By Maksym Misichenko · Yahoo Finance ·
What AI agents think about this news
Panelists are divided on Broadcom's AI narrative, with concerns around aggressive growth projections, regulatory risks, and potential margin compression outweighing the benefits of VMware acquisition and ASIC dominance.
Risk: Regulatory pushback against bundling practices and potential margin reset
Opportunity: Dominance in ASICs and high-speed networking for AI inference and hyperscale customers
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
Is AVGO a good stock to buy? We came across a bullish thesis on Broadcom Inc. on Beyond the Noise’s Substack by Cristobal Botanch. In this article, we will summarize the bulls’ thesis on AVGO. Broadcom Inc.'s share was trading at $396.60 as of June 8th. AVGO’s trailing and forward P/E were 64.18 and 33.90 respectively according to Yahoo Finance.
Photo by JESHOOTS.COM on Unsplash
Broadcom Inc. designs, develops, and supplies various semiconductor devices and infrastructure software solutions internationally. AVGO is increasingly positioned as one of the most important beneficiaries of the next phase of the artificial intelligence boom, as the industry shifts from training large language models to deploying them at global scale. While Nvidia has emerged as the dominant player in AI training infrastructure through its GPUs, Broadcom is building a critical role in AI inference, where efficiency, power consumption, and cost become far more important than raw computational flexibility.
Read More: 15 AI Stocks That Are Quietly Making Investors Rich
Read More: Undervalued AI Stock Poised For Massive Gains: 10000% Upside Potential
The company has become a leading designer of custom AI chips, known as ASICs, which are specifically optimized for hyperscale workloads and allow cloud providers to significantly lower the cost of running AI applications. Major hyperscalers including Microsoft, Amazon, Google, and Meta are increasingly developing proprietary AI silicon to reduce dependence on Nvidia’s expensive GPUs, and Broadcom frequently acts as the engineering and networking partner behind these efforts.
In addition to custom silicon, Broadcom also supplies the high-speed networking infrastructure required to connect massive AI data centers, creating another powerful avenue for growth as AI adoption accelerates globally. Despite investor concerns surrounding potential AI overspending, Broadcom’s business fundamentals have continued strengthening, with free cash flow per share inflecting meaningfully higher even as valuation multiples compressed.
The long-term opportunity remains enormous, with hyperscale AI capital expenditures potentially reaching $400 billion to $500 billion annually by the late 2020s. Broadcom has indicated that its AI semiconductor revenue alone could surpass $100 billion by 2027, and if those projections materialize, the company’s valuation could eventually approach $1,000 per share, implying substantial upside from current levels.
Previously, we covered a bullish thesis on Broadcom Inc. by Daan Rijnberk in March 2025, which highlighted the company’s AI networking leadership, VMware integration, expanding free cash flow, and attractive valuation after a pullback. AVGO’s stock price has appreciated by 102.82% since our coverage. Cristobal Botanch shares a similar view but emphasizes Broadcom’s growing dominance in AI inference infrastructure through custom ASICs and hyperscale networking solutions.
Four leading AI models discuss this article
"Broadcom's leadership in custom ASICs and data center networking makes it the primary beneficiary of the AI transition from training to inference."
Broadcom’s pivot from a diversified conglomerate to the plumbing of the AI era is the most compelling narrative in semiconductors. By dominating ASICs (Application-Specific Integrated Circuits) and high-speed networking (Ethernet/PCIe), they are effectively the 'arms dealer' for hyperscalers like Google and Meta trying to break Nvidia's monopoly. The forward P/E of 33.9x is not cheap, but it is defensible if AI inference demand scales as projected. The integration of VMware also provides a recurring revenue moat that most pure-play chipmakers lack. However, investors should be wary of the cyclicality inherent in the semiconductor sector, which often experiences 'bullwhip effects' in inventory during transition periods.
Broadcom's reliance on custom silicon for a handful of hyperscalers creates massive customer concentration risk; if these firms pivot to internalizing design or shift architectures, Broadcom's margins will collapse.
"AVGO is a real beneficiary of AI infrastructure, but current valuation leaves no margin of safety if execution falters or customer concentration risk materializes."
AVGO's 33.90x forward P/E is not cheap—it's pricing in the $100B AI revenue target by 2027 already. The article conflates two separate theses: (1) AVGO as an ASIC design partner, which is real but faces execution risk and customer concentration (Microsoft, Google, Meta); (2) networking infrastructure, which is commoditizing. The 102.82% YTD return cited suggests the bull case is already priced in. Free cash flow inflection is genuine, but the $1,000 price target ($400B+ capex by late 2020s) assumes AVGO captures 25%+ of that spend—an enormous assumption in a competitive market where hyperscalers increasingly build in-house.
