00:00Who is the hyperscaler Cristiano? I will tell you all about it on June 24th. It's going to be three
00:09days after my birthday. June the 24th. We've got to hold on to. But what does this signal about you
00:16have been a man who's committed to telling the market that you are expanding the market. It's not just about
00:21mobile phones while Qualcomm has really dominated the market share. But chips now in automobiles in IOT but really in
00:28the data center. What sort of town what sort of
00:30expansion of Qualcomm's business will this mean for you? Yeah look you know this is a great question and I
00:36want to maybe start the answer by saying you know why would not anybody you know bet on Qualcomm diversification
00:45efforts. We started as a mobile company. If you look at what we're doing right now we expanded into PC.
00:49We became one of the largest providers of semiconductor to automotive. We're going to industrial. We're going to robotics. We're
00:55going to broadband and networking. And I think the data center was a natural
00:59thing. We have been working on this for the past couple of years. We've been consistently saying we're building assets.
01:05We acquire companies like Alpha Wave. We have a pretty good CPU. A very good accelerator. We're not a small
01:12company. We ship 40 billion chips. So I actually I don't know why people would bet against Qualcomm but I
01:18think it's OK. We're just being on this road to execution like we have done on the other business. And
01:25we're very exciting because I think what's really happening right now as we enter this next phase
01:29of AI. We went from training to inference now agents which actually would generate demand for tokens. It provides an
01:37opportunity across the entire Qualcomm business. All of those devices where people are going to have agents running where there's
01:44phones and PCs. They're auto. They're going to go to an upgrade cycle. And then some of the assets that
01:50we build create opportunity for data center.
01:54Christiano specificity is so important here. You said and you can correct me if I'm wrong. You said it was
02:00an ASIC product which I think the markets looked at and is a little bit surprised. So you are not
02:08doing your own chip. You're not doing merchant silicon. Your own designs for market. You're doing an ASIC with someone.
02:16Christiano specificity. So we have said that our data center product and I need to make sure I don't front
02:21run my investor day. But I try our data. Yeah, I understand that my data center. My data center product
02:29offering includes three vectors. One is CPU. We are developing a CPU and we're very flexible. We can provide we
02:39can provide a full SOC. We can provide chiplets. So that's one asset. The other asset.
02:44It's an accelerator. It's an accelerator. We have the accelerator with innovative memory architecture. We don't need HBM for the
02:52disaggregated inference. The other one is the ability to do custom chip, which when we acquire Alpha Wave, they have
03:01both a lot of IP, which is important for custom ASIC, as well as a custom chip team. When you
03:06put everything together, it creates a lot of opportunities for Qualcomm. Now, here's what is important to understand.
03:12The data center market is highly concentrated. There are about six to eight companies that represent the majority of it.
03:22And what you see what's happening in the market right now, a lot of solutions are becoming bespoke.
03:26They're no longer just merchant solutions. So I think as Qualcomm entered the space, you should expect that we offer
03:34a lot of flexibility and we'll leverage our supply chain and our design capability to do bespoke products.
03:42And I think that's what we're doing.
03:44Four to six companies. That's interesting. So this depends on your definition of hyperscaler.
03:49Like, for example, Meta is not a cloud computing company, but it operates at hyperscale, its own data center infrastructure.
03:57So maybe that's what we're talking about here, the definition of what is or is not a hyperscaler.
04:03Look, I'm just going to cut to the chase. We're talking about a large hyperscalers. You should be thinking about
04:10one of the large cloud companies that we have today.
04:14It's a scale business. So I think that's what we're talking about.
04:19Ooh, getting closer. The potential customers, therefore, are four to six, but they're huge. And maybe they're even more than
04:27that.
04:28So just can you get specific about how big the opportunity is for this particular part of the business, how
04:34much you think it could ramp up revenue?
04:36So that's something you're able to push ahead of before the investment meeting.
04:40Look, we don't get I don't want to get ahead of our skis. We're executing. We're very pleased with the
04:46engagement.
04:47And what we said is it will be material in fiscal 27. Now, if you look at the scale of
04:53the Qualcomm revenue base, I think, you know, anything material to Qualcomm is probably in the multiple of billion dollars
05:01in fiscal 27.
05:02But look, we're just starting and we're very happy with the progress we're making.
05:07Cristiano, I think back very fondly to when you and I were on stage together at the Bloomberg Tech Summit.
05:12I think it was three years ago now. And we talked about the world where we switched to inference running
05:18on device.
05:19But actually, Qualcomm's story has really changed in that time, right?
05:22So in this world where we're moving to inference and we're moving to agentic, how are you positioning Qualcomm, right?
05:29The CPU is important, but you must think differently to how you did three years ago.
05:36Look, frankly, it's actually playing out exactly the way we thought.
05:43And I know we have a limited time, but when you start running agents and when you start running things
05:48like OpenClaw and a bunch of other claws,
05:53actually what is fascinating to see what's happening in the phone industry, every Android OEM right now is launching their
06:00own claw.
06:01And the way agents work, they actually operate your device for you.
06:06So what you're going to see at the end, you're going to see a lot of activity in the cloud
06:11and you're going to see also activity on the device.
06:14And it's all going to magically work.
06:16I almost like to say this.
06:18I am sure you have probably 200 apps on your phone or close to that.
06:22But we never have a discussion.
06:24We never have a discussion about how much each app is doing on the cloud or doing on the device.
06:30It just works.
06:31And I think that's how we see those agents working.
06:34They look at your data.
06:35They look at things on your device.
06:37They do things for you and they do stuff on the cloud.
06:39And I think that's why I believe this actually builds into the Qualcomm assets and the fact that we're present
06:45everywhere.
06:45Our engagement with the cloud companies right now, they look at our devices at endpoints for AI.
06:51And that's also exciting.
06:53I think we're talking with all the AI companies about their new and exciting devices.
06:57What's less exciting has been supply chain disruption.
06:58But you were really clear with investors and they liked it quickly with 30 seconds.
07:02Memory, you see an endpoint, at least in China.
07:06Yeah, look, the reality is everything needs memory and AI drives more memory demand.
07:13They drove a lot of demand at the data center.
07:15It caused a crunch on the consumer electronics.
07:19The good thing is we now see the bottom.
07:21As I said in the call, we're under shipping market demand.
07:24We look at customer activations of devices and sell true data.
07:29We see what the demand is.
07:31That's now we can call the bottom.
07:32And, you know, there's going to be more supply and more suppliers.
07:36And we'll see how this is going to play out.
Comments