NVIDIA GeForce RTX 3060 12 GB "Ampere" GPU Set for Return in 2024
NVIDIA is preparing to reintroduce its GeForce RTX 3060 12 GB "Ampere" graphics card, a model that originally launched two generations ago. This unexpected move is anticipated for June 2024 and comes as the rumored GeForce RTX 5050 9 GB edition faces delays. According to respected hardware leaker MEGAsizeGPU, NVIDIA has decided to pause the transition from the 8 GB RTX 5050 "Blackwell" to a 9 GB variant, focusing instead on bringing back the GeForce RTX 3060 12 GB for the budget GPU market.
Why NVIDIA Is Reviving the RTX 3060 12 GB
The GeForce RTX 3060 12 GB will return with its original 192-bit memory bus and will once again be manufactured using Samsung’s 8 nm DUV process. This is notable because the entire "Ampere" lineup was produced on this node, and its reappearance comes after NVIDIA’s recent shift to TSMC’s advanced 5 nm process for its "Ada Lovelace" and "Blackwell" GPUs. NVIDIA is now TSMC’s largest customer for 5 nm chips, making the decision to revisit Samsung’s 8 nm technology particularly interesting.
The RTX 3060 features 3,584 CUDA cores and 12 GB of GDDR6 memory, positioning it as a strong contender in the entry-level and mid-range segments. In comparison, the upcoming RTX 5050 is expected to use the newer GB207 die with 2,560 CUDA cores and 9 GB of GDDR7 memory. However, the RTX 3060’s use of GDDR6 may offer NVIDIA greater flexibility in sourcing memory, while reserving the more advanced and costly GDDR7 for higher-end models.
Strategic Manufacturing Decisions
The choice to bring back a two-generation-old GPU architecture in 2024 raises questions, especially since NVIDIA could have opted for a newer model like the RTX 4060. One possible explanation is that both the RTX 4060 and the upcoming RTX 5060 are produced on TSMC’s 5 nm-class (NVIDIA 4N) node, which is currently in high demand for the latest "Blackwell" GPUs and enterprise solutions. By reviving the RTX 3060, which is built on Samsung’s 8 nm DUV node, NVIDIA can allocate TSMC’s 5 nm capacity to its most advanced products.
Additionally, GPU designs are typically optimized for a specific manufacturing process. Adapting the RTX 3060 to a different node would require significant investment and engineering resources. By sticking with Samsung’s 8 nm process, NVIDIA avoids these upfront costs and leverages existing production capabilities.
Implications for the GPU Market
The return of the GeForce RTX 3060 12 GB "Ampere" GPU highlights NVIDIA’s strategy to address ongoing demand in the budget graphics card segment. With the RTX 5050 9 GB on hold and the RTX 4060 sharing manufacturing resources with newer models, the RTX 3060 offers a proven solution that can be produced efficiently. This move also underscores the complexities of GPU manufacturing and supply chain management, as NVIDIA balances innovation with market needs and production constraints.