If hyperscalers successfully internalize chip design and reduce Broadcom dependency, or if AI capex growth disappoints relative to hype, AVGO's multiple compresses sharply from 33.9x forward P/E with no multiple expansion cushion.
"Broadcom's AI revenue projections hinge on sustained hyperscaler capex that may face delays or reallocations not accounted for in the bullish thesis."
Broadcom's shift toward AI inference via custom ASICs and networking gives it a durable role alongside Nvidia, with hyperscalers like Microsoft and Google driving design wins. Free cash flow inflecting higher supports the narrative even as multiples compress from prior peaks. Yet the $100 billion AI semiconductor target by 2027 assumes uninterrupted $400-500 billion annual capex, ignoring customer concentration risk and the fact that custom silicon often carries lower margins than standard products. The 33.9x forward P/E already embeds aggressive growth after a 102% rally since March.
If hyperscaler AI budgets accelerate rather than moderate, Broadcom could exceed its own $100 billion target by 2027, making today's valuation look conservative despite the risks flagged.
"Broadcom's upside hinges on an outsized, sustained AI infrastructure capex trend that may not materialize, making the current valuation vulnerable to a downside surprise."
Broadcom's AI thesis rests on custom ASICs for inference and hyperscale networking, which could sustain high-margin growth if hyperscalers keep spending. The story benefits from Broadcom's software-enabled hardware moat and VMware synergy, plus accelerating free cash flow. Yet the bull case hinges on optimistic projections: AI semiconductor revenue >$100B by 2027 and hyperscale capex of $400–$500B annually, which may be too aggressive if deployment slows or customers push for homegrown silicon. Valuation at ~64x trailing and ~34x forward P/E leaves little margin for error; a sudden pullback in AI cycles or supply-chain frictions could compress multiples faster than fundamentals improve.
The strongest counter is that AI capex cycles can snap and Broadcom's AI revenue targets may prove unattainable if hyperscalers reallocate budgets or build more in-house silicon, undermining the growth thesis.
"Broadcom's valuation is supported by software bundling rent-seeking, which faces significant regulatory risk that the panel has ignored."
Grok and Claude focus on the $100B revenue target, but ignore the 'VMware tax.' Broadcom’s true moat isn't just silicon; it's the aggressive bundling of software. By forcing VMware onto existing networking clients, they extract rent that pure-play semiconductor firms cannot. This creates a non-cyclical cash flow buffer that protects the valuation during semiconductor downturns. The risk isn't just hyperscaler in-house design—it's regulatory pushback against their anti-competitive bundling practices, which could force a structural margin reset.
"VMware bundling creates regulatory liability that could force margin compression, not a durable moat."
Gemini's VMware bundling angle is real, but overstates its defensibility. Enterprise software bundling faces intense regulatory scrutiny post-Microsoft/Intel precedents. More critically: VMware's $61B acquisition price assumes 15%+ EBITDA margins persist. If hyperscalers successfully negotiate unbundling or migrate to Kubernetes alternatives, that 'non-cyclical buffer' evaporates fast. The regulatory risk Gemini flags actually undermines the moat, not protects it.
"VMware bundling links software margins directly to the same hyperscalers pushing in-house silicon, amplifying rather than offsetting concentration risk."
Claude correctly flags regulatory exposure on VMware, but both overlook how bundling could accelerate hyperscaler migration to open-source alternatives like Kubernetes. This would simultaneously erode the software margin buffer and intensify ASIC concentration risk, as the same customers driving custom silicon are the ones most incentivized to unbundle. The $61B acquisition's assumed 15%+ EBITDA margins then become a double liability rather than a moat.
"VMware bundling is not a durable moat; regulatory and competitive unbundling risks threaten Broadcom's non-cyclical cash flow if margins compress."
Responding to Gemini: VMware bundling looks attractive on cash flow, but it's not a durable moat. Regulatory scrutiny on bundling and potential antitrust settlements could force unbundling beyond VMware, eroding Broadcom's software cash flow. More conservatively, the real risk is execution and integration costs of VMware into Broadcom's software stack, plus pressure from hyperscalers seeking unbundled licenses and Kubernetes-based platforms. If VMware margins compress or unbundling accelerates, the non-cyclical cushion weakens faster than anticipated.
Panelists are divided on Broadcom's AI narrative, with concerns around aggressive growth projections, regulatory risks, and potential margin compression outweighing the benefits of VMware acquisition and ASIC dominance.
Dominance in ASICs and high-speed networking for AI inference and hyperscale customers
Regulatory pushback against bundling practices and potential margin reset