And so suffice it to say, I leave it at that. I have one hyperscale who we are shipping custom AI engines to today and leave it at that, if you don’t mind, okay? Now as far as customized switching, routing, sure. I mean, that happens. Most of the — many of the — those few OEMs, some OEMs who are supplying systems. Switch — systems, which are switches or routers and have their own custom solutions together with their own proprietary network operating system. That’s been the model for the last 20, 30 years. And today, 70% of the market is on merchant silicon. Not yet, I won’t say for not the network operating system, but certainly for the silicon is merchant silicon. So the message here is, there’s some advantages to doing a merchant solution here then to trying to do a custom solution as behavior or performance over the last 20 years have shown.
Matthew Ramsay: Thanks, Hock. Appreciate it.
Operator: Thank you. One moment for our next question. And that will come from the line of Christopher Rolland with Susquehanna. Your line is open.
Christopher Rolland: Hey, thanks for the question. So I think there’s been two really great parts of the Broadcom story that has surprised me. And the first is the AI upside. And the second is just the resilience of the core business and particularly storage and broadband in light of what have been kind of horror shows for some of your competitors who, I think, are in clear down cycles. So I’ve maybe been waiting for a reset in storage and broadband for a while, and it looks like Q4 gets a little softer here for you. Maybe you’re calling that reset a soft landing, Hock. So I guess maybe you can describe a little bit more for us what you mean by a soft landing. Does that mean that we have indeed landed here? Would you expect those businesses to be bottoming here at least?
And I know you’ve talked about it before, you guys have had tight inventory management. But is there perhaps even a little bit more inventory showing up — more inventory burn showing up for these markets? Or are the dynamics here, just all and demand that has started to deteriorate here? Thanks.
Hock Tan: Thanks. First and foremost, and you’ve heard me talked about in preceding quarter earning calls, and I continue to say it, and Kirsten reemphasized it today, we shipped very much to only end demand of our end customers. And we’re looking beyond in enterprise, even beyond and telcos even beyond OEM. We look to the end users, the enterprises of those OEM customers. We try to. Doesn’t mean we are right all the time, but we pretty much are getting very good at it. And we only ship to that. And what you’re seeing is — why I’m saying this — what you’re looking at, for instance, some numbers in broadband, some numbers in service storage that seems not quite as flat, which is why I made the point of purposely saying, but look at it collectively taking out generative AI.
My whole portfolio of products out there. It’s pretty broad, and it gets segmented into different end markets. And when we reach, I call it a plateau as we are in, you got a soft landing, as you call it, you never stay flat. There will be some products because of timing of shipments come in more and some ship out the wrong timing come in a bit lower. And in each quarter, we may show you differences, and we are showing some of it that differences in Q3 and some even in Q4. And that’s largely related to logistics timing of customer shipments, particular customers and a whole range of products that go this way. This one I referred in my remarks as revenues, which are puts and takes around a median. And that median also paints to highlight to you guys has sit around $6 billion, and it has been sitting around $6 billion since the start of fiscal ’23.
And as we sit here in Q4, it’s still at $6 billion. Now not quite there because there are some parts of it, they may go up, some parts of it go down. And that’s the puts and takes we talk about. And I hope that pretty much addresses what you are trying to get at, which is, is this — what’s — the fact that is it a trend? Or is it just a factor? And to use my expression, I call those flatters or puts and takes around a median that we’re seeing here. And I wouldn’t have said it if I’m not seeing it now for three quarters in a row, around $6 billion.
Christopher Rolland: Thanks, Hock.
Operator: Thank you. One moment for our next question. And that will come from the line of Edward Snyder with Charter Equity Research.
Edward Snyder: Thank you very much. Hock, I want to shift gears maybe a little bit here and talk about your expectations and actually indications from your customers about the integrated optics solutions that will start shipping next year. It seems to me by looking at what you’re offering and the significant improvement you get over performance and size. This would be something of great interest. Is it limited by inertia or architectural inertia by the existing solutions? Or what kind of feedback you’re getting? And why should we expect to see maybe because it’s rather a new market for you overall. You’ve not been in it before. So I’m just trying to get a feel for what your expectations are and why maybe we should start looking at this more closely.
Hock Tan: You should. I did. I made my investment, at least you should look at it a bit. I’m just kidding. But we have invested in silicon photonics, which is, I mean, literally integrating in one single solution packaging. As an example, our switch, our next-generation Tomohawk5 switch, which will start shipping middle of next year, what we call the program we call the Bailly, a fully integrated switch silicon photonic switch. And you’re right, very low power, and it’s — you make optics have always have optical and mechanical characteristics by sucking them into an integrated silicon photonics solution. You take away this failures on yield rates on mechanical, optical and translated to literally silicon yield rates. And so it’s much — we like to believe very reliable than the conventional approach.
So your question is, so why won’t more people jump in it? Well, because nobody else has done it. This — we are pioneering this silicon photonic architecture. And we’re going to — we have appeal done, a POC, proof of concept in selling Tomahawk 4 in a couple of hyperscale, but not in production volume. We now feel comfortable we have reliability data from those instances. And that’s why we feel comfortable to now go into production launch in Tomahawk5. But as people say, the proof is in the eating. And we will get it in one or two hyperscale who will demonstrate how efficient power-wise, effective it can be. And once we do that, we hope it will start to proliferate to other hyperscalers because they cannot do it. If one of them does it and reap the benefits of this silicon photonics solution, and it’s there.
You know it. I have indicated the power is simply enormous, 30%, 40% power reduction. And power is a big thing now in data centers, particularly, I would add, in generative AI data centers. That’s a big use case that could come over the next couple of years. All right.
Operator: Thank you. Thank you all for participating in today’s question-and-answer session. I would now like to turn the call over to Ms. Ji Yoo for any closing remarks.
Ji Yoo: Thank you, operator. In closing, we would like to highlight that Broadcom will be attending the Goldman Sachs Communacopia and Technology Conference on Thursday, September 7th. Broadcom currently plans to report its earnings for the fourth quarter of fiscal ’23 after close of market on Thursday, December 7th, 2023. A public webcast of Broadcom’s earnings conference call will follow at 2:00 P.M. Pacific. That will conclude our earnings call today. Thank you all for joining. Operator, you may end the call.
Operator: Thank you all for participating. This concludes today’s program. You may now disconnect.
