CLSA analyst Bhavtosh Vajpayee told CNBC on Wednesday that expectations for the AI semiconductor cycle to peak in 2028 are "utterly wrong." Vajpayee estimates that demand for GPUs and custom AI chips currently exceeds supply by approximately 73% and expects this shortage to persist until roughly 2030.
Vajpayee noted that the industry remains in the early stages of its AI buildout, making it difficult to identify which downstream companies will ultimately dominate. He believes upstream semiconductor and hardware companies will benefit from a worsening shortage as AI infrastructure spending expands. Additionally, he pointed out that large semiconductor companies trade at lower valuation multiples than many other parts of the AI ecosystem.
Regarding specific companies, Vajpayee highlighted Meta’s Muse as a notable development because it makes creating AI agents easier through simple language commands. He believes simpler interfaces could expand AI use among consumers who do not want to understand coding or complex development tools. He also expects similar advances to emerge in China, where companies including Alibaba are developing smaller AI models.
Vajpayee expects smaller AI models to push artificial intelligence beyond data centers and onto personal devices. While OpenAI is developing its own device, he pointed to Apple as particularly important because smartphones already sit at the center of consumers’ daily lives. He said the smartphone could become the primary gateway through which many consumers access AI services.
For now, CLSA sees clearer opportunities in the hardware supplying the AI boom than among model developers, where Vajpayee said it remains too early to determine the eventual winners.
According to Benzinga Pro data, Microsoft shares were up 0.60% at $501.00, Meta Platforms shares were up 0.90% at $743.24 and Apple shares were up 0.24% at $340.57 during premarket trading on Wednesday.
