Amid the global wave of artificial intelligence, Broadcom (AVGO), as a core supplier of data center custom chips and networking products, is playing an indispensable role. Its robust financial performance has driven a soaring stock price, with a year-to-date increase of 54% as of October 13. Extending the timeline, since the AI boom began at the end of 2022, the stock has surged by over 500%. However, there is ample reason to believe that Broadcom’s growth story for 2026 and beyond is far from over.
Broadcom has long maintained steady earnings growth, delivering market-beating returns for investors. This trend has been particularly pronounced over the past year, with its free cash flow growth continuously accelerating. In the first three quarters of fiscal year 2025, the company’s free cash flow increased significantly by 40% year-over-year, fully demonstrating that its profitability is rising in tandem with the strong growth of its AI accelerators and software business.
Providing even greater confidence to the market is its record order backlog—reaching $110 billion, far exceeding its $60 billion revenue over the past 12 months. This figure provides exceptionally high visibility for its revenue for the coming years. Against the backdrop of projected spending by hyperscalers on AI infrastructure reaching $350 billion this year, Broadcom, as a critical link in the industry chain, is undoubtedly poised to continue benefiting. Ample cash flow not only supports its continued investment in innovation but also lays a solid foundation for rewarding shareholders.
While NVIDIA’s GPUs have become the benchmark for AI training due to their powerful computing capabilities, Broadcom has carved out another vast arena through its unique strategic path. Beyond the essential Ethernet switching and networking products for data centers, its biggest opportunity lies in custom Application-Specific Integrated Circuits (ASICs), also known as XPUs.
These custom AI accelerators are designed for specific tasks. While they are less versatile than GPUs, they offer superior performance in execution efficiency and energy consumption control. Faced with increasingly high AI computing costs, many enterprises are willing to accept this trade-off, thereby driving strong demand for Broadcom’s ASIC chips. This is corroborated in its latest earnings report: third-quarter revenue hit a record $15.9 billion, a 22% year-over-year increase, with AI-related revenue surging 63% year-over-year to $5.2 billion, becoming the strongest growth engine.
Broadcom’s management is highly optimistic about the future outlook. CEO Hock Tan confirmed during the earnings call that the company’s share with three of its original major hyperscale customers continues to grow and raised expectations for next year’s AI business. A more significant signal is that Broadcom has added a fourth major hyperscale customer (widely speculated to be OpenAI), upgrading it to a “qualified customer,” with related AI racks already in production. This development directly adds another $10 billion to its substantial order backlog.
With the stock price soaring, market attention inevitably turns to its valuation. Broadcom’s current price-to-earnings ratio of approximately 88 may seem high, but when combined with its growth prospects, its more accurate Price/Earnings-to-Growth ratio is only 0.37, which is typically considered a sign of an undervalued stock. Simultaneously, its price-to-sales ratio also falls within a reasonable range. Considering the company’s vast market opportunities, solid customer orders, and its unique, hard-to-replace position within the AI ecosystem, Broadcom already possesses the full conditions to continue profiting from the long-term development of artificial intelligence. Its future growth trajectory remains clear and full of potential.
