenue and cloud growth.
Malaysia investigates a shipment of advanced chips
Malaysia Customs seized 30 servers containing chips valued at about 49.5 million ringgit ($12.19 million). The seizure followed an inspection of a warehouse in the Kuala Lumpur International Airport (KLIA) Free Trade Zone on October 15, 2026.
Customs director Aspalila Ag. Tuah said the shipment arrived from an Asian country under one air waybill. It was declared as electronic components and allegedly intended for re-export to another Asian country. Authorities have not identified the chips’ brands, countries of origin or intended destination. The investigation is ongoing.
“The syndicate also used Malaysia as a transit point to ensure there were no restrictions on the export of the servers to their next destination, ” Aspalila said.
Malaysia introduced permit requirements in 2025 for the export, transit and transshipment of high-performance, U.S.-origin AI chips. The rules rely on the Strategic Trade Act 2010 (STA 2010), which covers certain sensitive goods and technologies, including items with both civilian and military applications.
The Act’s “Catch-All Control” provision can apply to items not listed on Malaysia’s Strategic Items List. In some cases, anyone who knows or suspects that an item could be misused or used for a restricted activity must notify the authorities at least 30 days before the transaction.
Senior assistant director Masyita Ismail of the Investment, Trade and Industry Ministry’s Strategic Trade Secretariat said the chips were controlled as dual-use items, meaning goods that may have civilian uses as well as restricted applications.
“The Strategic Trade Act 2010 regulates dual-use items, which can be used in the development of weapons of mass destruction, ” Ismail said.
The seizure does not establish that the shipment involved U.S.-origin chips covered by Malaysia’s permit requirements. Authorities have also not confirmed where the servers were headed or who was behind the alleged operation.
The controls are part of a broader effort to restrict China’s access to advanced semiconductors. The U.S. Department of Commerce said its October 2023 export-control revisions aimed to limit China’s ability to obtain critical technologies for military modernization. The department described the measures as protecting U.S. and allied national security interests.
Generative AI reaches a broad audience in Taiwan
A Taiwan Fair Trade Commission (TFTC) survey found that 50.5% of Taiwanese adults had used a generative-AI tool in the previous six months. Among those users, 39.8% said they had paid for a subscription.
The TFTC conducted its “Behavioral Survey of Consumer Use of Generative Artificial Intelligence Tools” from May 6 to 28, 2026, and released it on October 29. It collected 1, 079 valid responses from people aged 18 and older across Taiwan’s 22 counties and cities. The findings describe the people surveyed, not every adult in Taiwan.
The survey reported tool-use figures of 56.6% for ChatGPT and 34.6% for Google Gemini. These should not be read as exclusive market shares. Many respondents had used or tried multiple tools, and the survey does not explain in detail how it handled that overlap.
Respondents mainly used generative AI to gain knowledge or learn, and for work. The leading reason they chose a brand was that its services or functions met their personal needs.
The TFTC said generative-AI use had “grown rapidly in recent years.” Separately, Microsoft estimated that 18.8% of the world’s working-age population, defined in its report as people aged 15 to 64, had used a generative-AI product in the second quarter of 2026, up from 17.8% in the first quarter. The Taiwan survey and Microsoft estimate cover different populations and use different measures, so their percentages are not directly comparable.
Alphabet’s sell-off puts AI spending under scrutiny
Alphabet shares fell more than 12% on October 29, 2026, marking the company’s worst single-day share-price performance on the U.S. market. The drop wiped out more than $500 billion in market value. Market coverage called it the second-largest single-day market-cap loss on record, after Nvidia’s $593 billion loss on January 27, 2025.
The sell-off came despite strong third-quarter results. Alphabet reported that revenue rose 25% to $127.6 billion, while Google Cloud revenue more than doubled to $30.7 billion, an increase of 103%.
Investors also had reasons to question the costs and returns. Search revenue grew 12%, below the 15% analysts had expected, according to the consensus cited in market reporting. Capital spending more than doubled year over year to $51.4 billion. Free cash flow, the company’s operating cash flow minus capital expenditures, was negative $9.1 billion.
The numbers tell two sides of the story. Google Cloud’s growth signals strong demand, while rising spending and negative free cash flow show the cost of building infrastructure to meet it. Cloud growth alone does not show whether Alphabet’s broader AI investments will generate returns large and lasting enough to justify the expense.
The sell-off spread to other technology shares. On October 29, ServiceNow fell 27%, Salesforce dropped 4.6%, and Workday slid 4%. On Friday, October 30, Intel fell roughly 29%, putting it on track for its worst day since 1974. Amazon was also reported to have recorded more than $42 billion in cash burn over the 12 months through September 2026. Without a more precise definition of that figure, it is difficult to compare it with Alphabet’s free-cash-flow measure.
The market moves have revived comparisons with the dot-com bust, when expectations for internet companies outpaced the profits many could deliver. That comparison is a warning, not proof that AI is in a bubble or that one has burst. A sell-off can reflect concerns about spending, valuations or future returns without settling the long-term case for AI. The ChatGPT Atlas launch has also put pressure on Alphabet and raised separate questions about privacy and competition.
Google Cloud’s 103% revenue growth points to substantial demand for cloud services. The open question is whether that demand can generate sustainable profits to cover the cost of the infrastructure under construction. The Taiwan survey and Malaysia seizure offer separate measures of AI’s spread and the competition over its hardware. Neither explains how the market views Alphabet’s returns.
Key questions and takeaways
-
Why did Malaysia seize the servers?
Customs said it suspected the 30 servers were part of a smuggling operation. They had been declared as electronic components for alleged re-export. Officials have not confirmed the chips’ origin or destination, or whether they were covered by Malaysia’s permit rules for certain high-performance chips.
-
How widely is generative AI used in Taiwan?
The TFTC survey found that 50.5% of Taiwanese adults surveyed had used a generative-AI tool in the previous six months. Of those users, 39.8% said they had paid for a subscription.
-
What do the Taiwan tool-use figures show?
The survey reported ChatGPT at 56.6% and Google Gemini at 34.6%. Respondents could use multiple tools, so the figures are not exclusive market shares.
-
Does Alphabet’s share-price fall prove the AI boom is over?
No. Alphabet’s third-quarter revenue rose 25% and Google Cloud revenue grew 103%, while higher capital spending and negative free cash flow raised sharper questions about returns. Whether future profits will justify the investment remains unresolved.
