According to a report by Taiwan’s Money DJ, there’s good news from TSMC regarding its 3nm node. Sources within the supply chain have disclosed that the number of new chip designs using the 3nm process, known as “New Tape-Outs” (NTOs), has surged. It’s confirmed that customers including MediaTek, AMD, NVIDIA, and Qualcomm will follow in Apple’s footsteps for mass adoption of the 3nm process in the next year (2024) and the subsequent year. By the second half of next year, the monthly production capacity for the 3nm family, including N3E, will increase from the current approximately 60,000 wafers to 100,000 wafers.

According to publicly available information from TSMC, the company began volume production of its first 3nm process node, N3, in the second half of last year. The enhanced version of the 3nm process, N3E, started production in the latter half of this year. There will also be extensions to the 3nm process, including N3P, N3S, and N3X. This year, Apple’s high-end A17 Pro chip for its iPhones was based on the initial N3 process.

Both TSMC and MediaTek previously announced their collaboration, with MediaTek developing new Dimensity products using TSMC’s 3nm process. The design phase, known as “Tape Out,” has been successfully completed, and mass production is scheduled for next year. Industry reports indicate that aside from Apple and MediaTek, AMD, NVIDIA, and Qualcomm are also confirmed to adopt the N3 family of processes. Intel is also on the list, with mass production planned for the year after next.

TSMC’s first-generation 3nm process currently has a monthly production capacity of about 60,000 wafers, serving Apple as its primary customer. TSMC has initiated a program known as “Continuous Improvement Plan” (CIP) for the 3nm process, referred to as N3B in the industry. Supply chain sources suggest that N3B’s capacity will be integrated into subsequent extended process nodes, such as N3E, which is expected to attract more customers. It is estimated that the overall 3nm monthly production capacity will reach 100,000 wafers by the second half of next year.

(Photo credit: TSMC)

According to a report by Taiwan’s Economic Daily, TSMC is aggressively expanding its advanced packaging capacity. Recently, they placed an additional 30% order for equipment with manufacturers, leading to a doubling of order volumes for companies in the interposer supply chain, such as UMC and ASE Group. Moreover, there are rumors of price increases on the horizon.

TSMC, responding to the robust production demands from major clients like NVIDIA, AMD, and Amazon, has not only expanded its CoWoS capacity as originally planned but has also added another 30% in new equipment. This move implies that once TSMC’s new advanced packaging capacity comes online next year, it will represent at least a doubling of the current production capacity.

Given the substantial demand for TSMC’s advanced packaging orders, coupled with the necessity of interposer in CoWoS for stacking logic computing ICs and high-bandwidth memory, there is an expectation of significant growth in interposer orders, likely exceeding a twofold increase compared to this year. Notably, semiconductor giants like UMC and ASE Group have already secured significant orders from TSMC for interposer layers and are currently in the mass production and delivery stages.

It is understood that UMC, after venturing into the advanced packaging market in recent years, has introduced packaging solutions applicable to IoT, automotive chips, and more, spanning from wafer bumping and wire bonding packaging to advanced 2.5D, 3DIC, and wafer-level fan-out packaging solutions. The most notable among them is the 2.5D silicon interposer layer solution, which, through collaboration with UMC and other specialized packaging facilities, has proven pivotal in securing major orders in the interposer layer market, particularly from NVIDIA.

Industry sources suggest that UMC has increased prices for super hot run interposer layer orders and initiated capacity expansion plans to meet customer demands, while ASE Group is also contemplating adjustments to its advanced packaging quotations.

(Photo credit: UMC)

According to a report by Taiwan’s Economic Daily, TSMC’s CoWoS advanced packaging capacity is running at full throttle. As they actively expand their production capabilities, there are reports of major customers like NVIDIA increasing their orders for AI chips. Additionally, industry giants like AMD and Amazon have rushed in with urgent orders.

In response to this urgent situation, TSMC is actively seeking equipment suppliers to expand its CoWoS machine procurement. Beyond TSMC’s existing production expansion goals, the company is further increasing its orders for equipment by an additional 30%, highlighting the ongoing fervor in the AI market.

It is reported that TSMC has sought assistance from equipment manufacturers such as Scientech, Allring, Grand Process Technology, E&R Engineering, and GP Group for this endeavor. They plan to complete the delivery and installation of the equipment by the first half of the coming year. The related equipment manufacturers are experiencing a surge in activity.

Industry sources reveal that TSMC’s CoWoS advanced packaging monthly production capacity is currently around 12,000 units. With their previous expansion efforts, they aimed to gradually increase this to 15,000 to 20,000 units per month. Now, with the addition of more equipment, they are looking at the possibility of reaching capacities of over 25,000 units per month, potentially even approaching 30,000 units. This substantial increase in production capacity positions TSMC to handle a significantly larger volume of AI-related orders.

Equipment providers have pointed out that NVIDIA is currently TSMC’s largest customer for CoWoS advanced packaging, accounting for 60% of the production capacity. Recently, in response to robust demand in AI computing, NVIDIA has increased its orders. Additionally, urgent orders from other customers such as AMD, Amazon, and Broadcom have started to pour in.

