Broadcom Q3 results confirm the stock as the number two player in AI as a combination of AI networking, custom silicon and AI software propelled total revenue to record levels. Semiconductor Solutions accelerated nine points to 26% YoY growth due to a rebound in AI accelerators (+63% YoY) and networking (170% YoY). While non-AI Semiconductor revenue remains weak and flat sequentially, VMware’s contribution helped prop up the segment.
Looking ahead, Broadcom’s Q4 guide implies acceleration into year-end, with total revenue expected to reach $17.4B. More importantly, we picked up on a subtle shift in commentary from CEO Hock Tan regarding AI revenue growth in FY26, which should place Broadcom firmly above $30 billion for next fiscal year. Tan stated: “[…] we now expect the outlook for fiscal 2026 AI revenue to improve significantly from what we had indicated last quarter.”
As we previously discussed this summer, demand for inference is booming. Broadcom’s edge goes beyond the fact that custom accelerators are often multiples cheaper than Nvidia’s GPUs for inference tasks – it's that custom silicon is increasingly performant with each generation. By optimizing algorithms (software), Big Tech can drive higher performance from large language models (LLMs) — which helps to drive down costs while also increasing output for specific workloads. For example, a rough idea as to how much it costs Nvidia to make merchant GPUs is estimated around $3,000 to $5,000 whereas the company charges $25,000 to $30,000 – hence the AI leader’s excellent margins. Reducing Nvidia’s high pricing power is what Big Tech is after and this can be accomplished both in the hardware costs but also through optimizing the workloads for specific use cases.
Big Tech is prominent in Broadcom’s custom silicon customer list, which includes Google and Meta. ByteDance reportedly emerged as the third customer last summer. Tonight, a fourth customer was announced for a $10 billion XPU order, which was likely either OpenAI or perhaps Apple, as both were heavily rumored to be prospective customers (note: customer name was not offiically confirmed)
“Last quarter, one of these prospects release production orders to Broadcom. And we have accordingly characterized them as a qualified customer for XPUs. And, in fact, has secured over $10 billion of orders of AI regs based on our XPUs.”
Perhaps most interesting, Hock Tan slipped into the opening remarks that their backlog is at $110 billion. I was glad when an analyst circled back to this comment as this backlog is astonishing, to say the least.
More details are below!
170% AI Networking Growth – Outpacing Street Models
Q3’25 Revenue was $15.95 billion, beating estimates for $15.82 billion, and reflecting top line growth of 22.0% YoY and 6.3% QoQ. This represents record quarterly revenue, driven by custom AI accelerators, networking switches, and VMware strength in Infrastructure Software.
Looking ahead, management provided Q4’25 guidance of $17.4 billion of revenue, implying 24% YoY growth and a slight uptick to 9% QoQ growth. This guide suggests that AI strength will more than offset flat non-AI revenue, while VMware remains a stable contributor.
Key Segments
Semiconductors were the main growth driver, fueled by AI accelerators and Ethernet networking. Infrastructure Software, which includes VMware acquisition, remained strong even though growth decelerated slightly when compared to prior quarters,
Semiconductor Solutions Revenue accelerated nine points to 26% YoY at $9.17 billion. AI Semiconductor revenue surged 63% YoY to $5.2B, showing re-acceleration after a slower Q2 (+46% YoY). AI now represents 57% of Semiconductor revenue and 32% of total company revenue.
Management guided Q4 AI revenue to $6.2B, which would represent ~19% sequential growth and eleven consecutive quarters of YoY growth. Though Broadcom did not lay out a FY25 AI revenue target, Q4 implies Broadcom is guiding for $19.9 billion in AI revenue for the year, up 63% YoY from $12.2 billion in FY24.
Non-AI Semiconductor revenue remained flat at ~$4B, indicating continued end-market softness outside of AI. As seen below, the gap between AI and non-AI revenue is widening as AI growth accelerates.
Infrastructure Software Revenue of $6.8B, ahead of guidance of $6.7B, up $190M QoQ from $6.6B in Q2’25 and up from $5.8B in Q3’24. $6.7B in revenue represents 43% of total revenue and 17% YoY Growth.
