The announcement of potential US export restrictions on artificial intelligence (AI) chips has ignited concerns within Thailand's technology and investment sectors.

Yupapin Wangviwat, Chief Financial Officer of Gulf Development (GULF), stated that, as of now, the company has not been affected by the prospective US policy.

Gulf's customer operations and the data center business under its subsidiary, Gulf Siam AI Data Centre (GSA DC)—a joint venture with AIS and NVIDIA—are reportedly progressing as planned.

The first phase of GSA DC's operations is scheduled to commence in mid-2025, with an expansion planned to achieve over 50 megawatts of service capacity in subsequent phases.

Furthermore, GULF is collaborating with Google Cloud to enhance capabilities in AI and cybersecurity.

Yupapin emphasized the company's target of achieving 20-25% year-on-year revenue growth by 2025, noting that while energy remains its primary revenue source, there is a growing contribution from new ventures like data centers and digital infrastructure.

She remarked, "So far, based on updates from our teams, we have not encountered any impact.

However, we continue to closely monitor the situation," and referred to the company's strategy of positioning itself as a regional digital infrastructure provider to mitigate external volatility.

The uncertainty surrounding the US policy has led sources within Thailand's tech sector to express that it may be premature to evaluate the full implications of the restrictions, particularly with variable factors such as the completion and enforcement of the regulations.

If Thailand limits its AI chip imports solely for domestic use without acting as an intermediary for China, it may avoid direct consequences from the restrictions.

However, concerns were raised that long-term perceptions of Thailand’s technological vulnerability could deter foreign investors, particularly given the country's efforts to promote its logistics, energy, and macroeconomic stability as attractive investment features.

In recent years, prominent tech companies, including Microsoft, Amazon Web Services, and Google Cloud, have made significant investments in developing data centers and cloud infrastructure in Thailand.

However, should the US enforce regulations mandating foreign data center providers to obtain certification from the US government, Thailand's appeal for investment in digital infrastructure could face increased scrutiny.

Some analysts believe that the immediate effects of the export controls would likely be limited, as Thailand’s data centers do not primarily aim to transfer technology to China.

Nonetheless, the necessity for transparent policies and safeguards to bolster investor confidence has been highlighted as critical.

Thailand's electronics sector is primarily engaged in packaging and testing components rather than the direct importation of advanced AI chips.

Significant repercussions are expected only if local companies are found to be indirectly facilitating the transfer of US-manufactured chips to Chinese clients—a connection that is often difficult to trace.

The complexities of AI chip importation are underscored by their production processes, which are typically custom-made for specific applications and involve production quotas from manufacturers like NVIDIA, alongside adherence to US export regulations.

Analysts indicate that Thailand currently imports only a minimal volume of AI chips, with most being incorporated into finished products such as smartphones and gaming consoles.

Despite this, stakeholders are recognizing the potential impact on the sector and calling for proactive measures, including the establishment of a national monitoring system for AI chip usage, enhanced strategic cooperation with the US and its allies, and a review of national digital strategies to mitigate dependence on any single country.

AI chips vary widely, with graphics processing units (GPUs) being the most prominent in the market, extensively utilized for training and executing AI models, particularly in deep learning scenarios.

GPUs, such as the NVIDIA A100, are noted for their energy consumption and cost.

Additionally, tensor processing units (TPUs), developed by Google, are tailored for accelerating machine learning workloads.

In contrast, central processing units (CPUs), while pivotal for general computing, are not optimized for AI tasks and operate at significantly lower performance levels compared to specialized processors.