Broadcom Inc.

NASDAQ Global Select
Back to all articles
Somewhat Bullish +45

Broadcom AI Chip Forecast Hits $115B Amid China Probe

AI integration Regulatory investigation AI-generated analysis

📈 Broadcom raised its fiscal 2027 AI-chip revenue forecast to roughly $115 billion and expects the figure to reach approximately $230 billion by fiscal 2028.

🤝 The company has visibility into over 10 gigawatts of future AI infrastructure commitments from key customers including Anthropic, OpenAI, Meta, Google, and others.

🇨🇳 Chinese authorities are examining Broadcom hardware deployed inside state-backed data centers, raising potential geopolitical risks regarding data sovereignty and supply chain dependency.

💻 Broadcom's business model focuses on custom ASICs designed jointly with hyperscalers for specific inference workloads, distinct from Nvidia's merchant GPU approach.

🏗️ The company is capitalizing on the structural shift in AI compute where hyperscalers prefer cost-efficient custom chips over general-purpose GPUs for stable inference loads.

🚀 Competitors like Qualcomm are entering the data center market with new AI200 and AI250 chips, while startups like Eliyan are solving interconnect bottlenecks.

📉 The review of Broadcom hardware in China represents a rare case where Chinese authorities might restrict an American chip designer's domestic presence rather than just reacting to US export limits.

🔍 Analysts expect Broadcom to face direct questions about the China review on its next earnings call, likely resulting in carefully worded responses that do not confirm specific restrictions.

Bullish Signals
  • Broadcom raised its fiscal 2027 AI-chip revenue forecast to approximately $115 billion and expects revenue to nearly double to $230 billion by fiscal 2028.
  • The company has secured visibility into more than 10 gigawatts of future AI infrastructure tied to major customers including Anthropic, OpenAI, and Meta.
  • Broadcom's custom ASIC business model allows it to capture value from specific customer inference workloads, effectively competing with Nvidia in the growing inference market.
  • The company is benefiting from a structural shift where hyperscalers are moving away from general-purpose GPUs toward cost-efficient custom chips for stable inference loads.
Risk Factors
  • Chinese authorities are examining Broadcom hardware deployed inside state-backed data centers, creating potential geopolitical risks regarding data sovereignty and supply chain dependency.
  • The review of Broadcom hardware in China represents a unique risk where Chinese authorities might restrict an American chip designer's domestic presence rather than just reacting to US export limits.
Full Analysis
Broadcom (AVGO) has significantly raised its fiscal 2027 AI-chip revenue forecast to approximately $115 billion, up from a prior estimate of over $100 billion, with expectations for the figure to nearly double to $230 billion by fiscal 2028. CEO Hock Tan highlighted strong visibility into more than 10 gigawatts of future AI infrastructure tied to major customers including Anthropic, OpenAI, and Meta, alongside confirmed production deployments at Google and Meta. The company's growth strategy relies on a custom ASIC business model where it partners with hyperscalers to design application-specific chips for stable inference workloads, rather than selling general-purpose merchant GPUs. This approach allows Broadcom to capture value from specific customer needs, such as ranking and recommendation inference, effectively competing with Nvidia in the growing inference market while maintaining its role in training. However, the company faces a significant geopolitical headwind as Chinese authorities are reportedly examining Broadcom hardware deployed inside state-backed data centers. This review, reported by Reuters on September 24, 2026, touches on data sovereignty and supply-chain dependency issues, potentially restricting an American chip designer's presence in domestic Chinese infrastructure and creating a unique risk profile distinct from standard export controls. Broadcom operates in a rapidly evolving ecosystem where competition is intensifying. Rivals like Qualcomm are entering the data center market with memory-heavy architectures aimed at inference cost efficiency, while startups like Eliyan are addressing critical chiplet interconnect bottlenecks. As hyperscalers diversify their supply chains to avoid reliance on a single supplier for AI compute, Broadcom's custom silicon share is expected to grow structurally, though it faces scrutiny regarding its strategic importance in global compute infrastructure.