Taiwan Semiconductor Manufacturing Company (TSMC) has been recognized as a prominent producer of advanced semiconductors for a considerable period.
According to a recent report from DigiTimes, its latest innovation, the performance of TSMC 3nm node, is outperforming expectations.
In Q4 2022, Taiwan-based semiconductor company TSMC commenced the mass production of chips utilizing its recently developed N3 or 3nm process node.
One company that could benefit significantly from the 3nm node is Apple, known for its tight control over its supply chain and reliance on TSMC to manufacture its processors. Despite being pricier than the existing 5nm node, TSMC’s 3nm process node is expected to have a notable customer in Apple. Apple has already ordered a the entire initial supply of N3.
The 3nm process node is better than initially anticipated because the technology will likely offer improved performance and energy efficiency compared to earlier expectations.
The version of this process node has higher expectations, which is expected to help TSMC to continue its dominance in the field of process node technology.
It will utilize FinFET technology with a gate pitch of 42nm and offer significant advancements in transistor density and power efficiency over the previous 5nm and 7nm nodes. The TSMC 3nm node is expected to be used for high-performance computing applications such as mobile devices, artificial intelligence, and high-performance computing.
Customers in the PC market are facing a twofold challenge, with the increasing expenses linked to adopting the latest semiconductor manufacturing nodes and a shrinking market with substantial drops in demand. Due to these circumstances, the cost factor is anticipated to be a significant concern for PC customers, who may require assistance.
However, Intel is delaying its orders for TSMC’s 3nm chips until the fourth quarter of 2024. But the company may only be ready to produce or release devices that utilize these chips at some point after this date.
The reasons for this delay have yet to be clarified, but they could be related to several factors, such as manufacturing challenges or changes in market demand. Still, it’s worth noting that the company has made a significant push in recent years to manufacture more of its chips in-house rather than relying on external foundries like TSMC.
This strategy shift may be playing a role in the decision to delay Arrow Lake’s production. Intel may need help to keep pace with TSMC’s rapid advancements in semiconductor manufacturing technology.
Intel intends to outsource the manufacturing of its 15th Gen Arrow Lake GPU to an external foundry. The N3 process node is considered the leading candidate for the GPU tile. This choice implies that Arrow Lake will probably be accessible in considerable quantities and ready to ship in the second half of 2025, indicating a substantial postponement in its launch.
Arrow Lake is the upcoming 15th generation of Intel’s product lineup, and it is expected to bring significant improvements over its predecessors. It is compatible with all systems that can support Meteor Lake, and it comes with upgraded Redwood Cove and Crestmont cores, which will be replaced with new Lion Cove and Skymont cores. This enhancement is expected to provide a considerable advantage, mainly due to the increased core count, projected to range from 40 to 48, including eight P-Cores and 32 E-Cores. This means Arrow Lake can handle more tasks simultaneously, improving performance and overall user experience.
In a surprising move, Intel has decided to skip the “Intel 4” node and opt for the 20A node for its upcoming Arrow Lake processors. However, it should be noted that despite this significant shift in node technology, both the Meteor Lake and Arrow Lake chips will still utilize TSMC’s N3 process node for additional core IPs, specifically the Arc GPU cores. Intel’s new 20A node is equipped with advanced RibbonFET and PowerVia technology, which is expected to provide a 15% improvement in performance per watt. The production of the first IP test wafers is anticipated to commence in fabs during the second half of 2022.
Overall, the news that TSMC’s 3nm node is outperforming expectations is undoubtedly good news for the industry. The prospect of faster, more efficient chips at a lower cost is enticing, and it’s clear that TSMC’s technology is at the forefront of this trend.
However, the news about Intel’s delay serves as a reminder that even the largest and most established players in the semiconductor industry are subject to the whims of technological progress. That competition in this field is as fierce as ever.