$AVGO · The Motley Fool · 2026-10-26

Although my second choice in the video, or my first choice, is what is called the "anti-Nvidia" company, Broadcom. Now, why is it called "Anti-Nvidia"? Because every time we hear that Broadcom's revenues are expected to double and double again, etc., the reason is always the desire of competitors to diversify their sources away from dependence on Nvidia, and that is why they turn to Broadcom.

They also use Marvell, but that's why it's called "Anti-Nvidia". Both players will continue to grow significantly in the foreseeable future, but today Broadcom, like Nvidia, is trading at 20 times or so future earnings.

Free cash flow, of course, grows every quarter. Based on management's projections, the AI portion of the business alone is expected to generate $115 billion next year, and that is projected to double in 2028 to $230 billion.

Again, that's just for the AI portion of the business. As for the rest, you just need to add it. And they still believe, I think, that their 2028 earnings per share target was $30 per share.

I believe the stock is currently trading at around $340 or $330 per share. So, even if you wanted to use a forward price-to-earnings multiple for 2028, and if you applied a 20x multiple to that amount of $30, you would arrive at $600 per share.

This is probably not what you would like to do, but just to illustrate that this is a company that is growing very rapidly. It is becoming more profitable day by day. And for me, if they are already setting AI revenue targets for next year and beyond, it means they have a clear vision for achieving those numbers.

This, again, represents the minimum. Just as Nvidia gave us 70% growth, I think they will give us 70% and maybe more. Here's the same thing: 115, 230, and possibly more for AI revenues. So, this is my choice.

Yes, yes, if you want to raise some of the risks associated with Broadcom, Neo, the first one, as I mentioned, is that Nvidia is such a giant. It is now clear who Broadcom's major clients are.

As I mentioned, there are a lot of cloud service providers, right? Amazon and Google are among the big players thanks to their Tensor Processing Units (TPUs). Then they also have other players in artificial intelligence labs.

They have a dead body too. So, there's a lot of pressure there. In artificial intelligence labs, OpenAI's "Jalapeno" chip appears to have performed really well, and we may begin to see commercial production of it.

But for me, one of the things I believe in, and we hear from the Nvidia team, is that building the chip is not even half the battle. Building the entire data center is paramount, along with owning that entire supply chain.

During the GTC conference in Berlin, Nvidia launched something truly remarkable, Neo. It was called the DGX Pod. They have had systems like this before, but this is the first time we have seen something resembling a shipping container that actually contains a pre-built Vera Rubin system solution inside it.

So, you can come, and they said they can reduce the time needed to get computing power after the data center is built from a few months to just 14 days. This is crazy. And if you are an OpenAI company, an artificial intelligence company, or a cloud company, New, then building these chips, paying for them, and taking a few months to connect and build everything means 4 or 5 months without any revenue.

Would I want to do that? Would my investor want me to deal with an entity that does not have real data? I think funding is a major problem at the moment for doing that. I think it exists.

Fortunately, most of the players who deal with Broadcom have sufficient liquidity, such as Google and Meta, but there are some players that I think will face a funding bottleneck, and the funders may say, "You know what?

If you really want our money, I feel more comfortable dealing with Nvidia." "Because I know that the revenue will start to be collected much faster, and therefore you will be able to pay me back much faster as well."

But regardless of that risk, I'm a big fan of Broadcom, Neil. I mean, it's another company that's trading at a forward price-to-earnings ratio of around 20 for the next twelve months.

I think it's probably still close to the early twenties with expectations of huge growth in artificial intelligence chips. They believe that next year it will reach $115 billion, and the year after that it will approach $230 billion.

These figures are very conservative on their part. They stated that they had a much larger demand than that. They are simply being conservative based on the amount of computing power that will actually be used.

For this reason, I feel they are being very cautious with those figures. Therefore, the opportunity to exceed expectations certainly exists. If I had to present one pessimistic viewpoint and one optimistic one, these would be the two points, Neil.

Yes, I agree with you. I don't think it can be described as a concentration risk, because I still believe that Google is a much bigger client for them than Anthropic or OpenAI.

But it is certainly a risk worth considering: if Anthropic and OpenAI do not grow as quickly as expected, they will not be able to pay Broadcom or even Nvidia for it. But I think that's also why we're seeing Nvidia playing the circular financing game.

I believe that Broadcom, too, with that $60 billion, aims to help accelerate this entire ecosystem. Some may not like this, but if you're a company about to generate hundreds of billions of dollars in free cash flow, and you see that you can help the business and make the ecosystem grow faster, why not?

Why not invest and push things further, especially if you're going to come out as a big winner from your own investment, of course? Yes, I certainly agree that there are some risks, but in the near future, and by near future I mean the next two years, I see Broadcom growing significantly from where it is now and becoming more profitable.