🇲🇾🤖 AI

Malaysia Weighs Huawei Partnership to Bolster National AI Infrastructure Ambitions

The government is considering utilizing Chinese-made AI chips to accelerate its domestic data centre roadmap despite potential diplomatic friction with the United States.

The Malaysian government is currently evaluating the integration of artificial intelligence chips manufactured by Huawei into its national infrastructure strategy. This consideration is part of a broader RM2 billion initiative aimed at positioning Malaysia as a premier hub for AI-driven data centres within the region.

According to a report by the original publisher, the potential shift toward Huawei hardware comes at a time when the United States has imposed significant export restrictions on high-end computing components. While the performance of Huawei’s AI accelerators is widely acknowledged to trail behind the industry-leading chips produced by US-based NVIDIA, the Chinese manufacturer has increasingly become a viable alternative for entities unable to access Western technology due to trade sanctions.

For the Malaysian administration, the decision presents a delicate balancing act between rapid technological scaling and international relations. Procuring Huawei hardware would provide a significant boost to the Chinese company’s sales revenue, a move that is likely to draw scrutiny and disapproval from the current Trump administration in Washington. The government finds itself in a precarious position, caught between the urgent need for competitive hardware to power its sovereign AI ambitions and the risk of straining critical trade and diplomatic ties with the United States.

The push to bolster AI infrastructure is a key pillar in the country’s economic growth strategy, which currently sees real GDP expanding at a robust 6.0% year-on-year. For Malaysian SMEs and local technology firms, the success of this initiative could determine the accessibility and cost of AI cloud services in the coming years. If the government opts for a more diversified hardware procurement strategy, local businesses might benefit from a broader range of service providers, though they may also face a fragmented technological landscape where compatibility between different chip architectures becomes a concern.

From a macroeconomic perspective, the decision to invest RM2 billion into this sector arrives during a period of relative stability, with headline inflation at 1.8% and unemployment holding at 3.0%. For the average Malaysian worker, the data centre initiative represents a potential pipeline for high-value jobs in the digital economy. However, the reliance on non-Western hardware could influence the long-term operational costs for companies building their digital ecosystems within Malaysia’s borders, potentially affecting the price points of consumer-facing digital services.

The move also arrives amid a complex global trade environment where Malaysia has historically maintained a neutral stance, balancing heavy investment from both Chinese and American firms. Historically, the nation has benefited from being a neutral ground for semiconductor manufacturing and testing. A pivot toward Huawei technology may signal a shift in this long-standing policy, raising questions about how existing American investors and partners in Malaysia’s tech sector will respond to such an alignment.

Looking ahead, observers will be watching for clear signals from the Ministry of Investment, Trade and Industry regarding the procurement criteria for these government-backed projects. The decision will likely depend on whether the performance gap between Huawei’s hardware and NVIDIA’s accelerators can be offset by lower costs, faster delivery timelines, or more flexible partnership terms.

Crucially, the full scope of any potential trade agreements or long-term commitments remains undisclosed. It is not yet clear whether the government intends to pursue a hybrid infrastructure model using both US and Chinese chips, or if it is leaning toward a complete procurement shift for these specific national projects.

Source

Originally reported by Lowyat.NET. Read the original report →

Join the conversation

We post stories like this all day on Threads. Discuss this story on Threads →

More in AI