Broadcom CEO Hock Tan on AI Revenue Targets
· dev
Broadcom CEO Addresses Anthropic’s Slowdown Push, Says AI Revenue Targets Haven’t Changed
In a recent essay, Anthropic CEO Dario Amodei advocated for slowing down frontier model development, sparking debate about the industry’s pace and its implications for companies like Broadcom that rely heavily on AI compute demand. Despite this unease among investors, Broadcom CEO Hock Tan remains confident in his company’s long-term revenue targets.
Tan’s optimism is rooted in his conviction that demand for compute infrastructure – particularly for inference, the day-to-day usage of AI models after they’ve been trained – will continue to drive growth. He points to the potential for generative AI to create significant value, much like the Industrial Revolution transformed society and commerce. This comparison highlights Tan’s view of AI as a catalyst for human progress.
However, this optimism raises questions about Broadcom’s business model sustainability. The company has forecasted $115 billion in AI semiconductor revenue by fiscal 2027, doubling to $230 billion by 2028. These targets rely on continued growth in compute demand – a trend that Amodei and others question. If the pace of frontier model development slows down, will Broadcom’s customers continue to need custom chips and networking gear?
Tan’s views on governance and safeguards for AI use are more nuanced than those of his CEO colleagues at Anthropic and OpenAI. While he agrees with Amodei about the need for restrictions, he downplays the risks of unchecked AI growth, likening it to a “live animal” that won’t run wild by itself.
The Broadcom-Anthropic relationship is noteworthy: Tan forecasts that Anthropic will become his company’s largest custom chip customer in 2027 and maintain that status through 2028. This dependence on a single customer highlights the risks of relying too heavily on a small set of high-profile clients.
As investors weigh their options, they should consider the broader implications of Broadcom’s bet on AI inference. If Tan is correct about long-term demand for compute infrastructure, his company will likely reap significant rewards. However, if Amodei and others are right about the need for moderation, Broadcom’s growth trajectory may be at risk.
The tech industry’s reliance on a small group of influential players has always been precarious. The current debate highlights the need for more nuanced thinking about AI development and deployment. Tan’s views offer a counterpoint to Amodei’s calls for restraint, but they also underscore the complexity of this issue.
Broadcom’s success will depend on its ability to adapt to changing circumstances – and convince investors that their bet on AI inference is not just a gamble, but a sound investment in human progress.
Reader Views
- QSQuinn S. · senior engineer
While Broadcom's Hock Tan is right to be bullish on AI compute demand, his company's forecasted $230 billion in revenue by 2028 relies on an increasingly uncertain premise. If Anthropic's Dario Amodei and others succeed in slowing frontier model development, companies like Broadcom risk overextending themselves with costly custom chip investments that might not materialize as anticipated. Tan's optimism assumes AI adoption follows a linear path, whereas the truth is likely more complex – requiring closer examination of market dynamics and potential for value capture beyond compute infrastructure alone.
- TSThe Stack Desk · editorial
While Broadcom CEO Hock Tan's confidence in his company's AI revenue targets is admirable, he seems to be playing with fire by counting on unchecked growth in compute demand. The potential risks of autonomous AI development are too real to ignore, and investors should be asking more questions about Broadcom's business model sustainability. It's also worth noting that the company's forecasted $230 billion revenue target in 2028 relies heavily on Anthropic's continued reliance on custom chips - a symbiotic relationship that could quickly turn toxic if industry dynamics shift.
- AKAsha K. · self-taught dev
It's ironic that Hock Tan is downplaying the risks of unchecked AI growth by likening it to a live animal that won't run wild on its own. Meanwhile, his company's business model heavily relies on this same technology accelerating without regulation. Broadcom's reliance on custom chip customers like Anthropic raises concerns about their ability to adapt if AI development does slow down. Tan needs to address how his company will maintain profitability in the face of potentially dwindling compute demand. The tech industry can't afford executives who are more focused on growth targets than responsible innovation.