Operating Leverage Offsets Gross Margin Compression
Broadcom delivered balanced profitability in Q3 FY25. While AI hardware growth slightly compressed GAAP and Non-GAAP gross margins, the Company leveraged its scale and VMware’s recurring revenue to expand operating and net margins YoY. Broadcom’s ability to maintain a 67% EBITDA margin during a period of rapid AI expansion demonstrates exceptional cost control and pricing power, positioning it well for the Q4 acceleration. See below for in-depth breakdown of GAAP and Non-GAAP margin figures:
GAAP gross margin of 67.1%, down 90 bps QoQ from 68% in Q2, and up from 63.9% in the same period last year. The margins were down sequentially due to custom silicon mix, which typically carries lower margins. Broadcom is flexing its ability to manage cost structures even as AI accelerators scale rapidly.
Non-GAAP gross margin of 78.4%, down 100 bps QoQ from 79.4% in Q2 and up 100 bps YoY. This is driven by the increase in AI hardware revenue, which has lower gross margins compared to VMware’s high-margin software business.
GAAP operating margin of 36.9%, down 190 bps QoQ from 38.8% in Q2, but up significantly from 29% in the same period last year. QoQ decline reflects seasonal opex and YoY improvement reflects strong operating leverage.
Non-GAAP operating margin of 65.5%, slightly up QoQ from 65.3% in Q2’25 and up 180 bps from 63.7% in Q3’24. Broadcom is maintaining tight opex control even as it invests heavily in AI and networking. Non-GAAP margin stability reflects strong execution in managing both segments.
GAAP Net Margin of 26.0%, down from 33.1% in Q2’25 and up from (14.3%) in Q3’24. Non-GAAP Net Margin of 52.7%, up from 51.9% in Q2’25 and up from 43.6% in Q3’24. Adjusted profitability is expanding as Broadcom benefits from both VMware’s recurring software revenue and AI-driven scale.
Adjusted EBITDA of $10.7B, compared to $10.0B in Q2’25 and $8.2B in Q3’24. This represents an Adjusted EBITDA margin of 67.1%, compared to 66.7% in Q2’25 and 62.9% in Q3’24.
Key Takeaways:
- AI Mix Impact on GM: As AI Semi revenue grows (now sits at 32% of company revenue) there is slight gross margin pressure because custom silicon and hardware products carry lower margins.
- Offset by Operating Leverage: Despite gross margin compression, non-GAAP operating margin expanded YoY thanks to scale and cost efficiencies.
- Stable EBITDA Margin: Maintaining a 67% EBITDA margin while growing 22% YoY highlights Broadcom’s unique profitability profile when compared to peers.
- Watch Forward Trend: With AI projected to reach $6.2B in Q4, mix shift could continue to pressure GAAP gross margins further. Look for growth from VMware and networking to keep non-GAAP margins steady or even up.
EPS Growth driven by Top-Line Growth & Economies of Scale
Non-GAAP EPS growth is the real story here: up 38% YoY, reflecting Broadcom’s ability to capitalize on AI strength with containing costs. GAAP EPS lagged slightly due to accounting for VMware’s amortization and elevated interest expense – not driven by any core business weakness.
EPS growth of 36.3% outpaced revenue growth of 22%, highlighting the company’s dual-engine model: AI hardware scale drives top-line growth while VMware’s recurring software offsets margin volatility and lifts EPS. Investors will appreciate the strong non-GAAP EPS beat and expansion in operating margins, especially given Broadcom’s premium valuation, where both EPS stability and growth are critical to sustaining multiples.
- GAAP EPS of $1.02, compared to $1.03 in prior quarter and ($0.40) in prior-year quarter.
- Non-GAAP EPS of $1.72, compared to $1.58 in prior quarter and $1.24 in prior-year quarter.
