AMD’s reduction of RDNA 4 lineup may be related to future production capacity allocation
Previously, reports intimated that AMD’s Navi 4x GPU series, crafted on the RDNA 4 architecture, would eschew the inclusion of any high-end models, thereby excluding appearances akin to the caliber of Radeon RX 8000 series cards mounted on Navi 21/31, reminiscent of the erstwhile RDNA or Polaris architecture. Following this, murmurs arose of RDNA 4 encountering a plethora of “random issues” during its developmental phase, propelling AMD to forgo the Navi 41 and Navi 42, retaining only the Navi 43 and Navi 44 chipsets, while concurrently redirecting more resources to expedite the development of the RDNA 5 architecture.
According to Bits & Chips, AMD’s strategy concerning the RDNA 4 architecture is also intertwined with its production capacity allocation. In future capacity planning with Taiwan Semiconductor Manufacturing Company (TSMC), AMD has ostensibly sacrificed the next-generation Radeon RX 8000 series cards in a bid to bolster the production volume of FPGAs and GPGPUs. Given that the Artificial Intelligence (AI) domain is currently in a burgeoning phase, AMD ardently wishes to align with the market’s momentum.
Dr. Lisa Su, AMD’s Chief Executive Officer, had previously postulated that AI presents a monumental growth opportunity, anticipating its exponential ascent over the ensuing decade. In last month’s financial teleconference, Dr. Su underscored the significance of the data center product line tasked with augmenting AI workloads. To cater to the ever-evolving performance requisites of the clientele in the Chinese sector and to vie for a larger market share, AMD might further proffer bespoke AI solutions, aiming to augment the sales of the Instinct product range.
It’s imperative to acknowledge that NVIDIA, AMD’s formidable competitor, has recently reaped considerable profits, capitalizing on the AI fervor. They too opted to divert the production capacity of their GeForce RTX 40 series cards toward GPUs used in various compute cards, such as the H100, to procure elevated returns. NVIDIA successively launched the A800 and H800 series compute cards, being pioneers in proffering specialized products tailored for the Chinese market, thus gaining a competitive edge. Perhaps, witnessing NVIDIA’s triumphant strategy kindled AMD’s inclination to recalibrate its product line focus.