(Photo credit: TSMC)

Fueled by an AI-driven inventory stocking frenzy across the supply chain, TrendForce reveals that Q2 revenue for the top 10 global IC design powerhouses soared to US $38.1 billion, marking a 12.5% quarterly increase. In this rising tide, NVIDIA seized the crown, officially dethroning Qualcomm as the world’s premier IC design house, while the remainder of the leaderboard remained stable.

AI charges ahead, buoying IC design performance amid a seasonal stocking slump

NVIDIA is reaping the rewards of a global transformation. Bolstered by the global demand from CSPs, internet behemoths, and enterprises diving into generative AI and large language models, NVIDIA’s data center revenue skyrocketed by a whopping 105%. A deluge of shipments, including the likes of their advanced Hopper and Ampere architecture HGX systems and the high-performing InfinBand, played a pivotal role. Beyond that, both gaming and professional visualization sectors thrived under the allure of fresh product launches. Clocking a Q2 revenue of US$11.33 billion (a 68.3% surge), NVIDIA has vaulted over both Qualcomm and Broadcom to seize the IC design throne.

Qualcomm’s Q2 took a hit as the Android smartphone sector grappled with dwindling demand and Apple’s modem pre-purchases resulted in a subdued seasonal rhythm. Consequently, their revenue slid by 9.7%, rounding off at about US$7.17 billion. Broadcom, while benefiting from AI-ignited demand for high-end switches and routers, faced headwinds with revenue drops in server storage, broadband, and wireless. The result was a second-quarter revenue that essentially mirrored the previous quarter at around US$6.9 billion.

AMD’s Q2 performance plateaued at about $5.36 billion, weighed down by a slump in gaming GPU sales and its embedded segment operations. Conversely, MediaTek, after several quarters of inventory recalibration, witnessed a resurgence with components like TV SoCs and Wi-Fi stabilizing. The added impetus of urgent TV orders and escalating shipments for mobile phones, smart platforms, and power management ICs boosted MediaTek’s Q2 to a solid US$3.2 billion.

Marvell, though buoyed by AI deployments in data centers, faced headwinds with a decline in On-Premise Servers (enterprise private clouds). End-user demand remained frail, and with sectors like data centers, telecom infrastructure, and enterprise networking facing revenue drops, Marvell’s Q2 took a 1.4% hit, culminating at roughly $1.33 billion.

Taiwan’s IC design stalwart Novatek flourished as customers replenished TV-related inventories and ushered in novel products such as OLED DDI. Realtek, drawing strength from supply chain restocking of PC/NB-centric ICs, reported quarterly growths of 24.7% and 32.6%, respectively. Yet, without substantial signs of a holistic revival in end-sales and inventory restocking, growth in H2 seems set to face challenges.

Will Semiconductor secured the ninth spot with a Q2 revenue of $528 million, registering a modest decline of about 1.9%. Hot on its heels is the US-based power IC maestro, MPS, with its Q2 revenue tallying up to $441 million—a slip of approximately 2.2%.

Peering into Q3, while inventory levels across companies paint a rosier picture than H1, a pervasive end-user demand slump urges caution. However, a silver lining emerges with CSPs, internet titans, and private firms flocking to generative AI and large language models. As these high-value AI offerings gain traction, TrendForce projects that the top ten global IC design giants will continue their double-digit ascent in Q3, potentially reaching record-breaking figures.

For more information on reports and market data from TrendForce’s Department of Semiconductor Research, please click here, or email Ms. Latte Chung from the Sales Department at lattechung@trendforce.com

According to a report by Taiwanese media TechNews, industry sources have indicated that Microsoft has recently reduced its orders for Nvidia’s H100 graphics cards. This move suggests that the demand for H100 graphics cards in the large-scale artificial intelligence computing market has tapered off, and the frenzy of orders from previous customers is no longer as prominent.

In this wave of artificial intelligence trends, the major purchasers of related AI servers come from large-scale cloud computing service providers. Regarding Microsoft’s reported reduction in orders for Nvidia’s H100 graphics cards, market experts point to a key factor being the usage of Microsoft’s AI collaboration tool, Microsoft 365 Copilot, which did not perform as expected.

Another critical factor affecting Microsoft’s decision to reduce orders for Nvidia’s H100 graphics cards is the usage statistics of ChatGPT. Since its launch in November 2022, this generative AI application has experienced explosive growth in usage and has been a pioneer in the current artificial intelligence trend. However, ChatGPT experienced a usage decline for the first time in June 2023.

Industry insiders have noted that the reduction in Microsoft’s H100 graphics card orders was predictable. In May, both server manufacturers and direct customers stated that they would have to wait for over six months to receive Nvidia’s H100 graphics cards. However, in August, Tesla announced the deployment of a cluster of ten thousand H100 graphics cards, meaning that even those who placed orders later were able to receive sufficient chips within a few months. This indicates that the demand for H100 graphics cards, including from customers like Microsoft, has already been met, signifying that the fervent demand observed several months ago has waned.

(Photo credit: Nvidia)