By Maksym Misichenko · Nasdaq ·
What AI agents think about this news
Broadcom's strong AI semiconductor growth is supported by custom chip demand from hyperscalers, but the stock's high valuation (68x earnings) and potential risks from insourcing by customers and regulatory scrutiny make the investment case uncertain.
Risk: Hyperscalers insourcing AI chip design and potential regulatory scrutiny of Broadcom's M&A strategy.
Opportunity: Broadcom's role as a 'picks and shovels' play in the AI infrastructure build-out.
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 →
Broadcom (NASDAQ: AVGO) may be one of the most compelling ways to invest in artificial intelligence (AI) right now thanks to its growing role in custom AI chips and networking hardware. As hyperscalers build out data centers to train models and run inference, more customers are turning to custom silicon that can lower costs and offer greater flexibility in designing AI stacks.
Where to invest $1,000 right now? Our analyst team just revealed what they believe are the 10 best stocks to buy right now, when you join Stock Advisor. See the stocks »
Broadcom has been a major beneficiary of this shift. The semiconductor giant has been delivering solid earnings results driven by its AI semiconductor revenue. The company is well-positioned in the AI supply chain, and if hyperscalers continue to favor custom chips, Broadcom could be one of the biggest winners of the AI race.
That growth story has already helped push the stock higher over the past year even as investors have occasionally reset their expectations.
Broadcom has traded in a range of $244 to $495 over the past 52 weeks. Over that period, the stock has rallied 63%. But after management's third-quarter guidance failed to clear elevated investor expectations, the stock dropped roughly 2% over the past month.
Today, the stock is currently trading at more than 68 times earnings. That puts Broadcom squarely at a premium to peers like Nvidia, which trades at around 32x. The reason for the disparity is that investors likely think Broadcom's earnings haven't peaked yet and that it can continue expanding by selling custom chips and networking hardware. Its AI revenue could have further room to grow compared with larger competitors that have already delivered solid growth in recent years.
To see why investors are willing to pay that premium, it helps to look at what Broadcom just delivered in its most recent quarter.
Based on its fiscal 2026 Q2 results, Broadcom delivered record revenue, operating profit, and free cash flow, fueled by rapid growth in AI semiconductor demand and improved operational efficiency.
AI semiconductor revenue rose 143% year over year to $10.8 billion driven by demand for custom accelerators and networking. Consolidated revenue climbed 48% to a record $22.2 billion, and adjusted earnings before interest, taxes, depreciation, and amortization (EBITDA) rose 52% to $15.2 billion, for a 69% margin.
Despite $231 million in capital expenditure (capex), free cash flow increased 60% to $10.3 billion, supported by its capital-light, outsourced manufacturing model.
With results like that, the next question is whether Broadcom can keep the pace up as the year progresses.
Hyperscalers are increasingly turning to Broadcom for custom AI silicon as they look for alternatives to Nvidia's graphics processing units (GPUs). Broadcom has partnered with Google, Anthropic, and OpenAI to develop custom chips.
Management expects Q3 AI semiconductor revenue of $16 billion, up more than 200% year over year, helping drive consolidated revenue up 84% to $29.4 billion. It also expects a non-GAAP operating margin of 67%.
If Broadcom beats its own guidance again, the stock could earn a higher valuation, depending on sustained AI demand and cost control.
All of that sets up the investing question: After a big run and a recent pullback, does the stock still make sense to buy?
A consensus among 42 analysts rates Broadcom a strong buy, a rating that has held over the past three months. The high target price implies a 56% upside from current levels.
Broadcom remains a high-margin infrastructure provider of key AI components. Its recent sell-off could be a healthy valuation reset, potentially offering more upside. If hyperscalers continue to favor custom chips, Broadcom could keep seeing solid growth in AI semiconductor revenue and exceed its guidance, potentially making the stock one of the higher-upside AI investments.
Before you buy stock in Broadcom, consider this:
The Motley Fool Stock Advisor analyst team just identified what they believe are the 10 best stocks for investors to buy now… and Broadcom wasn’t one of them. The 10 stocks that made the cut could produce monster returns in the coming years.
Consider when Netflix made this list on December 17, 2004... if you invested $1,000 at the time of our recommendation, you’d have $417,305! Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you’d have $1,293,148!
Now, it’s worth noting Stock Advisor’s total average return is 936% — a market-crushing outperformance compared to 209% for the S&P 500. Don't miss the latest top 10 list, available with Stock Advisor, and join an investing community built by individual investors for individual investors.
**Stock Advisor returns as of June 21, 2026. *
Rick Orford has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends Broadcom and Nvidia. The Motley Fool has a disclosure policy.
The views and opinions expressed herein are the views and opinions of the author and do not necessarily reflect those of Nasdaq, Inc.
Four leading AI models discuss this article
"Valuation currently prices in persistent, outsized AI capex growth; a material slowdown in AI spending or a shift to competitors could cause meaningful downside despite the current growth runway."
Broadcom's exposure to AI via custom accelerators and networking hardware underpins a strong near-term growth story as hyperscalers scale AI workloads. The Q2 AI semiconductor revenue of $10.8B (+143% YoY) and a guided $16B for Q3 imply a high-growth trajectory that justifies a premium multiple. Yet the narrative rests on fragile assumptions: AI demand staying durable, hyperscalers sticking with Broadcom over Nvidia or rivals, and margins staying robust as mix shifts. The stock trades north of 68x earnings vs Nvidia ~32x, suggesting a big re-rating risk if AI capex slows or competitive dynamics shift. Also, AI revenue is a subset, not the whole engine.
Strongest case against: AI spending could be cyclical or peak; Broadcom's AI revenue is concentrated and could be choppy, so any slowdown or substitution of Broadcom’s offerings could trigger a sharp multiple drawdown. If Nvidia maintains a dominant position, Broadcom’s premium looks increasingly fragile.
"Broadcom's transition to a custom silicon powerhouse provides a superior, margin-accretive moat compared to the commoditization risks facing general-purpose GPU providers."
Broadcom’s pivot to custom ASIC (Application-Specific Integrated Circuit) design for hyperscalers is a masterstroke of business model evolution, effectively insulating the firm from the cyclical volatility of general-purpose GPUs. While the 68x P/E ratio looks steep, it reflects the market pricing in a structural shift where Google, Meta, and others prioritize internal silicon to optimize TCO (Total Cost of Ownership). The 69% EBITDA margin is staggering, proving that Broadcom’s 'design-to-manufacture' model is the most efficient way to capture AI infrastructure spend without the massive capex burden of building fabs. Provided they maintain their lead in high-speed networking, Broadcom is the 'picks and shovels' play for the next phase of the AI build-out.
The primary risk is hyperscaler 'in-sourcing' fatigue or a sudden pivot back to merchant silicon if Nvidia’s next-generation Blackwell or Rubin architectures significantly outperform custom designs on power-per-watt efficiency. Furthermore, the 68x valuation leaves zero room for error; any deceleration in hyperscaler capex will lead to a violent multiple compression.
"Broadcom's AI revenue growth is real, but the company's role as a custom-chip enabler may be self-liquidating as hyperscalers build internal design and manufacturing capabilities."
Broadcom's 143% YoY AI revenue growth and $16B Q3 guidance are real and material, but the 68x forward P/E deserves scrutiny. The article frames custom chips as a durable moat, yet omits a critical risk: hyperscalers are building *their own* silicon (Google TPU, Amazon Trainium, Meta MTIA). Broadcom's role may be transitional—providing design wins and manufacturing partnerships while customers develop in-house capabilities. The 69% EBITDA margin is exceptional, but assumes sustained pricing power. If hyperscalers achieve vertical integration, Broadcom risks commoditization or margin compression within 18-24 months. The stock's 63% YoY rally already prices in significant upside; the recent 2% dip is noise, not a reset.
