AVGO Q3 2024 Earnings Analysis

Alex, these numbers are pretty remarkable. Broadcom posted $13.1 billion in revenue, up 47% year-over-year. Operating profit was up 44%. But here's the thing - when you strip out VMware, which they acquired last year, the underlying semiconductor business only grew 4% year-over-year. So this is really a story of two businesses firing on different cylinders.

Absolutely. VMware contributed $3.8 billion in revenue this quarter, and Hock Tan was pretty excited about their transformation strategy. They're aggressively moving customers from perpetual licenses to subscription models, specifically pushing something called VMware Cloud Foundation - that's their full virtualization stack. What's impressive is they booked over 15 million CPU cores of this product, representing 80% of total VMware bookings. That translated to $2.5 billion in annualized booking value, up 32% from the prior quarter.

Exactly - classic Broadcom playbook. They brought VMware's operating expenses down to $1.3 billion from $1.6 billion in Q2. Tan said they're on track to hit their target of $8.5 billion in adjusted EBITDA within three years, and maybe even exceed it by fiscal 2025.

AI is absolutely driving their semiconductor business. Tan said AI revenue will hit $3.5 billion in Q4, bringing the full year to $12 billion - up from their previous guidance of $11 billion. That's huge growth. The interesting part is the mix - about two-thirds is custom AI accelerators and one-third is AI networking. These are the chips that power AI data centers for the big hyperscalers like Google, Amazon, Microsoft.

This was fascinating, Alex. Tan actually said he's changed his mind on this. He used to think general-purpose merchant chips would win - that's typically how the semiconductor industry works. But now he believes the big cloud companies have the scale and financial resources to justify building their own custom AI accelerators. His logic is compelling: if GPUs are more important than engineers to these companies right now - and he literally said that - then controlling your own silicon destiny makes sense. He sees this trend accelerating, though it'll take time.

Yeah, but here's the key - Tan believes they've hit bottom. Non-AI bookings were up 20% year-over-year in Q3, which is a strong leading indicator. Networking revenue outside of AI was up 17% sequentially, even though it was still down 41% year-over-year. The company sees different recovery timelines by segment - server and storage are showing early signs of improvement, wireless should get a boost from new device launches, but broadband remains weak due to telecom spending cuts.

One analyst asked about the shift from cloud to enterprise in AI spending, wondering if that might hurt Broadcom since they're focused on hyperscalers. Tan's response was basically "not our problem" - they're laser-focused on the big cloud players and don't really play in the enterprise AI market. Another interesting exchange was about M&A. When asked about future acquisitions, Tan was very clear: his hands are full with the VMware transformation and he wants to see that through over the next year or two.

For Q4, they're guiding $14 billion in revenue - that's 51% year-over-year growth. For the full fiscal year 2024, they're now expecting $51.5 billion in revenue with adjusted EBITDA at 61.5% of revenue. The company also reinstated quarterly guidance going forward, which signals they feel the post-VMware integration chaos is behind them.

I see three big themes. First, AI sustainability - can they maintain this $12 billion AI revenue run rate and grow it further? They have three meaningful custom chip customers now, but the question is whether demand stays strong. Second, the VMware transformation - if they can successfully move customers to higher-value subscription models while cutting costs, this becomes a cash generation machine. Third, the cyclical recovery in non-AI semiconductors. If Tan's right about hitting bottom, we could see some nice tailwinds as enterprise IT spending normalizes.

Well, they're carrying $72 billion in debt from the VMware acquisition, though they're actively paying it down. Free cash flow margins have compressed due to higher interest expenses and tax mix issues - that's something to monitor. Also, while the AI growth is impressive, it's concentrated among a few large customers. That creates both opportunity and risk.

They're uniquely positioned in two massive technology transitions - the shift to custom AI silicon and the move to software subscriptions. The AI business gives them exposure to what might be the biggest technology wave we've seen, while VMware provides steady, recurring revenue. Tan has consistently delivered on his acquisition integration playbook, and the early VMware results suggest this deal is following that pattern. If the non-AI semiconductor recovery materializes as expected, you've got multiple growth drivers firing simultaneously.

Everything discussed today is AI-generated analysis for educational purposes. Past performance doesn't guarantee future results. Please do your own due diligence. But this quarter really showcased Broadcom's transformation from a traditional semiconductor company to something more diversified and strategic. The AI tailwinds are real, the VMware integration is progressing well, and they may have found the bottom in their cyclical businesses.