SEA Daily Brief: BNM Fines Three Money Changers, Kenanga's 'Second Inning' for AI, and ISEAS Maps ASEAN's Split Over AI's Climate Cost

Bank Negara Malaysia Fines Three Money Changers RM40,400 Over Due Diligence and Sanctions-Screening Failures

Bank Negara Malaysia imposed administrative penalties totalling RM40,400 on three licensed money-changing businesses following supervisory examinations, under the Money Services Business Act 2011. Wawasan Ilham (M) Sdn. Bhd. was fined RM22,000 for failing to conduct customer due diligence on a money-changing transaction. SMJ Teratai Sdn. Bhd. was fined RM6,900 for not exercising reasonable care in identifying and verifying an authorised company representative. Ruhanmas Forex Sdn. Bhd. was fined RM11,500 for having no sanctions screening mechanism and failing to screen customers against Malaysia's Domestic List and the UN Security Council sanctions lists.

All three firms have paid and put corrective measures in place, including staff training, strengthened internal controls, updated procedures and, in Ruhanmas Forex's case, integration of a commercial sanctions database service. The amounts are small, but the findings are basic: a missing screening mechanism is a design failure, not an edge-case miss.

Kenanga Says the AI Cycle Has Entered Its Second Inning, With Power Availability as a Gating Factor

Kenanga Research, speaking at the Huatai Securities investment forum in Shanghai, said the market is moving "from an AI narrative towards evidence of execution and earnings conversion." It describes 2023-2025 as the first inning, dominated by GPUs and high-bandwidth memory, and 2026 onward as a second inning centred on inference, AI agents and more diversified infrastructure demand.

Kenanga lists five bottlenecks: compute, storage, optical connectivity, power and equipment such as advanced packaging. On power it said "the availability of power could determine which projects proceed." Its screening principle is "capacity expansion beats price increases," which it reads as a sign of structural rather than temporary demand. Among Malaysian names it cited NATGATE, which is targeting a November 2026 production start in optical transducers, while keeping a neutral sector stance.

ISEAS Survey Finds ASEAN Divided on Whether AI Helps or Harms the Environment

The ISEAS Southeast Asia Climate Outlook Survey Report 2026, released October 2, polled 3,706 respondents across eleven ASEAN member states, including Timor-Leste for the first time. On AI's environmental impact, 47.9% responded positively (27.2% "somewhat good", 20.7% "clearly good"), 30.2% negatively (17.2% "somewhat bad", 13% "clearly bad") and 21.9% said they lacked enough knowledge to judge.

Views vary by country. Laos (45.7%), Cambodia (33.7%) and Vietnam (33.1%) were the most likely to say AI is clearly good, while Brunei was the most sceptical at 35.4% clearly bad against 1% clearly good. Roughly one-fifth of respondents in Indonesia, Malaysia, the Philippines, Singapore, Thailand and Vietnam were unsure. Higher education correlated with scepticism: 11.4% of master's holders and 11.8% of doctoral holders said clearly good, versus 38% of those with no formal education. The report attributes this to greater familiarity with AI and possibly its energy demands.

Eyes on the Day Ahead

Nothing concrete is scheduled for the next 24 hours, so this looks to the next few days. Singapore's consultation on writing its stablecoin framework into the Payment Services Act closes on October 16, and XREX Group's Stablecoin Summit takes place in Singapore on October 8. Watch for further BNM enforcement notices and for how Malaysian data-centre suppliers describe power constraints in upcoming results.