We came away from our recent meeting with GAMUDA feeling assured about its near-term outlook, particularly in the data centre space. Key takeaways are as follows: (i) robust pipeline of data centre projects, (ii) limited impact of U.S. chip export restrictions, and (iii) resilient orderbook growth. Maintain BUY with an unchanged SOP-derived TP of RM6.30/share.

Robust Pipeline of Data Centre Projects

Despite the U.S.’s restrictions on artificial intelligence (AI) chip exports to Tier-2 countries, GAMUDA remains cautiously optimistic about the outlook for data centre (DC) projects, supported by a strong pipeline of potential rollouts. Based on SemiAnalysis’ estimates, global AI-driven DC capacity is projected to reach approximately 20GW by 2025 and is expected to triple by 2027. Meanwhile, GAMUDA’s key client, “X”, a leading U.S. hyperscaler, is targeting to achieve a global DC capacity of 18–22GW by 2027 based on our channel check.

The U.S.’s 7% cap on hyperscalers’ total computational power per country (CPPC) and the 25% aggregate limit for all Tier-2 countries translate to an estimated 1.2–1.4GW of new DC capacity to be established by hyperscaler X within a single Tier 2 countries by 2027, based on our projections. Despite these constraints, Malaysia remains a highly attractive destination for DC hyperscalers due to its political friendly environment and comparatively lower costs for power and water.

Using hyperscaler X’s DC 1 project in Elmina Business Park, with a capacity of approximately 80MW, as a benchmark, the total project value for a new DC is estimated at around USD2bn per 100MW, with construction costs typically accounting for 15–20% of this total value. Assuming hyperscaler X proceeds with its initial plan to expand its DC presence in Malaysia, this would translate into a potential rollout of at least 1.1GW of new DC projects. This expansion represents a collective investment of c.USD22.4bn (equivalent to RM100.9bn), with gross construction costs of c.USD4bn (equivalent to RM18bn). These projections highlight considerable opportunities for GAMUDA to capitalise on the potential DC expansion of its key client during 2025–2027.

GAMUDA has forged a strong partnership with hyperscaler X and is strategically positioned to secure these additional DC projects through direct negotiations. This confidence is bolstered by GAMUDA’s recent acquisition of 389 acres of land in Port Dickson, Negeri Sembilan, which is estimated to have the capacity to accommodate up to 10 new DCs, assuming 40 acres are required for every 100MW of DC developed. Additionally, GAMUDA’s technical expertise and operational capacity enable it to manage 8–10 DC projects simultaneously, underscoring its readiness to handle the growing demand.

Limited Impact of U.S. Chip Export Restrictions

The U.S.’s 7% CPPC cap is anticipated to have a limited impact on DC expansions, given the urgent global demand for hyperscalers to scale up their DC capacity to meet large-scale cloud storage requirements. While the U.S.- based hyperscalers are mandated to maintain at least 50% of their total AI computing power within the U.S., we believe this will not hinder their efforts to establish DCs in Tier-2 countries. This strategy is driven by the need to diversify concentration risks, benefit from lower operating expenses, and enhance global connectivity.

In line with this, GAMUDA is proactively engaging with other U.S.-based MNC hyperscalers to broaden its project portfolio. This effort is further supported by GAMUDA’s innovative “bundled land + power + water” integrated solution, which offers cost-effective and efficient setups tailored to the needs of DC offtakers.

Additionally, the U.S. export restrictions currently apply only to a few selected GPUs, a crucial component for specific types of DCs, particularly those supporting AI/machine learning and high-performance computing. The restrictions do not affect general-purpose CPUs or lower-end GPUs typically used for standard computing tasks, such as cloud storage, web hosting, and enterprise applications. In response, GAMUDA’s key client, hyperscaler X, is actively developing its own AI chips, which are expected to enter mass production by 2026. These in-house chips will enable hyperscaler X to reduce its reliance on major chipmakers while meeting its large-scale cloud storage requirements without breaching the 7% cap. As these chips are not subject to the current restrictions, they will likely facilitate hyperscaler X’s DC expansion in Malaysia.

This continued growth in DC projects and associated construction opportunities is expected to bolster GAMUDA’s construction orderbook, which will support GAMUDA’s target total unbilled orderbook of RM45bn by the end of CY25.

Resilient Orderbook Growth

Following its recent win of the Penang LRT segment 1 project, GAMUDA plans to proactively tender for segment 3, which involves system works and rolling stock. According to our report dated 14 January 2025, we estimate the cost for segment 3 is approximately RM3bn–RM4bn. The tender, which opened in mid-December 2024, is set to close on 14 April 2025, with results expected by the end of CY25. GAMUDA is considered a frontrunner for this project, leveraging its strong technical expertise demonstrated in similar projects in Australia.

In addition, the rollout of the Upper Padas water supply scheme, valued at approximately RM5bn, is anticipated to be finalised by 3QCY25. With GAMUDA holding a 75% stake in the project, this is expected to bolster its effective orderbook by another RM3.5~4bn.

Regionally, GAMUDA continues to target growth in Australia, driven by largescale, high-margin renewable energy projects. Notably, the company has been shortlisted for a pumped hydro project, with outcomes expected by 1QCY25. Overall, our FY25 job replenishment assumption of RM20bn for GAMUDA remains intact, supported by these projects in the pipeline.

Impacts

We Maintain Our FY25-27F Earnings Forecasts Unchanged.

Valuation

We reiterate our SOP-derived target price of RM6.30, inclusive of a 3% ESG premium based on our 4-star rating. Maintain Buy call on the stock. We believe the recent heavy selloff in GAMUDA’s share price offers an attractive buying opportunity for investors. This view is supported by the company’s unchanged fundamentals, underpinned by resilient earnings visibility and a robust pipeline of projects. Notably, its exposure to data DC represents less than 7% of the unbilled order book as of now.

Potential downside risks include: (i) a narrower quota than the initial 7% of total CCPC in each Tier-2 country, (ii) delays in the rollout of new DC projects by its key hyperscaler client, and (iii) more stringent regulations and caps imposed by the U.S., including the potential suspension of AI chip exports to Malaysia.

Source: TA Research - 20 Jan 2025