AMD reiterated its 2022 targets for roughly 60% year over year sales growth

[hist78]

Intel is talking about supply chain constraints and customer inventory adjustment that will negatively impact already depressed Intel revenue. At the same time AMD is talking about 60% YoY revenue growth!

I'm scratching my head because both of them might be right.

Some highlights from the AMD Financial Analyst Day held on June 9, 2022:

~ The company said its total addressable market is now $300 billion, vs. $79 billion at its 2020 analyst day. AMD expanded its scope with its recent acquisitions of Xilinx and Pensando Systems.

~ AMD predicted improving profit margins and free cash flow as well as compound annual sales growth of 20% for the next three to four years.

~ AMD reiterated its 2022 targets for roughly 60% year over year sales growth and a gross margin of 54%.

Investor's Business Daily: AMD Stock: Advanced Micro Devices Raises Targets.

AMD Raises Targets Amid Growing Data Center Opportunity

AMD painted a bright picture of its growing market opportunity, especially in data centers, at an analyst day presentation. AMD stock fell.

www.investors.com

 

[blueone]

I'm not believing Xilinx and Pensando are going to be such impressive revenue generators. In 2021 Xilinx revenue was less than $4B. FPGAs are great for validating and debugging chip designs, and enabling software development to proceed before sample chips are ready, but for algorithmic processing use they have that special combination of being slower than ASICs, more expensive, and consuming a lot of power. Pensando has a lot of competition in the so-called DPU space, and they have very impressive competition in DPUs from Nvidia. I do believe AMD may kick Intel's butt in x86 CPUs, but Intel may yet have a surprise up its sleeve. (I doubt Intel's near-term prowess, but I leave getting assertive about market predictions to financial analysts.)

Skepticism aside, I can't believe AMD would make such bold forward-looking statements without having a lot of confidence. Lisa Su is impressive, and I can't believe she'd sanction unlikely financial upside predictions.

 

[hist78]

I'm not believing Xilinx and Pensando are going to be such impressive revenue generators. In 2021 Xilinx revenue was less than $4B. FPGAs are great for validating and debugging chip designs, and enabling software development to proceed before sample chips are ready, but for algorithmic processing use they have that special combination of being slower than ASICs, more expensive, and consuming a lot of power. Pensando has a lot of competition in the so-called DPU space, and they have very impressive competition in DPUs from Nvidia. I do believe AMD may kick Intel's butt in x86 CPUs, but Intel may yet have a surprise up its sleeve. (I doubt Intel's near-term prowess, but I leave getting assertive about market predictions to financial analysts.)

Skepticism aside, I can't believe AMD would make such bold forward-looking statements without having a lot of confidence. Lisa Su is impressive, and I can't believe she'd sanction unlikely financial upside predictions.

"FPGAs are great for validating and debugging chip designs, and enabling software development to proceed before sample chips are ready, but for algorithmic processing use they have that special combination of being slower than ASICs, more expensive, and consuming a lot of power."

For variety of reasons, such as performance, features, flexibility, and cost, FPGAs are frequently used in many military and telecommunication applications. The F-35 Lighting II stealth fighters and Samsung 5G carrier network use Xilinx FPGAs.

 

[blueone]

For variety of reasons, such as performance, features, flexibility, and cost, FPGAs are frequently used in many military and telecommunication applications. The F-35 Lighting II stealth fighters and Samsung 5G carrier network use Xilinx FPGAs.

I'm aware of many FPGA applications; they're even used in some audio amplifiers. To put these niches into perspective, F-35 production is currently about 150 aircraft per year, though I would suspect FPGAs are also popular in civilian aircraft avionics, which are another low-volume, high margin, high cost product category. My point was that FPGAs are not a huge market, and I'm very skeptical of forecasts which predict rapid expansion.

 

[Portland]

Nvidia became what it is because of the demand for GPUs were multiples of people's best guesses. FPGAs could be the same thing.

 

[blueone]

Nvidia became what it is because of the demand for GPUs were multiples of people's best guesses. FPGAs could be the same thing.

That's a stretch for the short and medium terms. It was difficult to foresee data science as a big market for GPUs, and it took a lot of software development over several (perhaps many) years in Nvidia to make data center applications such a big market for GPUs. (Now bigger than the client graphics market.) For that kind of growth to happen for FPGAs would mean cross-platform software models like SYCL and OneAPI would have to really progress and become far more widely used. That's possible, but the trend is not likely to explode next year, which is what I was expressing doubts about in earlier posts.

 

[hist78]

We can discuss the FPGA's impact on the future growth of AMD and Intel. But most importantly AMD and Intel are competing in many same market segments. Why Intel is trending down their already depressed revenue while AMD is talking about 60% explosive growth in 2022?

If those headwinds pointed out by Intel such as supply chain constraints, customer inventory adjustment, and China Covid-19 lockdown are all true, how can AMD look for 60% growth in the same 2022 on the same planet Earth?

I searched it on Internet and haven't found any meaningful explanation. Actually most those analysts are focused on Intel's talking points and ignoring AMD, Nvidia, TSMC, and many other semiconductor companies' amazing performance and forecast.

I think they are either lazy or believe Intel represents the whole industry.

 

[Tables 777]

I'm not believing Xilinx and Pensando are going to be such impressive revenue generators. In 2021 Xilinx revenue was less than $4B. FPGAs are great for validating and debugging chip designs, and enabling software development to proceed before sample chips are ready, but for algorithmic processing use they have that special combination of being slower than ASICs, more expensive, and consuming a lot of power. Pensando has a lot of competition in the so-called DPU space, and they have very impressive competition in DPUs from Nvidia. I do believe AMD may kick Intel's butt in x86 CPUs, but Intel may yet have a surprise up its sleeve. (I doubt Intel's near-term prowess, but I leave getting assertive about market predictions to financial analysts.)

Skepticism aside, I can't believe AMD would make such bold forward-looking statements without having a lot of confidence. Lisa Su is impressive, and I can't believe she'd sanction unlikely financial upside predictions.

Given Lisa's history of underpromising and overdelivering, she is likely setting the company up for future beats and raises.