The Six Five team discusses Meta’s plan to diversify its approach to Silicon.
Watch the clip here:
If you are interested in watching the full episode you can check it out here.
Disclaimer: The Six Five Webcast is for information and entertainment purposes only. Over the course of this webcast, we may talk about companies that are publicly traded and we may even reference that fact and their equity share price, but please do not take anything that we say as a recommendation about what you should do with your investment dollars. We are not investment advisors and we do not ask that you treat us as such.
Transcript:
Daniel Newman: Without further ado, Patrick, you wrote an awesome article this week on Meta, Facebook, Meta, Facebook, getting more and more into first-party silicon. Let’s start there.
Patrick Moorhead: Yeah, it’d be great. Meta’s silicon strategy has been a mystery forever and not just because they changed their name to Meta, but before that Facebook and Instagram, and I had really… This just spell in my lap, the opportunity to talk with Jason Taylor, who runs all of infrastructure for Meta. As running infrastructure, as you and I have talked about, has a lot to do with having the right silicon and that silicon can be a combination of first-party silicon that you do yourself. Some can be co-designed with somebody else, but most of that silicon are from vendors like Marvell, Intel, NVIDIA, AMD, Qualcomm, folks like that.
The fact that I got kind of first dibs on this was really made my day, and really the driver for this is that Meta made a key hire, Dr. Alexis Bjorlin, and she was at Broadcom most recently, but she had spent many years at Intel, so a real heavy hitter. The big question is, why do they need to bring a heavy hitter to Meta to do infrastructure? Well, it’s pretty simple. Well, first of all, their infrastructure is huge, and in my conversation with Jason Taylor, he really gave kudos to the current vendors in relation to machine learning training with regards to like computer vision, text-to-speech, natural language translation.
He really seems to be running into a brick wall on what they really, I don’t know, make their business on, which is recommendation engines. We see it on Netflix. We see it on Amazon. You click this, tee up this, you buy this, we think you might like this. That’s some pretty hardcore type of training that goes into it and, quite frankly, Jason said, “Looking at the roadmaps out, there’s not enough bandwidth here. We don’t have enough network bandwidth. We don’t have enough memory bandwidth.”
Dr. Bjorlin is being brought in to not only manage and work with companies that Jason was very complimentary of, the Intels, AMDs, the Marvells, et cetera, but to potentially create their own in-house chip for this. The only inkling publicly that Facebook had ever given is they did do their own in-house transcoding chip for video. You look at the amount of hours that you spend or kids spend watching videos on Instagram as an example. It just goes to show, Daniel, it reinforces that silicon is eating the world. We were talking about it a decade ago.
People were kind of laughing at us, but I think it’s really safe to say that if you have special things that you have to do that can’t be provided by the merchant silicon people and you have the capability, you might have to create or partner to do your own silicon. What I mean by partnering, AMD and Marvell both have custom capabilities that they do, so the hyperscaler or the console game maker like a Microsoft and Sony don’t have to do it on their own.
Daniel Newman: Yeah, you bring up a great point, and this is one of more and more stories I think that’s going to come out of giant tech where they’re going to be looking at their own designs. I think you made a great point yesterday. Apple had some dues about their expanded silicon capabilities and you point out the difference between design and production. I think that’s an important nuance for the market to understand, but the ability because of arm instruction sets now to be a designer of your silicon and then work with companies to produce it on your behalf, is it increasingly becoming impossible?
We’re seeing it being done with Apple. We’re seeing it being done with AWS. You talk about recommender engines for Facebook, I mean, look at Alibaba. They’ve been doing this for a while. They’ve been building out for this exact purpose. When you’re shopping on Amazon, when you’re shopping on Alibaba, when you’re surfing on Facebook or on different social networks, the ability to create recommender engines that meet all of your needs specific to your platform are going to become increasingly important.
Of course, you have frameworks like Merlin from NVIDIA that do meet a lot of these requirements and can help certain organizations to achieve their goals for recommender engines. Some companies say, “You know, that’s great, that’s good. That’s a good starting point, but we want something that’s even more specific to what we’re trying to do in our platform,” And the world has increasingly become able to deliver that.
These companies that have a lot of money, a lot of resources, they can hire the right people like we’ve seen with AWS, like we’ve seen with Alibaba, like we’re seeing with Microsoft, they’re going to be able to do it. That might end up being Netflix and that might end up being Starbucks in the future. “Might say, we want to build an ASIC that’s just for people who drink our coffee and want a very specific experience on our app,” but they might use something that’s off the shelf. That’s to be seen.
I have a great article that was really thoughtful. There’s a lot going on here. The number of entrants into silicon is going to continue to change and application-specific silicon, programmable silicon is going to continue to change the landscape of competition. It’s going to put more pressure on your traditional players, but it’s also going to… All players are going to have the opportunity to grow as the overall demand for chips continues to grow. It’s not a zero-sum game. It’s an additive opportunity and we’re going to see more players enter, and especially these players like Facebook, they’re not going to be selling what they’re going to design to other people, but they will in itself become significant players because of the amount of market that they impact.
Author Information
Daniel is the CEO of The Futurum Group. Living his life at the intersection of people and technology, Daniel works with the world’s largest technology brands exploring Digital Transformation and how it is influencing the enterprise.
From the leading edge of AI to global technology policy, Daniel makes the connections between business, people and tech that are required for companies to benefit most from their technology investments. Daniel is a top 5 globally ranked industry analyst and his ideas are regularly cited or shared in television appearances by CNBC, Bloomberg, Wall Street Journal and hundreds of other sites around the world.
A 7x Best-Selling Author including his most recent book “Human/Machine.” Daniel is also a Forbes and MarketWatch (Dow Jones) contributor.
An MBA and Former Graduate Adjunct Faculty, Daniel is an Austin Texas transplant after 40 years in Chicago. His speaking takes him around the world each year as he shares his vision of the role technology will play in our future.
