Broadcom is set to report earnings in the early hours of Thursday; could it lead a rebound in semico
While the market is still debating whether general-purpose GPUs are sufficient, hyperscalers have already voted with their orders: custom AI accelerators and high-speed Ethernet are the true battlegrounds for the next phase of AI infrastructure.$Broadcom (AVGO.US)$ The "chip arms dealer" behind these tech giants is set to release its fiscal 2026 Q3 earnings report after the market close on September 2 (Eastern Time).Institutions expect Broadcom to generate revenue of $29.435 billion in fiscal 2026 Q3, an 84.52% year-over-year increase; EPS is projected at $2.55, up 199.53% year-over-year.Rather than simply chasing general-purpose computing power, Broadcom's core investment highlights are more focused on AI ASIC/XPU custom chips, Ethernet chips, and infrastructure software centered around VMware.
![While the market is still debating whether general-purpose GPUs are sufficient, hyperscalers have already voted with their orders: custom AI accelerators and high-speed Ethernet are the true battlegrounds for the next phase of AI infrastructure.$Broadcom (AVGO.US)$ The "chip arms dealer" standing behind these tech giants will release its fiscal 2026 Q3 earnings report after the U.S. market close on September 2.Institutions expect Broadcom to achieve revenue of $29.435 billion in fiscal 2026 Q3, an 84.52% year-over-year increase; expected EPS is $2.55, a 199.53% year-over-year increase.Rather than simply chasing general-purpose computing power, Broadcom's core investment highlights are more focused on AI ASIC/XPU custom chips, Ethernet chips, and infrastructure software centered around VMware.[Clap] 🔎 There are three core highlights to watch regarding Broadcom this time: First, can AI custom chips remain in a state of "excess demand"? According to Broadcom's official earnings report and conference call disclosures, AI semiconductor revenue grew significantly year-over-year in the previous quarter, with strong and expanding visibility into AI orders. Management explicitly stated that demand for AI semiconductors remains "unmet." The company is advancing multi-generation xPU, TPU, and AI networking collaborations with several tech companies including Google, Anthropic, OpenAI, and Meta, with custom AI accelerators becoming a core growth engine. For Broadcom, the key focus is not just on how much can be delivered this quarter, but also on the next...](https://nnqimage.futunn.com/sns_client_feed/999982/20260831/web-1788139102828-9QGouFCX3N.png/big?area=2&is_public=true&imageMogr2/ignore-error/1/format/webp)
🔎 There are three key points to watch regarding Broadcom this time:
First, can demand for AI custom chips remain "outstripping supply"?
According to Broadcom's official earnings report and conference call disclosures, AI semiconductor revenue saw significant year-over-year growth last quarter, with strong and expanding visibility into AI orders. Management explicitly stated that demand for AI semiconductors remains "unsatisfied." The company is advancing multi-generation xPU, TPU, and AI networking collaborations with tech firms such as Google, Anthropic, OpenAI, and Meta, making custom AI accelerators a core growth engine. For Broadcom, the focus is not just on delivery volume this quarter, but also on the client matrix and order depth for the next phase of custom chips.
Second, can networking chips become the "high-speed nervous system" of AI clusters?
AI training and inference are not just a race for computing power; how data flows at high speed within high-performance clusters is equally critical. According to Broadcom's product announcements, the Tomahawk 6 3nm switching chip supports industry-leading throughput and is targeted at AI/ML, general-purpose cloud, and enterprise data centers, making it one of the top-performing Ethernet switching chips globally. Beyond the general GPU ecosystem, the combination of "Ethernet interconnects + custom accelerators" is becoming Broadcom's key differentiator in competing for AI infrastructure.
Third, how will software and "new computing infrastructure" unlock long-term growth potential?
According to Broadcom's earnings press release, revenue from the infrastructure software business continues to grow, with VMware Cloud Foundation focusing on enhancing infrastructure efficiency and security while supporting enterprise AI inference workloads. Meanwhile, Broadcom is partnering with investors like Apollo and Blackstone to establish an AI SPV platform, planning to deploy large-scale computing capacity over the coming years to support cutting-edge AI labs such as OpenAI and Anthropic. This signifies that Broadcom is evolving from a chip supplier into an integrator of AI computing infrastructure.
This time, the market's focus may not just be on Broadcom's quarterly results, but also on whether its custom AI chips and networking chips can continue to support the next phase of its growth narrative.
So, the question arises:
✅ What is the visibility of AI semiconductor orders and the level of customer concentration?
