ASMPT (00522), the global leader in semiconductor packaging equipment, has pushed back against growing investor concerns over an AI bubble, with its CEO, Huang Zida, asserting that long-term demand for AI-driven equipment remains robust.
In an interview on Wednesday, Huang Zida stated that as new AI systems incorporate more chips, chip architecture will continue to evolve and transform. This, he explained, will necessitate more back-end packaging processes, a core component of ASMPT's business. "We want the entire industry to grow, not just in value, but also in the volume of chips that need packaging," he said. "That's why, despite stock market fluctuations, we remain broadly optimistic."
Huang Zida, who is set to retire next month, also dismissed concerns about the risk of Chinese domestic competitors eroding market share through aggressive price cuts to advance localization. He noted that the company will continue to "double down" on research and development, targeting clients willing to pay for advanced packaging technology rather than simply seeking discounts.
These comments come after the company's Hong Kong-listed shares dropped as much as 9.7% on second-quarter earnings that fueled investor worries about overinvestment in AI capacity. While ASMPT's second-quarter net profit jumped 155%, it still fell short of market expectations for even higher growth. However, its quarterly revenue exceeded analyst estimates, and its growth forecast for the current quarter outpaced analyst projections.
AI Equipment Demand Drives Orders Higher
A team led by Citi analyst Kevin Chen wrote in a report: "We see strong future demand for advanced packaging, while the mainstream semiconductor business remains steady." Huang Zida indicated that while orders surged 98% year-over-year, they have not yet fully converted into billable revenue due to extended lead times for certain materials, which have constrained shipments.
According to Huang Zida, ASMPT is actively expanding its supply chain network to shorten lead times and fulfill orders faster, thereby accelerating revenue recognition.