“Many investors in AI don’t understand what they’re putting their money into,” says Hock Tan, CEO of chipmaker Broadcom. According to him, expectations surrounding AI are sometimes so high that investing in it is more like gambling in a casino.
During a conference in Singapore, Tan criticized the investment frenzy in the AI sector. According to Bloomberg, he had particular reservations about the viability of companies that develop AI models. These companies are investing billions in chips, data centers, and model training. However, profitability remains elusive.
Tan noted that OpenAI reached annual revenue of $50 billion in September. While this points to a promising revenue model, it offers no guarantee of future profits. After all, the costs of training AI models remain high. An interesting detail: Broadcom itself works closely with OpenAI. The chip manufacturer supported OpenAI in developing a chip for AI inference.
Companies are wasting money on experiments
Tan was also critical of companies’ use of AI. According to him, many companies are experimenting with AI without knowing what those investments will yield. AI model developers need to find more customers who actually get a return on investment from the technology. “Otherwise, the sector’s growth risks stagnating.”
Another challenge is capital availability. Tan warned that financing could well become the most significant limiting factor for further growth. Intel CEO Lip-Bu Tan (photo), who also participated in the discussion, agreed. He said investors are taking a closer look at the cost of capital and the returns on AI investments.
Chip shortages will persist for a long time
Hardware availability remains a problem. According to SiliconANGLE, Hock Tan expects the current chip shortages to ease around 2028 or 2029. Nevertheless, the CEO said that demand for computing power will continue to outstrip available supply through 2030.
The top executives thus painted a picture of an industry that is not only running up against technical limits. Even if production capacity increases, the question remains whether investors will continue financing the massive expansion of AI infrastructure.