Over the past two years, AMD has achieved incredible success in the PC and server markets. The chipmaker went from being non-existent in the server space to conquer more than 10% of the share in less than a handful of years. Similarly, the desktop processor market has also been reacting very positively to the company’s Ryzen products with many regions seeing complete domination over rival Intel parts. AMD also re-entered the mobile processor market with Renoir last year, following it up with Cezanne and plans to launch Rembrandt early next year.
AMD CFO Devinder Kumar during the Deutsche Bank Technology Summit laid down the chipmaker’s plans and priorities for the coming years. It looks like AMD is laser-focused on the server and mobile segments. This means that the Epyc-SP will get preferential treatment in terms of roadmap execution, supply, and marketing. Side-by-side, the mobile markets will be targeted with a yearly cadence.
Update: There was a mistake in translating the quote from the source site. The original quote reads as follows:
We apologize for any confusion caused and will be more careful with foreign sources from here on. Even though the company didn’t directly say this, from what we’ve seen this is indeed the case. Products with higher margins will be prioritized such as the Epycs and higher-end Ryzens alongside the consoles, and then the Radeon GPUs. Hopefully, this will change next year.
According to rumors, mass production of Rembrandt has already started, and the lineup is on track for a Q1 2022 launch. If you’ve forgotten, Rembrandt will be the successor to Cezanne in the mobile processor market. Although it’ll feature the same Zen 3 core architecture, it’ll upgrade the GPU quite massively.
Our highest priority is to secure revenue share, so there will be a sequential strategy in terms of products, with server products and mobile processors first, followed by high-end desktop products. Devinder Kumar, AMD CFO
This inevitably means that the desktop and DIY space will take the backseat, as we’ve seen with Zen 3. We’re yet to see a budget SKU priced under $200 or an affordable eight-core option. This could be an opportunity for Intel to capitalize on AMD’s shortcomings with the Alder Lake launch later this year.
Team Blue will finally be able to match AMD on the multi-threaded front, all the while crushing it in the single-threaded department. Regardless, the latter isn’t sitting idle. There are plans to launch Zen 3D for both the server and consumer markets with 2-3x more L3 stacked V-Cache. These chips should be potent in gaming workloads while boosting investor confidence and awarding packaging leadership to Su and Co.
On the downside, AMD is entirely reliant on TSMC for its supply and as more and more chipmakers go fabless, the need for an alternative is growing evermore. With Intel, NVIDIA, and AMD all set to rely on TSMC’s 5nm (N6 for Xe-HPG) node for the next generation of their products, competition will be fierce than ever. We’ll have to wait and see if AMD can procure sufficient supply for all its upcoming products or whether some segments will suffer at the cost of others.
Finally, let’s not talk about the GPU space. For the time being, AMD is only invested on a PR basis with most higher-end RDNA 2 parts being absent from the market. Given the slimmer profit margins, I don’t expect that will change by much. You can read what Kumar said about the company’s graphics business in the below post: