The Six Five team discusses Altera Spin-out and New Mid-range.
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 ask that you do not treat us as such.
Transcript:
Patrick Moorhead: Let’s jump into Intel’s Altera did a spin out and brought new mid-range product out. What’s with all the excitement of FPGAs now?
Daniel Newman: Look, Pat, there’s a couple of things going on. This is a reminder. Altera was a company that Intel acquired years back. Became their PSG group. And now just this last what, was about a year ago, they announced that they were going to spin the FPGA business back out a little bit in the style of what they did with Mobileeye. Where they spun it out but yet still track and hold the majority of the stock. The way it’s pronounced is Altera and Intel Company, but it is standing alone.
It is being run by Sandra Rivera. Sandra is someone that we’ve worked very closely with. Most recently ran the DCAI business for Intel. She’s at the helm. They are planning to, it was pretty clear from the onset of this that this will be another public company. I think a lot of this comes down to, Pat, the desire. Let’s talk about a company like Lattice Semiconductor, which 13, 14 of the last 15 quarters has shown remarkable growth, high margins and has operated really in the low to mid-end range.
So all these other companies like Xilinx and Altera played a little bit higher up in the range. And I think Altera sees the opportunity. Pat Gelsinger sees the opportunity, Sandra Rivera sees the opportunity. That the field programmable, which is the FP, is an important design right now as companies… The ASIC, sorry, there’s so much to talk about, man, my brain is running all over the place.
Patrick Moorhead: Dude, its great.
Daniel Newman: It’s this really sexy topic right now. The Rockchip, the Inferentia chip, the SambaNova chip, and then, of course you’ve got the TPUs, you’ve got the suspected NVIDIA getting into some of this themselves. You’ve got Microsoft doing it and peers partnering it. And the thing about the ASIC is it’s really hard to do, takes a long time, and if you make a mistake, it’s not very forgivable. Not to mention Intel, what they’re doing with Geti.
The field programmable isn’t going to be necessarily the replacement of that, but for certain workloads it is an incredibly robust way to build something that has some flexibility. Think about the in-between, the full programmability of a TPU, think about something that’s more like a more specific chip use case that still leaves an amount of programmability into it. The opportunity to do this, at the high end, you can do almost anything you want with this.
It’s very expensive. Xilinx was a $40 billion acquisition for AMD. This is a very valuable type of technology. But in the mid and the lower range, being able to do a number of different of programmable things, easy to incorporate AI, be able to play in that segment, be able to operate on top of open frameworks like PyTorch and TensorFlow, being able to work against standards, whether you’re building for connectivity and ethernet, or memory and CXL.
There’s just a lot of capabilities top to bottom in this. I think they see the big opportunity in the business. They see the growth in the category. They actually see the fact that there isn’t a known large competitor in the wild. There’s large competitors that are business units inside of businesses now, but now they can be a 100% FPGA-focused company in the wild that’s putting mega resources.
And I don’t think this is so much shots fired at our friends at Lattice Semiconductor, but I do think that this is a clear signal of intent that they see them coming up the stack and how successful the company’s been able to perform and saying, we’ve got great assets, we’ve got great historic research, we’ve got a strong go-to-market sales force, and we see value in programmability.
And I think it’s also just showing that the category as a whole has a lot of upside, like I said, across many things. Low power, networking, communications. And then of course I mentioned AI. I’ll leave the launch of new products to you in the mid-range, Pat, leave you with that and of course whatever you want to add about the overall launch of the business.
Patrick Moorhead: So just stepping back here, what are FPGAs used for? Two things. They’re the on-ramp to ASICs, meaning a lot of the time, or a CPU design. Before you harden something and you go from simulation to creating an FPGA to then creating hardened silicon. For instance, when I was at AMD and we did Opteron, we created an FPGA design, which was gigantic by the way, to get closer in off of simulation before you spend a ton of money hardening that silicon. FPGAs are also used as the main brain. It can be used as compute, it can be used as IO, it can be used in many different AI applications.
And then when you add on top of that putting an FPGA plus some hardened silicon like for IO or an AI block, it gets even more interesting. And the trick has always been to come up with the right software because it was hard to program these things. And sorry, the other application can be when you have a standard that’s midstream, like all the initial 5G silicon that went into Wireless Edge were all FPGAs.
And then after the standard got fully baked, all those chips got, most of those chips got transferred to an ASIC. But that’s where it sits in. It’s a lot more programmable than an ASIC. And its also time to market is probably a year quicker than an ASIC. But it’s not as, sorry about that folks, not as programmable as a-
Daniel Newman: No sneezing on the air.
Patrick Moorhead: … as a CPU. But anyways, with that background, I’m going to give hats off to Lattice who really I think kicked off all of this competitive action. And I don’t think we would’ve seen this activity from Altera without it. Because Lattice came in with brand new designs on the low power, low compute capability, and they also launched a brand new mid-range that they’re getting a ton of designs on.
So I think that really motivated the rest of the industry to move here. Altera has four different layers. They’ve got 9, 7, 5 and coming soon is 3. Agilex 3 competes directly with what Lattice has today at their low power. And the Agilex 5, which is mid-range, is now broadly available. Competition is good and they’ve got 3, 5, 7, and 9 for all these. I guess the question is how quickly can Altera get refocused on this and wake up every morning that this is all they do?
And the reason I bring this up is because Altera for many years inside of Intel, the thesis was we were going to put FPGAs inside of Xeon processors. And that didn’t end up panning out very well. How quickly can the company get focused on this? We’re going to have to see and then, by the way, AMD Xilinx, what do you have up your sleeve, and are you going to focus and bring out some new designs on modern manufacturing techniques?
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.