✅ What impact will supply chain diversification by core customers like Google have?
✅ Can the recovery in non-AI semiconductors sustain itself?
✅ Can VMware software continue to serve as a cornerstone of profitability?
✅ Regarding the AI SPV platform jointly established with Apollo and Blackstone, what are the funding and delivery timelines?
🏆 Activity 1: Bull Circle Insight Champion ✨
Broadcom's earnings report is making a major splash. With dual engines driven by custom AI chips and networking chips, will it bolster confidence in AI infrastructure after the results are released?
Share your thoughts in the comments below:
✅ At least 30 words ✅ Original viewpoint ✅ Complies with community guidelines
Meet the requirements to participate and share an equal split of 10,000 points!
🏆 Activity 2: Options Strategy Exchange ✨
Stock volatility tends to intensify around earnings announcements—creating a strategic window for options positioning.How do you plan to trade? Share your options strategy in the comments below!Write at least 30 characters with sound reasoning to earn 66 points(Points stack on top of those from Activity 1)
Note: All activities above will end at 04:00 Beijing Time on September 3; rewards will be distributed uniformly after the conclusion of this earnings season.
Click to reserve:
![While the market is still debating whether general-purpose GPUs are sufficient, hyperscalers have already voted with their orders: custom AI accelerators and high-speed Ethernet are the true battlegrounds for the next phase of AI infrastructure.$Broadcom (AVGO.US)$ The "chip arms dealer" standing behind these tech giants will release its fiscal 2026 Q3 earnings report after the U.S. market close on September 2.Institutions expect Broadcom to achieve revenue of $29.435 billion in fiscal 2026 Q3, an 84.52% year-over-year increase; expected EPS is $2.55, a 199.53% year-over-year increase.Rather than simply chasing general-purpose computing power, Broadcom's core investment highlights are more focused on AI ASIC/XPU custom chips, Ethernet chips, and infrastructure software centered around VMware.[Clap] 🔎 There are three core highlights to watch regarding Broadcom this time: First, can AI custom chips remain in a state of "excess demand"? According to Broadcom's official earnings report and conference call disclosures, AI semiconductor revenue grew significantly year-over-year in the previous quarter, with strong and expanding visibility into AI orders. Management explicitly stated that demand for AI semiconductors remains "unmet." The company is advancing multi-generation xPU, TPU, and AI networking collaborations with several tech companies including Google, Anthropic, OpenAI, and Meta, with custom AI accelerators becoming a core growth engine. For Broadcom, the key focus is not just on how much can be delivered this quarter, but also on the next...](https://nnqimage.futunn.com/sns_client_feed/999982/20260828/web-1787900012833-zWbmwNtXML.webp/big?area=2&is_public=true&imageMogr2/ignore-error/1/format/webp)
![While the market is still debating whether general-purpose GPUs are sufficient, hyperscalers have already voted with their orders: custom AI accelerators and high-speed Ethernet are the true battlegrounds for the next phase of AI infrastructure.$Broadcom (AVGO.US)$ The "chip arms dealer" standing behind these tech giants will release its fiscal 2026 Q3 earnings report after the U.S. market close on September 2.Institutions expect Broadcom to achieve revenue of $29.435 billion in fiscal 2026 Q3, an 84.52% year-over-year increase; expected EPS is $2.55, a 199.53% year-over-year increase.Rather than simply chasing general-purpose computing power, Broadcom's core investment highlights are more focused on AI ASIC/XPU custom chips, Ethernet chips, and infrastructure software centered around VMware.[Clap] 🔎 There are three core highlights to watch regarding Broadcom this time: First, can AI custom chips remain in a state of "excess demand"? According to Broadcom's official earnings report and conference call disclosures, AI semiconductor revenue grew significantly year-over-year in the previous quarter, with strong and expanding visibility into AI orders. Management explicitly stated that demand for AI semiconductors remains "unmet." The company is advancing multi-generation xPU, TPU, and AI networking collaborations with several tech companies including Google, Anthropic, OpenAI, and Meta, with custom AI accelerators becoming a core growth engine. For Broadcom, the key focus is not just on how much can be delivered this quarter, but also on the next...](https://nnqimage.futunn.com/sns_client_feed/999982/20260828/web-1787900026867-Itys5orO6V.webp/big?area=2&is_public=true&imageMogr2/ignore-error/1/format/webp)
Risk Disclaimer: The above content only represents the author's view. It does not represent any position or investment advice of Futu. Futu makes no representation or warranty.Read more
Comments (80)
to post a comment
26
8
