Broadcom Inc. (AVGO.US) CEO pushes back on "AI brake" theory: AI semiconductor revenue target unchanged, aiming for $230 billion by 2028.

The debate over whether to hit the brakes on frontier AI model development continues to simmer across the industry. After NVIDIA Corporation (NVDA.US) CEO Jensen Huang directly rejected the "AI brake" theory at an event, Broadcom Inc. (AVGO.US) CEO Hock Tan also sought to downplay related concerns. In a recent interview, Hock Tan responded to concerns that a potential slowdown in frontier AI model development could impact the chipmaker's business, emphasizing that the company remains committed to its long-term revenue targets. Hock Tan's remarks came as investors were rattled by an article published over the weekend by Anthropic CEO Dario Amodei, who advocated for slowing the pace of model development. Amodei's views were also endorsed by OpenAI CEO Sam Altman and Elon Musk, prompting investors in AI infrastructure suppliers to reassess their expectations for computing demand during Monday's trading session. As a result, Broadcom Inc. shares fell 4.8% its custom chip business counts Anthropic as one of its most important customers. The semiconductor ETF-iShares (SOXX.US), which covers chip stocks, dropped 5.6%. Shares of companies selling other data center components also took a heavy hit. In response, Huang made clear on Monday at the All-In Summit in Los Angeles, while discussing AI safety with the host, that he rejected predictions that "AI could destroy the world," calling such claims "not scientifically grounded." A week earlier, Huang also made a pointed remark: panic around AI safety may simply be creating business opportunities for the cybersecurity industry "what better way to create demand than to create a problem?" When renowned financial commentator Jim Cramer asked on a program whether any factor in the AI slowdown debate had made Hock Tan reconsider Broadcom Inc.'s AI semiconductor forecasts for fiscal 2027 and 2028, he replied: "No, not at all." Hock Tan reassured the market from a demand and revenue perspective, saying, "We see that whether it's computing infrastructure for frontier large model development or AI inference computing for deployed products, demand remains very strong, and I believe this demand is highly sustainable." On Broadcom Inc.'s fiscal 2026 third-quarter earnings call on September 2, Hock Tan had predicted that the company's AI semiconductor revenue would reach $115 billion in fiscal 2027 and double again to $230 billion in fiscal 2028. This 2028 performance target, which exceeded market expectations, was a major highlight of the earnings report. It is understood that Broadcom Inc.'s AI revenue consists of two parts: custom AI accelerators and networking chips used in AI systems. Hock Tan also mentioned that Anthropic is expected to become Broadcom Inc.'s largest custom chip customer in 2027 and maintain that position in 2028 which explains why Broadcom Inc. investors are so sensitive to Amodei's remarks. Previously, Alphabet Inc. Class C (GOOGL.US) had been Broadcom Inc.'s largest custom chip customer, with the two companies co-designing Alphabet Inc. Class C's tensor processing units (TPUs). In Monday's interview, Hock Tan was particularly optimistic about inference computing demand (referring to computing power used in everyday calls after AI models are trained). "On the training side, I can't really judge, but once you turn models into commercial products and deploy them in inference scenarios, the demand I see will continue to maintain extremely strong growth," Hock Tan said. Additionally, Amodei's article over the weekend further intensified the debate over whether the AI industry is advancing too quickly in developing powerful models. Amodei proposed a three-step plan aimed at slowing the pace of development without "sacrificing commercial advantage or America's leadership in AI." Hock Tan said he agrees with Amodei's view that certain restrictions need to be placed on AI, but suggested his own concerns about the direction of technological development are not as strong. "Like any tool, it's very important to govern and safeguard how we use it," Hock Tan said. But he also offered his own perspective: "AI is not a living organism that will run wild and out of control." Hock Tan instead emphasized AI's potential to boost productivity. "AI, generative AI, the deployment of those frontier models... will create tremendous value," Hock Tan said, likening it to the Industrial Revolution that began in 18th-century Britain. "Ultimately, it is still a tool that will elevate our society and humanity to a higher standard of living."