Record Free Cash Flow Conversion
Broadcom converted 44% of revenue into free cash flow, placing it among the top-tier semiconductor companies for cash generation. Strong OCF growth driven by AI semiconductor momentum and VMware’s software contribution. Broadcom continues to de-risk its balance sheet, reducing debt by $3B sequentially while building cash reserves and modestly increasing inventory ahead of anticipated Q4 AI revenue acceleration. With high FCF conversion and a strong balance sheet, Broadcom should have ample flexibility to return capital to shareholders and fund future AI growth initiatives, reinforcing its premium valuation. See below for an breakdown on cash flow figures:
- GAAP operating cash flow of $7.2B, up 10% QoQ compared to $6.5B in Q2’25 and 57% YoY $4.6B in Q3’24. This represents an operating cash flow margin of 44.9%, up from 43.7% in Q2’25 and 36.7% in Q3’24.
- Free Cash Flow of $7.0B, compared to $6.41B in Q2’25 and $4.5B in Q3’24. This represents a free cash flow margin of 44.0%, compared to 42.7% in Q2’25 and 35.6% in Q3’24.
- Cash and Cash Equivalents of $10.7B, up from $9.5B as of Q2’25 and up from $10.0B as of Q3’24.
- Debt of $64.2B, down from $67.2B in Q2’25, and down from $69.9B in Q3’24.
- Inventory of $2.2B, up from $2.0B in Q2’25, and up from $1.8B in Q3’24.
Valuation
Broadcom currently trades at a premium to Nvidia, sitting at 19x forward revenue vs Nvidia 17.6x. Based on forward earnings, AVGO trades at 38x forward earnings, 13% above Nvidia and 18% above semi-industry average.
This premium suggests that the market is pricing in 70%+ AI revenue CAGR through FY27, above management’s current 60%. The implication here: Any sign of AI growth slowing or continued margin pressure could trigger a valuation reset.
Earnings Call Q&A
Hock Tan Hints that Strong AI Growth is Incoming
Last quarter, Hock Tan had stated the following: “And reflecting this, we may actually see an acceleration of XPU demand into the back half of 2026 to meet urgent demand for inference on top of the demand we have indicated from training. And accordingly, we do anticipate now our fiscal 2025 growth rate of AI semiconductor revenue to sustain into fiscal 2026."
This evening, the following was stated:
“we now expect the outlook for fiscal 2026 AI revenue to improve significantly from what we had indicated last quarter.”
When asked later to clarify, Tan reaffirmed our understanding that it would mean above 60% growth and it was also stated the majority of the growth would come from XPUs with networking’s share of AI revenue declining next year due to XPU strength.
“Let's answer the first part first, if I could be so bold as to suggest to you, when I — last quarter when they said, "Hey, the trend of growth of '26 will mirror that of '25which is 50%, 60% year-on-year. That's really all I said. I didn't — but of course, it comes up 50%, 60% because that's what '25 is.
All I'm saying, if you want to put another way of looking at what I'm saying, which is perhaps more accurate is we're seeing — the growth rate accelerates as opposed to just remain steady at that 50%, 60%. We are expecting and seeing 2026 to accelerate more than the growth rate we see in '25. And I know you love me to throw in the number at you, but we are not supposed to be giving you a forecast for '26. But best way to describe it, it will be fairly material improvement.”
Back of the napkin math:
Given that Hock Tan disclosed that Broadcom had secured $10B+ of orders related to the new fourth customer, current growth projections may materialize far too low. We have some estimates from HSBC placing just ASICs revenue at $28.3B (+128% YoY) with only $2.5B contribution from customers outside GOOG and META.
The Street is at $19.9B for ASICs in FY26; assuming the Street is still at $28-29B AI revenue for FY26, this places Networking at $8-9B.
However, assuming XPU strength and strong networking demand (as seen throughout the ecosystem) can drive an acceleration to 70% YoY in FY26, this projects $33.8B in AI revenue up from $19.9B estimated in FY25 given the implied guide.
This ramp is supported by estimated ~80% increase in CoWoS allocation from ~83K in FY25 to ~150K in FY26, after Broadcom reportedly increased orders last week from 120-125K expected in FY26.
Notably, Hock Tan did say the additional $10 billion would be recognized in Q3 of FY2026 during the call.
$110 Billion Backlog
Broadcom stating a $110 billion backlog was not on my Bingo card tonight. Wow! That is certainly a strong statement in terms of Broadcom’s prospects for continued AI growth. It’ll take me a day or two to process what that could mean for the next 1-3 years, assuming some of this is supply constrained.
