By Maksym Misichenko · Nasdaq ·
What AI agents think about this news
The panel consensus is bearish on Broadcom's long-term margins and the feasibility of achieving $100B in annual AI revenue by FY2027. Key risks include execution challenges, customer concentration, and potential margin erosion due to pricing pressure and in-house design moves by hyperscalers. The non-AI cyclical drag on Broadcom's legacy segments also poses a significant valuation risk.
Risk: Execution challenges and customer concentration in custom ASIC deals, as well as potential margin erosion due to pricing pressure and in-house design moves by hyperscalers.
Opportunity: None identified
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 and Advanced Micro Devices compete in AI chips, but have very different strategies and products.
Broadcom's custom silicon game plan has yielded several blockbuster partnerships with $100 billion in annual sales on the way.
The stock is more expensive than AMD's, but it's worth paying that premium.
Nvidia still dominates the market for chips used in artificial intelligence (AI) data centers. But AI is such a tremendous opportunity that second place isn't so bad in this case. Statista estimates that the AI chip market will continue to grow, reaching $333 billion by the end of the decade.
That leaves plenty of room for other companies to make investors quite a bit of money over the coming years. Broadcom (NASDAQ: AVGO) and Advanced Micro Devices (NASDAQ: AMD) would be two prime candidates. Both have made inroads with their AI chips, but when it's all said and done, one stands out above the other. Here's why Broadcom is likely the better AI chip stock to own.
Will AI create the world's first trillionaire? Our team just released a report on the one little-known company, called an "Indispensable Monopoly" providing the critical technology Nvidia and Intel both need. Continue »
The million-dollar question is, how does a smaller company compete with the industry giant? Advanced Micro Devices, or AMD for short, is primarily competing head-to-head with Nvidia in general-purpose AI chips. To its credit, AMD has gotten some traction; data center revenue grew by 57% year over year to $5.8 billion in the first quarter of 2026.
AMD is naturally going to see some opportunities from the AI hyperscalers who don't want to put all their eggs in Nvidia's basket. AMD recently announced plans to supply Meta Platforms with 6 gigawatts of its Instinct GPUs, including a custom version for the first gigawatt. But it's unlikely that AMD will ever threaten Nvidia's stranglehold, since Meta and other Nvidia customers have already leaned heavily on Nvidia's CUDA software.
To make a dent in Nvidia's moat, you need to take a new path. Broadcom has done just that with its XPU chips. Instead of emphasizing general-purpose AI chips, Broadcom is co-designing custom chips for each customer's AI workloads. That creates efficiency gains and can also build a stickier relationship with that customer. Broadcom is working with Anthropic, Alphabet, Meta Platforms, and OpenAI, among others, on custom silicon. It is a big opportunity as compute shifts from training to inference, where efficiency matters more.
Wall Street sold off Broadcom hard for its second-quarter earnings, mainly because its third-quarter AI revenue guidance came in below expectations. However, CEO Hock Tan reiterated the company's long-term expectation of $100 billion in annual AI chip sales starting in fiscal year 2027. Broadcom generated $10.8 billion in AI revenue in the second quarter, so there's still significant growth ahead as these custom silicon projects roll out.
Given Broadcom's list of high-profile AI customers and continued momentum toward $100 billion in annual AI chip revenue, it's hard not to like Broadcom's competitive position more than AMD's at this point. Investors may notice that Broadcom stock is more expensive than AMD's, but the premium seems justified.
Analysts see far faster growth ahead for Broadcom compared to AMD, and the valuation gap between the two isn't that wide. Broadcom is where you want to put your money, especially after this post-earnings dip.
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 $439,847! Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you’d have $1,342,065!
Now, it’s worth noting Stock Advisor’s total average return is 968% — a market-crushing outperformance compared to 211% 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 6, 2026. *
Justin Pope has positions in Alphabet and Meta Platforms. The Motley Fool has positions in and recommends Advanced Micro Devices, Alphabet, Broadcom, Meta Platforms, 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
"Broadcom's $100B AI revenue target by 2027 is highly uncertain and the stock is pricing in a best-case outcome."
Nvidia still dominates AI datacenter chips; AMD is gaining ground but faces a crowded field. Broadcom's XPU strategy hinges on deep co-design with a handful of hyperscalers and the transition from training to inference. While that could unlock stickier, higher-margin revenue, the path to $100B annual AI chip sales by 2027 is unproven and likely lumpy across customers. Execution risk, margin discipline, and customer concentration could cap upside if AI compute demand cools or if Nvidia maintains CUDA software moat. The article glosses over total company exposure outside AI and the risk of multiple compression if AI capex slows.
If Broadcom nails several large, durable XPU deals and maintains high incremental margins, the stock could re-rate well before 2027, making the bearish view too pessimistic.
"Broadcom’s valuation premium is predicated on the assumption that hyperscalers will outsource custom silicon design indefinitely rather than internalizing the IP to capture higher margins."