If custom silicon adoption accelerates faster than in-house development timelines, Broadcom could sustain 50%+ growth for 3+ years, justifying the valuation multiple on a discounted cash flow basis.
"The 68x multiple already prices in flawless custom-AI execution that remains unproven at the volumes needed to justify it."
Broadcom's AI semiconductor revenue jump to $10.8B (143% YoY) and $16B Q3 guide look strong, but the 68x earnings multiple versus Nvidia's 32x assumes custom-chip wins with Google, Anthropic, and OpenAI will scale without hiccups. Hyperscaler concentration, potential delays in tape-outs, and the capital-light model's sensitivity to any volume shortfall are underplayed. The recent 2% post-guidance dip signals investors are already dialing back expectations. If Q3 misses or networking attach rates disappoint, the re-rating risk is asymmetric.
Even with slower custom-chip ramps, Broadcom's established networking franchise and VMware diversification could still deliver 30%+ earnings growth, justifying part of the premium over pure-play GPU names.
"The real risk is revenue concentration and hyperscaler insourcing; the 68x multiple embeds continued AI spend by a tight group—any shift in capex or in-house silicon could trigger sharp re-rating."
Gemini's bullish take on the 'picks and shovels' thesis glosses over concentration risk. The real flaw is revenue concentration: if hyperscalers start insourcing more quickly or slow AI capex, Broadcom's AI revenue could be far choppier than the EBITDA margin implies, yet the stock trades at 68x forward earnings. A material cyclic/competitive shock could trigger a sharper multiple re-rating than any QoQ margin improvement, especially if customers push for longer-term price concessions or alternatives.
"Broadcom's proprietary networking IP creates a high-moat dependency that makes in-house hyperscaler silicon development significantly harder than a simple design-to-manufacture transition."
Claude, your 'transitional' thesis ignores the massive switching costs of Broadcom’s IP ecosystem. Hyperscalers aren't just buying chips; they are buying Broadcom’s proprietary SerDes and PCIe switch IP, which are the backbone of data center connectivity. Developing in-house alternatives that match this efficiency is a multi-year engineering nightmare, not a 24-month pivot. The real risk isn't commoditization—it's the regulatory scrutiny of their M&A-heavy growth strategy, which could limit future inorganic expansion.
"Switching costs buy time, not permanence; the margin compression risk if Broadcom becomes a contract manufacturer rather than a design partner remains underpriced."
Gemini's switching-cost argument is overstated. SerDes and PCIe IP are valuable, but hyperscalers have already demonstrated willingness to absorb multi-year engineering costs (Google TPU, Amazon Trainium). The real question: does Broadcom's *current* margin advantage persist if customers internalize design? Nobody's addressed whether Broadcom's 69% EBITDA margin survives if they shift from custom-chip design partner to pure-play contract manufacturer. That's the commoditization risk Claude flagged—and it's structural, not cyclical.
"Networking IP stickiness with custom chips could preserve margins longer than the 18-24 month commoditization risk implies."
Claude's commoditization timeline underplays the integration between custom ASICs and Broadcom's SerDes/PCIe networking IP, which hyperscalers cannot easily replicate without losing connectivity performance. This linkage could keep attach rates and pricing power intact even if design wins migrate in-house, limiting margin compression to the networking segment rather than the full 69% EBITDA base. The 68x multiple still embeds too little buffer for any volume shortfall in either area.
Broadcom's strong AI semiconductor growth is supported by custom chip demand from hyperscalers, but the stock's high valuation (68x earnings) and potential risks from insourcing by customers and regulatory scrutiny make the investment case uncertain.
Broadcom's role as a 'picks and shovels' play in the AI infrastructure build-out.
Hyperscalers insourcing AI chip design and potential regulatory scrutiny of Broadcom's M&A strategy.