Post by Minh Nguyen Van

🚀 Why NVIDIA May Choose TSMC's A16 Instead of Just 2nm NVIDIA's next-generation Rosa CPUs, expected to arrive in 2029, are reportedly being evaluated on TSMC's 2nm family of process technologies, including the company's A16 node. 🔍 While 2nm naturally attracts attention, what I find even more interesting is the technology behind A16. Unlike conventional process nodes, A16 introduces Super Power Rail, TSMC's backside power delivery architecture. This fundamentally changes how power is distributed inside the chip. ⚡ Instead of sharing the same side of the silicon for both signals and power, A16 separates the two. Signal routing remains on the front side, while power delivery moves to the back side of the wafer. The result is a more efficient design that reduces power congestion and enables higher performance without increasing chip area. 📊 According to TSMC, compared with N2P, A16 offers: • 8–10% higher performance • 15–20% lower power consumption at the same performance • Around 10% higher logic density 💡Perhaps more importantly, these improvements come without changing the overall layout footprint. This explains why backside power delivery is becoming one of the industry's most anticipated packaging and process technologies. 🏭 If the reports are accurate, NVIDIA's Rosa would join a growing list of next-generation processors that leverage TSMC's most advanced manufacturing technologies. Meanwhile, AMD's EPYC Venice and Intel's Nova Lake are already expected to utilize TSMC's N2P process. 🔥 The competition in AI computing is no longer defined solely by CPU architecture. Process technology, power delivery, and advanced packaging are becoming equally important in determining how much performance can be extracted from each generation of silicon. #NVIDIA #TSMC #A16 #2nm #Semiconductor #CPU #AIInfrastructure #DataCenter #Arm #Packaging #VDO #VDODistributor #MinhNguyenVan #CEOVDO VDO JSC, 🔸 [email protected] 🔸 https://dis.vdo.com.vn