We downgrade Gamuda to a HOLD (from BUY). The Group's share price has fallen by 13.8% after US announced new restrictions on AI chip exports on 13-Jan. Though we believe current data centre-related projects remain intact, we believe orderbook replenishment prospects are likely to be affected and hence revise our earnings forecasts by -2% to -7% for FY26- FY27. Additionally, we expect valuations to see further correction as negative sentiment lingers from the overhang. We revise our RNAV-based target price (TP) to RM4.70/share (from RM5.70) which implies a CY26 PER of 20x, or >+2 SD to its 10- year mean of 16x.

• AI chip restriction to reduce DC market size. We believe the move will reduce the addressable market for data centres (DC) in Malaysia. In our strategy report De-risking from AI theme (15- Jan), we highlighted that the notional chip cap of 50k translates to a power requirement of only 76.5MW vs. Tenaga Nasional Bhd's projected maximum DC-related demand of 4.7GW (actual load utilisation of 248MW).
• Downwards revision on future DC-related orderbook replenishment. Though we believe Gamuda will not face risks of cancellations for existing DC projects as end-clients are not China-related companies, we are concerned over the Group's prospects in FY26-FY27. Recall, management has begun scaling its DC operations to 8 teams which can cater to works worth up to RM16bil which we believe this will be crucial due to the lack of visibility over new public infrastructure projects in Malaysia. As such we revise our orderbook replenishment assumptions from DC projects by 60%, bringing our value to RM22.5bil and RM18.7bil for FY26 and FY27 respectively. Correspondingly, this will lead to earnings decline of -2% to - 7%. For reference, DC-projects currently accounts for 8.3% (or RM2.5bil) of the group's total orderbook as at 1QFY25.
• Correction for construction segment likely, revise TP downwards. We believe there is further room for correction as share price is still at +72% vs its 1-year performance whilst PER valuation appears to be stretched at 18.3x PER or >+2 SD against its 10-year mean. Admittedly, without the DC catalyst, upside from our sector re-rating thesis now appears capped. Hence, we revise our target PER for the construction segment to 16x (from 22x prior). This brings our RNAV-based TP downwards to RM4.70/share (from RM5.70) which implies a PER of 20x. We believe this is justified based on Gamuda's growing orderbook size relative to its historical performance.

Source: AmInvest Research - 24 Jan 2025