Guozhen AIGlobal AI field notes and model intelligence

Realtime AI News

Broadcom beats Q3 estimates and forecasts AI semiconductor sales to quadruple within two years

Broadcom beat market expectations in its latest quarter, with revenue up 86% year over year to $29.59 billion on the strength of AI semiconductors. The company now expects AI chip sales to roughly double to about $115 billion in fiscal 2027 and reach $230 billion in fiscal 2028, with Anthropic and OpenAI custom-chip deals at the center of that growth.

Published
博通Q3业绩超预期,预计AI半导体销售额两年扩大4倍:Anthropic与OpenAI成新增长引擎
Image source: broadcom.cn

Broadcom turned in third-quarter results that beat market expectations, powered by demand for AI semiconductors, and laid out a forecast calling for roughly quadrupling AI chip sales within two years, according to South Korea's Maeil Business Newspaper. Analysts cited in the report say that after Google, Anthropic and OpenAI are speeding up in-house AI chip development, rapidly expanding Broadcom's position in an AI semiconductor market long dominated by Nvidia.

On the numbers, third-quarter revenue rose 86% year over year to $29.59 billion, above the $29.36 billion analysts expected. Net income came in at $13.09 billion, more than triple the $4.14 billion a year earlier, and earnings per share of $3.32 topped the $3.24 consensus. AI semiconductors drove the growth, with quarterly revenue of $16.7 billion exceeding market forecasts.

For the near term, Broadcom guided fourth-quarter AI semiconductor revenue to $21.7 billion, while its total fourth-quarter revenue guidance of $34.8 billion came in slightly below the roughly $35 billion market consensus.

The long-term outlook is far more aggressive. CEO Hock Tan said on the post-earnings call that AI semiconductor sales should rise to about $115 billion in fiscal 2027, double, then double again to $230 billion in fiscal 2028. He added that the company has six customers, four of which will be very large, and that demand currently exceeds supply capacity.

Anthropic and OpenAI sit at the center of that growth. Broadcom expects to supply Anthropic with AI chips at roughly 5GW scale in 2027 and an additional 10GW in 2028. Anthropic is expected to surpass Google as Broadcom's largest custom AI chip customer next year. The OpenAI partnership is also scaling quickly: the co-developed Jalapeño AI accelerator is slated to reach about 1.3GW in 2027, with first- and second-generation chips together exceeding 5GW by 2028, and OpenAI is close to finishing its second-generation chip design while preparing a third-generation product with Broadcom.

Google and Meta remain core customers. Broadcom plans to supply Google with custom processors worth tens of billions of dollars annually for the coming years, and to deliver Meta a third-generation custom chip by the end of 2027 along with volume production of Meta's own MTIA accelerator for AI inference and recommendation workloads.

The report frames the custom-chip push as big AI companies trying to reduce dependence on Nvidia GPUs and improve cost and power efficiency, a market Broadcom captures by designing chips to customer specifications and organizing production. Tan claimed that the latest chip co-developed with Google performs equal to or better than Nvidia's next-generation Vera Rubin when running AI models.

What to watch: whether the 5GW Anthropic deliveries land on schedule in 2027, how quickly OpenAI's Jalapeño scales, and whether Google keeps expanding its orders. If the $230 billion target materializes, Broadcom becomes the most important AI compute supplier outside Nvidia, and the clearest signal yet that frontier labs want their own silicon.

Why it matters

Broadcom's guidance extends the custom-silicon story from Google to Anthropic and OpenAI; if the fiscal 2028 target holds, the AI compute supply chain gains a credible second pole beside Nvidia, and frontier labs' chip autonomy takes another step forward.

BroadcomAI chipsAnthropicOpenAI
Back to realtime news

Nearby Updates

All

09/03, 08:54

Tech CEOs and Trump officials push G20 for lighter AI rules at Chapel Hill ministerial

Top U.S. tech CEOs and Trump administration officials pressed the G20 Innovation Ministerial in Chapel Hill on September 2 to embrace lighter AI regulation, arguing that excessive restrictions would slow innovation. G20 ministers reached consensus on the U.S.-backed Carolina Principles for Emerging Technologies, even as the gathering underscored a widening U.S.-Europe split over AI governance.

09/03, 09:46

Zhaogang and Tencent WorkBuddy unveil NextB2B, China's first B2B trade AI agent

Zhaogang Group has teamed up with Tencent WorkBuddy to launch NextB2B, billed as China's first AI agent for the B2B trading industry. The release pairs an industrial trading platform with an enterprise agent product, signaling that AI agents are moving into core transaction workflows.

09/03, 09:54

Token consumption surges, AI agents go live in batches: the AI race in listed banks' half-year reports

Listed banks' half-year reports are exposing a new competitive front: AI, with token consumption surging and AI agents going live in batches, according to a Red Net report. The disclosures suggest large-model usage inside Chinese banks is shifting from pilots toward measurable, production-scale workloads.

09/03, 09:59

Doubao returns to phones as ByteDance bets on AI agent devices

ByteDance's AI assistant Doubao is heading back into the phone market in the form of AI agent devices, according to a Pandaily report published on September 3. The move signals that ByteDance wants its assistant to go beyond apps and compete at the level of dedicated agent hardware.