Broadcom (AVGO.US) CEO Responds to "AI Brakes" Theory: AI Semiconductor Revenue Target Remains Unchanged, Aiming for $230 Billion by 2028

According to Zhitong Finance APP, the debate around whether to hit the brakes on cutting-edge AI model development continues to ferment throughout the industry. Following NVIDIA (NVDA.US) CEO Jensen Huang's outright rejection of the "AI slowdown" view at a recent event, Broadcom (AVGO.US) CEO Hock Tan has also sought to downplay related concerns.

In a recent interview, Hock Tan responded to worries about a possible slowdown in advanced AI model development impacting the chipmaker's business and emphasized that the company still adheres to its long-term revenue targets.

Hock Tan made these remarks as investors grew uneasy over an article published last weekend by Anthropic CEO Dario Amodei, which advocated for slowing down model development. Amodei's viewpoint was also backed by OpenAI CEO Sam Altman and Elon Musk, prompting investors in AI infrastructure suppliers to reassess expectations for computing power demand during Monday’s trading session.

As a result, Broadcom’s stock price fell 4.8%—its custom chip business counts Anthropic as one of its most important customers. The semiconductor ETF covering chip stocks—iShares (SOXX.US)—declined by 5.6%. Shares of companies selling other data center components also suffered heavy losses.

In response, Jensen Huang explicitly dismissed predictions that "AI could destroy the world" when discussing AI safety issues with a host at Monday's All-In Summit in Los Angeles, arguing that such claims "have no scientific basis." The previous week, Huang had made a telling comment: panic around AI safety could simply be creating business opportunities for the cybersecurity industry—"What could create more demand than inventing a problem?"

When renowned financial commentator Jim Cramer asked Hock Tan on a show if anything about the AI slowdown debate would make him reconsider Broadcom’s AI semiconductor forecasts for the 2027 and 2028 fiscal years, he replied: "No, absolutely not."

From the perspectives of demand and revenue, Hock Tan reassured the market, saying, "What we see is that, whether it’s computing power infrastructure for the R&D of cutting-edge large models or AI inference compute for applied products, demand remains extremely strong, and I believe such demand is highly sustainable."

At Broadcom’s Q3 FY2026 earnings call on September 2, Hock Tan predicted that the company’s AI semiconductor revenue would reach $11.5 billion in FY2027, and would double to $23 billion in FY2028. This above-market-expectation 2028 target was also a major highlight of the earnings report. Reportedly, Broadcom’s AI revenues are composed of custom AI accelerators and networking chips for AI systems.

Hock Tan also mentioned that Anthropic is expected to become Broadcom’s largest custom chip customer by 2027 and maintain that position in 2028—explaining investors’ high sensitivity to Amodei's remarks. Previously, Google (GOOGL.US) was Broadcom’s largest custom chip customer, with both parties co-designing Google’s tensor processing units (TPUs).

In Monday’s interview, Hock Tan was particularly optimistic about the demand for inference compute (referring to compute required for AI model inference in daily use after training completion).

"I'm not in a position to judge the training aspect, but once models become commercial products and are deployed in real-world inference scenarios, the demand I see will continue to show extremely strong growth," Hock Tan said.

Additionally, Amodei’s article last weekend further intensified debates on whether the AI industry is moving too quickly in developing powerful models. Amodei proposed a three-step plan aimed at slowing development without "sacrificing commercial advantages or America’s leading position in AI."

Hock Tan responded that he agrees with Amodei regarding the need to impose some restrictions on AI, but implied that his own concerns over the direction of technological development are not as strong.

"Just like with any tool, it’s very important to govern and ensure how we use these tools," Hock Tan stated.

But he also stressed his opinion, "AI isn't a living being that will run wild and out of control."

Hock Tan then emphasized the potential of AI to enhance productivity. "AI, generative AI, and the deployment of those advanced models... will create tremendous value," he said, comparing it to the Industrial Revolution that began in Britain in the 18th century. "Ultimately, it is still a tool that will enable our society and humanity to reach a higher standard of living."