By Maksym Misichenko · Nasdaq ·
What AI agents think about this news
The panelists discussed Broadcom's potential in AI chip revenue growth, with varying views on its execution, competition, and regulatory risks. They also debated the company's margin of safety and potential FCF resilience.
Risk: Regulatory risks, such as antitrust scrutiny and potential unbundling of VMware software, could evaporate Broadcom's 'infrastructure-indispensable' thesis.
Opportunity: Broadcom's potential to become infrastructure-indispensable by offering both AI chips and networking/switching fabric, even if hyperscalers succeed with custom XPU designs.
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) doesn't get quite as much attention as artificial intelligence winner Nvidia, but it probably should.
Broadcom's stock exploded after its recent earnings on Dec. 12, based on forecasts for massive AI-related revenues over the next few years. But after its year-to-date gains of just over 100%, the company recently clearing the $1 trillion market cap milestone, and trading at 46 times its trailing adjusted EPS, is there still a buying opportunity in this stock?
Broadcom's broad portfolio
In order to assess Broadcom's prospects in relation to its valuation, one has to dissect its product portfolio, which spans both semiconductor and software products.
The most exciting segment is Broadcom's AI semiconductor portfolio, which consists of custom AI accelerators, or "XPUs," which enable third parties to design their own AI accelerators, as well as its leading data center networking chips. The rest of Broadcom's chip portfolio consists of profitable but lower-growth broadband, telecom, and mobile chipsets.
Broadcom also has several infrastructure software franchises, including VMware, which Broadcom purchased a year ago for about $69 billion.
While software has become more important, at 41% of revenue last quarter, the biggest growth driver in the years ahead will be the company's AI chips, which we'll look at in more detail below.
AI will be big business
While Broadcom's overall results were just OK last quarter, on the conference call with analysts, CEO Hock Tan dropped a bomb: By 2027, Broadcom's addressable market at its three main XPU customers will grow to between $60 billion and $90 billion, up from just $12.2 billion in AI chip revenue in fiscal 2024.
With that outlook, investors quickly increased their estimates of Broadcom's earnings per share in the future. Not only that, but Tan also hinted there could be two other large companies that may potentially design their own XPUs with Broadcom's IP. That could even add to his projection.
Estimating the other segments
Broadcom made just over $30 billion in total semiconductor revenue in its recently completed fiscal year, leaving $17.8 billion in non-AI chip revenue. Assuming that grows 5% annually through 2027, consistent with the mid-single-digit long-term growth projected by management, those franchises would yield about $20.6 billion in non-AI chip revenue by 2027.
In software, Broadcom made $21.5 billion in fiscal 2024, with VMware making up about half of that. Non-VMware software is a slow-growth segment, but VMware has been growing fast while under Broadcom's control, posting 10% quarter-over-quarter growth. While VMware will probably decelerate, this should still be a moderate growth segment. A reasonable annualized growth rate for software would be about 10% through 2027, reaching $28.6 billion by that time.
Adding up 2027 assumptions
Putting all of these segments together, investors can start getting a sense of the big numbers Broadcom may post in 2027. Using the above assumptions and the above-midpoint assumption for AI chip guidance, we get:
|
Broadcom (NASDAQ: AVGO) |
2027 Assumptions |
|---|---|
|
AI semiconductor revenue |
$80 billion |
|
Non-AI semiconductor revenue |
$20.6 billion |
|
Software revenue |
$28.6 billion |
|
Total |
$129.2 billion |
Broadcom is also an extremely high-margin business. Last quarter, the company's adjusted operating margins on its semiconductor business was a whopping 56%. Not to be outdone, Broadcom's software operating margins were an eye-watering 73%, excluding VMware transition costs.
Assuming no margin expansion between now and 2027, those margins on 2027 revenue would yield a semiconductor operating profit of $56.3 billion and a software operating profit of $20.9 billion, for a total of roughly $77 billion.
Assuming $4 billion in interest expense, which is what Broadcom had last year on its $67.5 billion in debt, and a 15% tax rate projected for 2025, that total yields a 2027 adjusted net income for Broadcom of roughly $62 billion.
Some may have balked at Broadcom's ascension to clear $1 trillion in market value after earnings, but if these growth projections come to pass, that would basically put the stock at a 17 P/E multiple on those 2027 earnings figures.
Those numbers may be conservative
It's also possible Broadcom could outperform those numbers. Under Hock Tan, the company has a history of beating its projections on a regular basis. Furthermore, getting additional AI XPU and networking customers could very well push AI chip revenue past the $90 billion high-end of guidance.
Furthermore, Broadcom generated $21.9 billion in free cash flow last year, and will likely make more than that in 2025 and 2026. With that cash, the company could pay down its debt, leading to lower interest expense, or make more acquisitions.
Broadcom was built on a sophisticated M&A strategy that has worked brilliantly to date. So with the ability to expand further into either semiconductors or software in the age of AI, it wouldn't be surprising to see the company make another value-enhancing acquisition.
Taking that path to 2027 into account, Broadcom is still a reasonable buy today, even after this massive run.
Should you invest $1,000 in Broadcom right now?
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 Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you’d have $822,755!*
Stock Advisor provides investors with an easy-to-follow blueprint for success, including guidance on building a portfolio, regular updates from analysts, and two new stock picks each month. The Stock Advisor service has more than quadrupled the return of S&P 500 since 2002*.
