Friday, September 11, 2026
AI 인프라 · 뉴스 & 분석
반도체·하드웨어리포트
반도체·하드웨어 · 리포트

Broadcom은 AI 실리콘 수요에 힘입어 강력한 3분기 실적을 보고하며, 지속적인 두 자릿수 매출 및 이익 성장에 대한 가이던스를 제시했다.

맞춤형 AI 가속기 및 네트워킹 장비에 대한 하이퍼스케일러와 뉴클라우드의 지속적인 지출을 확인하여 장기적인 인프라 자본지출 사이클을 입증함.
업계 전문지Slicast · September 5, 2026 · 글로벌 · 출처: Data Center Dynamics
중요도 80

Broadcom posted strong, silicon-fueled results for the third quarter of fiscal 2026—and forecast further growth—overshadowing robust earnings from its VMware-led Infrastructure Software division. Overall Q3 revenues surged 86 percent year-on-year (YoY) to $29.6 billion, driven by a 127 percent increase in the Semiconductor Solutions division, which totaled $20.8 billion. Infrastructure Software operations posted a more modest 29 percent YoY increase to nearly $8.8 billion, with the segment’s share of total revenues declining from 43 percent last year to 30 percent this year.

On a generally accepted accounting principles (GAAP) basis, overall net income surged more than 200 percent to nearly $13.1 billion. Broadcom generated $14.2 billion in cash over the three-month period while deploying just $500 million in capital expenditures.

The company’s revenue trajectory remains heavily anchored to AI infrastructure. CEO Hock Tan stated that Broadcom expects AI networking revenues to double in fiscal 2027 to $115 billion before doubling again in fiscal 2028 to $230 billion. “We have secured the supply to meet this outlook,” Tan told investors during the earnings call. “This AI revenue guidance through [2028] is being provided to give you the trajectory of our growth, and that demand for compute continues to be extremely strong. As a result, I have to say we are very much on target to exceed $30 in earnings per share in fiscal 2028.”

Tan highlighted ongoing AI data center projects, including a recently expanded agreement with Google to develop and supply future Tensor Processing Units (TPUs) and AI networking equipment. Work with Anthropic includes the upcoming one-gigawatt (1GW) Ironwood facility, another five-gigawatts of TPU Version 8i scheduled for fiscal 2027, and a “clear line of sight to deliver another incremental 10 gigawatts.” “Even as we expect Google to grow for us, Anthropic is on track to become our largest XPU customer in 2027, and sustain that in 2028,” Tan added.

Regarding OpenAI, Tan noted that Broadcom’s work on the AI data movement-optimized “Jalapeño” chip remains on track for a planned 1.3-gigawatt deployment in 2027. “Together with OpenAI, we are deep in development of the next-generation XPU beyond Jalapeño, which is approaching tape-out in 2028,” Tan said. “We have line of sight for OpenAI to deploy over five gigawatts of Jalapeño and its successor generation of XPU, which would make OpenAI our second largest XPU customer.”

Addressing investor concerns regarding supply chain constraints, Tan characterized Broadcom’s forecasts as “conservative,” noting that demand continues to outpace supply. “We have secured supply, and we think it's the right number to put it to [$115 billion], and if times, if circumstances change and our ability to scale more supply chain, of course, we will uplift, but at this point that's our best outlook,” Tan said. “The same thinking applies to 2028, when we give you that outlook of $230 billion, this is real demand we believe based on what's available, what data center sites, locations are ready [in] 2028 with respect to our customers, the size of what we have and against the supply chain we have in leading-edge wafers, substrates, and [high-bandwidth memory]. … So this is, again, a carefully structured outlook that we believe we can achieve.”

Broadcom’s networking division also drew praise, with Tan positioning the vendor as the “leader in AI networking,” supported by its 100 Tbps Tomahawk 6 switch and the recent tape-out of its 200 Tb/s Tomahawk 7. The company continues to advance its optical digital signal processors (DSPs), electro-absorption modulated lasers (EMLs), and continuous-wave (CW) indium phosphide (InP) lasers. “We continue to invest and invest heavily to provide the broadest and most leading-edge AI portfolio,” Tan added. “In fact, our AI networking revenue is expected to grow just as fast as XPUs over the next few years, reflecting our excellent progress with this key group of [large language model] customers.”

Outside of AI, Broadcom’s non-AI semiconductor business reported a modest 5 percent YoY revenue increase to $4.2 billion for the quarter, remaining flat sequentially. Tan forecast a similar 5 percent YoY increase for the segment in the coming quarter.

Finally, Tan provided a brief update on the VMware-linked Infrastructure Software business, noting it is tracking at a 15 percent annual recurring revenue (ARR) growth rate. While actual revenues are expected to dip slightly to $8.7 billion next quarter, he emphasized that VMware’s newly launched Private AI Cloud platform will help capture expanding “enterprise consumption of AI,” opening “a new opportunity for our Infrastructure Software business.”

원문 보기