An analyst did ask for clarity and since it’s a rather large number to provide, especially since it’s primarily AI-driven, I’m quoting it in full for you below:
Stacy Rasgon. Bernstein & Co.
I was wondering if you could parse out this $110 billion backlog. Did I hear that number right? Could you give some color on the makeup of it — like how far does that go — and like how much of that $110 billion is AI versus non-AI versus software?
Kirsten Spears CFO:
Well, yes, Stacy, we generally don't break up back on digital to give you a sense of how strong the business is as a whole for the company, and it's largely driven buying AI in terms of growth. Software continued to add on a steady basis. And non-AI, as I indicated, has grown double digits. Nothing compared to AI, which has grown very strongly. Give you a sense, perhaps fully 50% of it at least is semiconductors.
Stacy Rasgon, Bernstein:
Okay. And it's fair to say that semiconductor piece, it's going to be much more AI than non AI.
Hock Tan, President:
Right.
Custom Silicon is Progressively Gaining Market Share
During the discussion, an analyst asserted that custom silicon could surpass GPUs in terms of market share. That’s a tall order given Nvidia will typically be 1-2 years ahead of any custom programs (thereby offering an advantage to those who remain with their GPUs) and the two companies will likely end the year nearly $180B apart in AI revenue with AVGO around $20B and NVDA over $200B (could see $212B).
However, I’ve also argued inference will provide an opening for Broadcom and AMD to meaningfully compete on AI accelerators. Therefore, I’m all ears and we will be watching this closely as we move along 2026-2028.
Harsh Kumar, Piper Sandler:
Hock, congratulations on all the exciting AI metrics and thanks for everything you do for Broadcom and sticking around. Hock, my question is, you've got 3 to 4 existing customers that are ramping. As the data centers for AI clusters get bigger and bigger, it makes sense to have differentiation, efficiency, et cetera, therefore, the case for XPUs. Why should I not think that your XPU share at these 3 or 4 customers that are existing will be bigger than the GPU share in the longer term?
Hock Tan, CEO:
It will be. It's a logical conclusion, Harsh, you're correct. And we are seeing that step by step. As I say, it's a journey. It's a multiyear journey because it's multigenerational, because these XPUs don't stay still either. I'm doing multiple versions, at least 2 versions, 2 generation versions, for each of these customers we have. And with each newer generation, they increase the consumption, the usage of the XPU. As they gain confidence, as the model improves, they deploy it even more. […] And that's why I say we progressively gained share.”
Conclusion:
Broadcom’s AI networking growth of 170% YoY and sharp rebound in custom accelerators at 63% YoY was certainly a highlight. In addition to strong AI growth, the company maintained a 67% EBITDA margin despite lower-margin hardware scaling, while record free cash flow of $7.0B highlighting the company’s financial strength. With AI now making up a third of total revenue and Q4 revenue projected at $6.2B, Broadcom is slated to accelerate XPUs as a primary beneficiary of the inference era.
The nod toward accelerating growth in 2026 is why the stock went from flat to +5% AH, as the CEO seemed to hint that 70% AI growth or higher is not out of the question. Essentially, there are still four months to go in 2025 and Broadcom is gearing up for a strong 2026 already. The backlog of $110 billion was a “sit up in your seat” moment as it will force analysts to ponder – just how long will it take to work through that backlog? One can safely assume the backlog will only grow from here, as Broadcom is communicating they are preparing to be a strong contender to Nvidia toward the end of this decade.
That’s a wrap! The I/O Fund’s earnings season is officially coming to a close. Keep an eye out for deep dives over the coming weeks plus coverage of notable earnings reports on stocks we don’t own.
Please note: The I/O Fund conducts research and draws conclusions for the company’s portfolio. We then share that information with our readers and offer real-time trade notifications. This is not a guarantee of a stock’s performance and it is not financial advice. Please consult your personal financial advisor before buying any stock in the companies mentioned in this analysis. Beth Kindig and the I/O Fund own shares in AVGO at the time of writing and may own stocks pictured in the charts.
Recommended Reading: