An analyst projects Broadcom will generate $230 billion in AI-related revenue by 2028, primarily fueled by custom silicon and networking demand from OpenAI and Anthropic.
Demand for AI infrastructure is outpacing supply as major artificial intelligence labs accelerate development of custom accelerators, strengthening Broadcom’s position as a leading alternative to Nvidia in the sector. The semiconductor giant shows little sign of slowing, with CEO Hock Tan projecting that AI chip revenue will double in each of the next two years as demand from customers including Anthropic and OpenAI intensifies.
Based in Palo Alto, California, Broadcom expects AI semiconductor revenue to reach approximately $115 billion in fiscal 2027, up from roughly $58 billion projected for 2026. The company anticipates another doubling to $230 billion in fiscal 2028, underscoring the momentum of its custom AI accelerator and networking chip business. During the third quarter, AI chip sales more than tripled to $16.7 billion, lifting total revenue 86% year over year to $29.59 billion and beating consensus estimates of $29.36 billion. Adjusted profit came in at $3.32 per share, surpassing estimates of $3.24.
Despite the robust quarterly results, Broadcom’s fourth-quarter revenue forecast—which implies 93% growth—fell short of Wall Street expectations. The miss pressured AVGO shares, which declined 1.3% in overnight trading late Wednesday. Addressing the outlook, Tan stated: “In 2027, we have secured the supply to again double AI revenue to approximately $115 billion. Our demand actually exceeds this outlook, and we will work to improve supply.” He added regarding the following year: “In 2028… we have line of sight for fiscal 2028 AI semiconductor revenue growth to again double to $230 billion. Here again, we have secured the supply to meet this outlook.”
Broadcom has emerged as one of the primary beneficiaries of the AI infrastructure boom beyond Nvidia, with its custom accelerators, or XPUs, gaining traction among major AI firms building proprietary silicon. Anthropic is expected to become Broadcom’s largest XPU customer in 2027, planning to deploy an additional 5 gigawatts of Broadcom-designed TPU v8i chips that year, followed by another 10 gigawatts in 2028. OpenAI is also poised to become a major client, with Broadcom forecasting the ChatGPT developer could deploy more than 5 gigawatts of its next-generation Jalapeno XPU in 2028. To support these targets, Broadcom has secured supplies for wafers, substrates, and memory, and plans to begin deploying its Singapore substrate facility in fiscal 2027 to alleviate a critical supply bottleneck.
Tan emphasized the strategic importance of these partnerships, noting: “These two guys, Anthropic and OpenAI, I mean, these are… it’s like you have two geniuses in the middle of outer Mongolia. They need to go to college to fulfill their potential,” suggesting that while both companies possess enormous potential, they require access to Broadcom’s chips and infrastructure (“college”) to fully realize it. He continued: “We do what we can to help them, and part of it is creating sources of financing to help these companies with the leading-edge frontier models in the world be able to play in the same playing field and be able to offer these great technology products to the world.”
Industry analysts acknowledged the strong execution while flagging lingering challenges. Patrick Moorhead, CEO and principal analyst at Moor Insights & Strategy, described Broadcom’s third quarter as “strong, but not flawless,” arguing that the key takeaway was successful AI execution rather than the modest revenue beat. He pointed to the $16.7 billion in Q3 AI revenue and $21.7 billion Q4 outlook as proof that the company is effectively converting its AI backlog into sales. Moorhead also highlighted Broadcom’s comprehensive AI infrastructure advantages, spanning custom XPU co-design, Ethernet switching, SerDes, optics, packaging, and foundry access. However, he cautioned that the results “do not prove broad NVIDIA displacement.” Identifying remaining headwinds, Moorhead noted: “The wrinkle is software, and the release does not resolve customer concentration, Google multisourcing, bookings conversion, AI networking mix, first-generation XPU yields or the fiscal 2027 AI bridge.”
Market participants reacted cautiously optimistically to the guidance. Gary Black, Managing Director at Future Fund, observed an improvement in AVGO’s after-hours trading “as investors responded favorably to the company’s AI rev growth guidance.” He added, “As the AI Rev strength becomes clearer to investors, I expect the stock to recover somewhat.” On retail platforms, AVGO became the top trending ticker on Stocktwits, with sentiment shifting from “bullish” to “extremely bullish.” Traders expressed strong conviction, with one stating: “$AVGO Broadcom just posted the most massive financial results in company history — bears can say whatever they want — the share price will go higher regardless.” Another wrote: “$AVGO it takes time for the earnings results and outlook to filter into the market; this forward outlook is insane… The stock is too cheap, just like Nvidia. It’s going higher.”
Notwithstanding the positive sentiment, AVGO stock declined during the two trading sessions preceding the Q3 report and remains up just 6.5% year to date. By comparison, the iShares Semiconductor ETF (SOXX) has gained over 65% over the same period.