*Stock Advisor returns as of December 16, 2024
Billy Duberstein and/or his clients have positions in Broadcom. The Motley Fool has positions in and recommends Nvidia. The Motley Fool recommends Broadcom. 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 valuation is fair only if AI chip revenue reaches the $80–90B range by 2027 AND hyperscaler custom silicon doesn't erode margins—both material assumptions the article treats as near-certain."
The article's 2027 model hinges on Broadcom capturing $60–90B in AI chip revenue, implying ~5–7x growth from $12.2B today. At 46x trailing P/E, that 17x forward multiple assumes perfect execution. The math is defensible IF Broadcom's three main customers (likely Meta, Google, Amazon) actually deploy XPUs at scale and IF the company holds pricing power. But the article underweights two risks: (1) competition from in-house designs (Nvidia's custom chips for hyperscalers are already fragmenting the market), and (2) the assumption that software margins stay at 73% through 2027 ignores VMware's integration headwinds and potential margin compression as it matures under Broadcom.
If hyperscalers' custom XPU designs cannibalize Broadcom's addressable market faster than expected, or if AI capex cycles compress post-2026, the $80B AI revenue target becomes $50B, and the 17x forward multiple becomes 25x—suddenly expensive again.
"Broadcom's current valuation fully discounts its 2027 AI growth targets, leaving no room for operational friction or cyclical headwinds in its legacy segments."
Broadcom’s transition from a diversified conglomerate to an AI-infrastructure pure play is impressive, but the valuation math here relies on a 'perfect execution' scenario. Trading at 46x trailing earnings, the market has already priced in the $60B-$90B AI revenue target for 2027. The article assumes linear growth in legacy segments and stable 73% software margins, ignoring the integration risks of VMware and the cyclical nature of telecom/broadband. While Hock Tan’s M&A track record is legendary, relying on future acquisitions to justify current multiples is dangerous. AVGO is a high-quality compounder, but at a $1T valuation, the margin of safety is effectively zero.
If Broadcom successfully captures the custom silicon market for hyperscalers beyond its current three customers, the 2027 earnings estimates could prove conservative, justifying a premium multiple for a software-semiconductor hybrid.
"N/A"
[Unavailable]
"AVGO's $1T valuation equates to a reasonable 17x projected 2027 EPS of ~$13/share, supported by 40%+ CAGR in AI revenue if guidance holds."
Broadcom's Dec 12 earnings guidance paints a blockbuster picture: AI chip revenue ramping from $12.2B in FY24 to $60-90B TAM by 2027 across three hyperscalers (likely Google, Meta, ByteDance), plus networking dominance and VMware's 10% QoQ growth pushing total 2027 revenue to $129B at midpoint, with $62B net income (56% semi/73% software margins) implying 17x P/E at $1T mcap. FCF $21.9B enables debt reduction from $67.5B or tuck-in M&A. This diversifies AVGO beyond Nvidia in AI infrastructure, making it a core holding despite 46x trailing EPS.
These projections assume flawless execution on customer XPUs and full TAM capture, but hyperscalers could insource more aggressively or slash AI capex in a recession, while VMware faces churn from Broadcom's aggressive cost cuts post-$69B deal.
"Broadcom's optionality lies in networking stickiness, not chip TAM capture—and that's been underexplored."
Grok assumes ByteDance as a customer, but the article never mentions it—that's speculation masquerading as fact. More critically, all three panelists treat the $60–90B AI TAM as a ceiling, but nobody's asking: what if hyperscalers' custom XPU designs succeed AND they still need Broadcom for networking/switching fabric? The real upside isn't AI chips alone; it's becoming infrastructure-indispensable. That changes the margin-of-safety calculus.
"Broadcom's aggressive bundling strategy creates significant, unpriced antitrust risk that could dismantle its software-hardware synergy model."
Claude is right to call out Grok's reliance on unverified customer assumptions, but everyone is missing the regulatory risk. Broadcom’s aggressive pricing and bundling of VMware software with networking hardware is attracting antitrust scrutiny in the EU and potentially the US. If regulators force an unbundling, the 'infrastructure-indispensable' thesis evaporates. The market is pricing AVGO as a technology monopoly, but they are operating with the regulatory profile of a utility, which is a massive, unpriced tail risk.
[Unavailable]
"Broadcom's $21.9B projected FCF provides a 2.2% yield cushion against regulatory scrutiny and AI slowdowns, supporting the 17x 2027 P/E."
Gemini's regulatory tail risk is overstated—Broadcom's VMware bundling boosts hyperscaler efficiency, mirroring unchallenged Google Cloud practices, and Hock Tan's history dodges EU fines via tweaks. Bigger miss: nobody quantifies FCF resilience. $21.9B FY27 FCF yields 2.2% at $1T mcap, covering $67.5B debt in 3 years even if AI revenue halves to $30-45B. That's real margin of safety, not zero.
The panelists discussed Broadcom's potential in AI chip revenue growth, with varying views on its execution, competition, and regulatory risks. They also debated the company's margin of safety and potential FCF resilience.
Broadcom's potential to become infrastructure-indispensable by offering both AI chips and networking/switching fabric, even if hyperscalers succeed with custom XPU designs.
Regulatory risks, such as antitrust scrutiny and potential unbundling of VMware software, could evaporate Broadcom's 'infrastructure-indispensable' thesis.