The article correctly identifies the divergence between AMD’s 'merchant' model and Broadcom’s 'ASIC' (Application-Specific Integrated Circuit) model. However, it glosses over the margin compression risks inherent in custom silicon. While Broadcom (AVGO) secures 'stickier' revenue, it effectively acts as a high-end design house for hyperscalers who are notoriously ruthless on pricing once they achieve scale. AMD (AMD) faces a brutal software moat via Nvidia’s CUDA, but its MI300 series is becoming a viable 'good enough' alternative for inference. If the AI market shifts from massive training clusters to specialized inference, Broadcom’s efficiency plays win, but if hyperscalers bring more design in-house, Broadcom’s long-term margins could face significant headwinds.
Broadcom’s reliance on custom silicon for a handful of hyperscalers creates massive concentration risk; if Alphabet or Meta decide to pivot their architecture, Broadcom has no 'merchant' product to fall back on.
"Broadcom's $100B AI target hinges on flawless multi-customer custom ramps that carry higher concentration and execution risk than the article acknowledges."
The article pushes Broadcom as the clear winner via custom XPU deals with Meta, Alphabet, OpenAI and Anthropic, targeting $100B annual AI revenue from FY2027. Yet it downplays the post-earnings selloff on soft Q3 guidance and the inherent limits of bespoke silicon: each design ties Broadcom to one customer's roadmap, raises non-recurring engineering costs, and caps addressable volume versus AMD's general-purpose Instinct GPUs. AMD's 57% data-center growth and Meta's 6GW commitment show hyperscalers still hedging Nvidia with off-the-shelf parts. Broadcom's premium valuation therefore embeds execution perfection that custom programs have historically struggled to deliver.
Custom designs could still deliver higher margins and stickier revenue once ramped, making the current dip an entry point the article correctly flags.
"Broadcom's $100B FY2027 revenue target requires 9x growth in 18 months with no margin for execution risk, and the stock's premium valuation leaves limited upside if the timeline slips even one year."
The article's $100B Broadcom AI revenue forecast by FY2027 requires scrutiny. Broadcom generated $10.8B in Q2 2026, implying ~9x growth in 18 months—aggressive even for AI. The custom silicon thesis is real, but the article underplays execution risk: design cycles slip, customers hedge across vendors, and hyperscalers increasingly build in-house (Google's TPU, Amazon's Trainium). AMD's CUDA disadvantage is real, but the article ignores that Meta's 6GW commitment is *custom* silicon too—proving AMD can differentiate. Broadcom's post-earnings selloff suggests the market already priced in optimism; the premium valuation leaves little room for delays.
Broadcom's custom silicon moat is stickier than AMD's commodity approach, and the $100B target may actually be conservative if inference workloads explode faster than expected—in which case the stock's premium is justified and undershoots upside.
"Hyperscalers' pricing pressure and in-house design moves threaten Broadcom's long-run margins, even with bespoke XPUs, challenging the 2027 AI revenue target."
Gemini overstates the stickiness of Broadcom's margins by relying on a few hyperscalers. The real risk is pricing pressure and ever more in-house design moves from Google, AWS, and others; once a standard accelerates, Broadcom's custom ASICs could see margin erosion, not expansion. A $100B AI revenue target by FY2027 hinges on volume and pricing power that may fade as customers internalize the cost of bespoke IP. Bearish bias on Broadcom long-run margins.
"Broadcom's valuation is vulnerable because it relies on aggressive AI growth to mask structural stagnation in its legacy non-AI business segments."
Claude, you hit on the critical blind spot: the $100B revenue target is mathematically aggressive, yet we are all ignoring the 'non-AI' drag. Broadcom’s legacy segments—storage and broadband—are currently stagnant. If AI growth doesn't perfectly offset the cyclical weakness in these divisions, the valuation multiple will contract violently. We are debating AI margins while assuming the rest of the business remains a stable cash cow, which is an increasingly dangerous assumption in this macro environment.
"Non-AI weakness amplifies Broadcom's custom silicon concentration risks, hastening valuation compression."
Gemini flags non-AI cyclical drag correctly, yet this risk directly magnifies the custom ASIC concentration limits I highlighted: when storage and broadband weaken, hyperscalers facing the same macro pressure can delay or cancel the few XPU programs that underpin the $100B target. Broadcom then has no diversified merchant revenue to cushion the blow, unlike AMD, turning a soft offset into outright multiple compression that the recent selloff only previewed.
"Macro pressure on hyperscalers' own capex budgets poses greater risk to Broadcom's AI revenue than the company's legacy segment weakness."
Gemini and Grok correctly identified non-AI cyclical drag, but they're conflating two separate risks. Storage/broadband weakness is a *valuation* problem (multiple compression), not a *revenue* problem for the $100B AI target itself. Broadcom's custom ASIC concentration is the real threat: if hyperscalers delay XPU programs during macro weakness, that directly kills the $100B forecast. The legacy business doesn't need to stay strong—it just needs to not force customer capex cuts.
The panel consensus is bearish on Broadcom's long-term margins and the feasibility of achieving $100B in annual AI revenue by FY2027. Key risks include execution challenges, customer concentration, and potential margin erosion due to pricing pressure and in-house design moves by hyperscalers. The non-AI cyclical drag on Broadcom's legacy segments also poses a significant valuation risk.
None identified
Execution challenges and customer concentration in custom ASIC deals, as well as potential margin erosion due to pricing pressure and in-house design moves by hyperscalers.