Broadcom (AVGO.US) shares declined in after-hours trading following its Q4 earnings release, as the chipmaker's AI backlog figures slightly disappointed investors.
For the fiscal fourth quarter ended November 2, Broadcom reported revenue of $18 billion, up 28% year-over-year and beating analyst estimates of $17.5 billion. Non-GAAP earnings per share rose to $1.95, surpassing the $1.87 consensus. Semiconductor segment revenue jumped 34.5% to $11.07 billion, while infrastructure solutions grew 19% to $6.94 billion.
Despite the strong results, shares fell over 5% after CEO Hock Tan revealed the company currently holds $73 billion in AI product backlog for delivery over the next six quarters—a figure some investors found underwhelming. Tan later clarified this represents a "floor," stating: "We expect significant upside as more orders come in. Lead times could range from six months to a year depending on specific products."
The CEO disclosed Broadcom secured $11 billion in Q4 orders from AI startup Anthropic PBC, building on Q3's $10 billion commitment. Another $1 billion customer deal was signed, though the buyer wasn't named. However, Tan warned that AI product sales are compressing overall margins.
For fiscal Q1 ending February 1, Broadcom guided revenue of approximately $19.1 billion versus the $18.5 billion consensus. AI semiconductor revenue is projected to double year-over-year to $8.2 billion. The company raised its quarterly dividend by 10% to $0.65 per share.
Broadcom's stock has surged 75% year-to-date as investors bet on its AI growth potential, pushing its forward P/E to about 42x—far above its 10-year average of 17x. The valuation now exceeds all "Magnificent Seven" tech stocks except Tesla, suggesting even strong earnings may trigger profit-taking.
The company continues gaining share in Nvidia-dominated (NVDA.US) AI chips through custom silicon designs for hyperscalers. Q4 AI chip revenue grew over 70%, accelerating by more than 10 percentage points sequentially. Recent partnerships with major AI players like OpenAI (using Broadcom's custom chips and networking components) and Anthropic (leveraging Google Cloud TPUs containing Broadcom designs) have fueled optimism.
Analysts note Broadcom's diversified portfolio—spanning communications chips, networking gear, and software—positions it well for AI infrastructure buildouts. Its upgraded networking equipment addresses critical data transfer needs as AI models grow more complex. Custom AI processor collaborations with Google and Meta Platforms (META.US) provide alternatives to Nvidia's GPUs.
CEO Tan stands to receive 610,521 Broadcom shares if AI revenue hits $90 billion by fiscal 2030—a long-term target reflecting the company's ambitious AI roadmap.
Comments