Amidst the AI artificial intelligence frenzy, NVIDIA's founder and CEO, Jensen Huang, coming to Taiwan has become the center of attention in the media. Even the activities of domestic IC design giant MediaTek had to take a backseat. Earlier, NVIDIA and MediaTek announced their collaboration in the automotive electronics market, causing the media conference of MediaTek to lose focus. The attending reporters were only concerned about when Jensen Huang would arrive at MediaTek's venue. Consequently, MediaTek had no choice but to reveal related information on the spot, stating that MediaTek's flagship mobile processor, the Dimensity 9300, would be officially unveiled in October 2023, igniting competition against Qualcomm's Snapdragon 8 Gen 3.
The fourth quarter of each year marks the beginning of the competition between MediaTek and Qualcomm in the next-generation flagship mobile processors. In 2022, MediaTek's Dimensity 9200 went head-to-head with Qualcomm's Snapdragon 8 Gen 2, followed by the Dimensity 9000 against Snapdragon 8 Gen 1 in 2021. Therefore, this sequel to the ongoing drama will continue to unfold in 2023, with MediaTek's Dimensity 9300 facing Qualcomm's Snapdragon 8 Gen 3. Although this news was expected in the market, the early exposure indicates MediaTek's eagerness to attract attention.
According to previous information, MediaTek's next-generation flagship mobile processor, named Dimensity 9300, will be manufactured using TSMC's N4P process. It will feature a tri-cluster architecture with ultra-large cores, big cores, and efficiency cores. The ultra-large core is expected to be based on Cortex X4, the big core might be Cortex A715, and the efficiency core might be Cortex A515. As for the first release, it is expected to be obtained by the Chinese smartphone brand Vivo for their X100 model.
The Dimensity 9300 utilizes TSMC's advanced N4P process. According to official data, compared to the previous generation Dimensity chip manufactured using the 5-nanometer (N5) process, performance is improved by 11% with the N4P process, and compared to the N4 process, performance is improved by 6%. The N4P process also increases power efficiency by 22% and transistor density by 6%. Since the N4P process belongs to the same family as the 5-nanometer process, it can be easily transferred to the 5-nanometer platform, reducing customer R&D costs and providing faster and more energy-efficient updates to the 5-nanometer platform.