BTC $68,430.61 -6.98%
ETH $1,919.02 -6.17%
BNB $630.79 -6.55%
XRP $1.00 -3.60%
SOL $78.33 -5.48%
TRX $0.3104 -2.14%
DOGE $0.0633 -5.55%
ADA $0.1671 -3.15%
BCH $214.73 -6.12%
LINK $9.68 -3.89%
HYPE $70.30 -2.10%
AAVE $109.87 -7.09%
SUI $0.7678 -6.03%
XLM $0.1424 -2.51%
ZEC $891.91 -8.86%
AAPL $356.18 -0.03%
AMZN $267.92 +0.50%
GOOGL $300.56 +1.78%
MSFT $520.48 -0.16%
META $624.77 -0.69%
NVDA $164.34 +2.13%
TSLA $398.44 +3.09%
SNDK $922.44 -2.49%
INTC $61.13 +1.23%
SPCX $140.50 +0.73%
MU $793.85 +1.26%
AMD $575.27 +2.21%
BTC $68,430.61 -6.98%
ETH $1,919.02 -6.17%
BNB $630.79 -6.55%
XRP $1.00 -3.60%
SOL $78.33 -5.48%
TRX $0.3104 -2.14%
DOGE $0.0633 -5.55%
ADA $0.1671 -3.15%
BCH $214.73 -6.12%
LINK $9.68 -3.89%
HYPE $70.30 -2.10%
AAVE $109.87 -7.09%
SUI $0.7678 -6.03%
XLM $0.1424 -2.51%
ZEC $891.91 -8.86%
AAPL $356.18 -0.03%
AMZN $267.92 +0.50%
GOOGL $300.56 +1.78%
MSFT $520.48 -0.16%
META $624.77 -0.69%
NVDA $164.34 +2.13%
TSLA $398.44 +3.09%
SNDK $922.44 -2.49%
INTC $61.13 +1.23%
SPCX $140.50 +0.73%
MU $793.85 +1.26%
AMD $575.27 +2.21%

city

All
Article
Flash

first_img Supply chain Chuanhuida slows down chip orders, MediaTek and World Semiconductor are expected to gain more production capacity

The Industrial and Commercial Times reported on November 10 that the supply chain has recently indicated that NVIDIA intends to slow down chip orders. The report stated that NVIDIA's CEO Jensen Huang admitted to infidelity and apologized in a letter to employees on the 8th, U.S. time. The board has initiated an independent investigation to clarify whether it has been targeted by foreign intelligence activities. Foreign media disclosed that some companies authorized to handle the H200 are suspected of being shell companies, raising concerns that chips may be transshipped to mainland China through shell companies and third locations.The report indicated that stricter export reviews and the accumulation of H200 processors in Hong Kong have increased the risk of NVIDIA cutting orders, with upstream supply chain players such as TSMC and ASE being the first to bear the brunt. The supply chain revealed that there have also been rumors of NVIDIA cutting orders in wafer foundry recently, stating that this would benefit other chip manufacturers in obtaining more capacity, including AMD and cloud service providers' self-developed chips. ASIC manufacturers indicated that as long as TSMC can provide capacity, end customers will fully take on the orders. Industry sources pointed out that if NVIDIA releases capacity, companies like WorldChip, Creative, and MediaTek are expected to gain more capacity support.MediaTek estimates that the first AI accelerator aims for approximately $16 billion by 2027 and has already secured a second customer. WorldChip has also obtained a second ASIC order from a major American company and anticipates that revenue will gradually expand next year and the year after. Demand for AMD's data center business continues to exceed existing supply. AMD CEO Lisa Su stated in a recent earnings call that large cloud service providers are expanding EPYC deployments, with increased demand for AI accelerators and cabinets. The Instinct MI455X has been shipped in large quantities, and multiple cloud customers and computing power buyers have activated the first batch of Helios.

Meritz stated that storage orders from 2027 to 2028 far exceed capacity

Meritz Securities, a South Korean securities company, pointed out that the orders received by storage chip manufacturers for 2027 to 2028 far exceed their production capacity. Even if the Democratic Party in the United States gains an advantage in the midterm elections, leading to delays or adjustments in some data center projects, it is unlikely to fundamentally change the supply shortage of storage chips.Meritz stated that the order delivery rate for storage chips this year (the ratio of available production capacity to order volume) has remained at a low level of 80% to 85%, driving a significant increase in storage chip prices. It is expected that the supply-demand gap will further widen in 2027, and the order delivery rate may drop to around 80%. During a recent visit to South Korea, OpenAI CEO Sam Altman requested SK hynix and Samsung Electronics to provide customized storage products such as HBM for 2027 to 2030, with demand exceeding the available production capacity of both companies.Community user Teng Yan added that Samsung stated during a conference call with Meritz on October 29 that 10 large customers have signed multi-year storage supply agreements, and an equal number of customers are in negotiations. The demand for storage customers has also expanded from cloud service giants and AI server manufacturers to more fields. Therefore, the slowdown in U.S. data center approvals will only affect some orders and is unlikely to change the overall tight supply-demand situation.

first_img Delphi Digital: Electricity is a greater constraint on AI computing power deployment than chips

On October 29, the cryptocurrency research institution Delphi Digital released a report titled "The AI Boom Has a Power Problem," authored by Accelxr. The report states that deployable AI computing power depends on chip supply, as well as sites that are powered, cooled, and connected to the grid. Electricity is harder to supplement in the short term than chips, as chip production capacity can expand quarterly, while new power generation and transmission take years.The report cites data indicating that the PJM capacity auction has cleared at the price cap for three consecutive times, with the last two still falling short of reliability targets by 6.6 to 6.8 GW, marking the first time such a gap has occurred in the history of this auction. The scale of large load queues in Texas ERCOT increased from 63 GW at the end of 2024 to 474 GW in June, with about 90% being data centers; the governor requested an audit of the data centers in the queue in August, and ERCOT subsequently suspended power approval for data centers and cryptocurrency miners of 75 MW and above, including 17 projects that had passed stability studies.The report also states that U.S. data centers are expected to account for about 4.7% of the nation's electricity consumption in 2024, with LBNL predicting an increase to 9.5% to 15.3% by 2030, and EPRI estimating a maximum of 17%. In a 1 GW facility, chip depreciation is about $2.8 billion to $3.3 billion per year, equivalent to about $8 million per day. SemiAnalysis predicts that by 2028, the U.S. will have over 40 GW of data center capacity utilizing behind-the-meter power sources; the delivery cycle for large gas turbines is about five years, with GE Vernova, Siemens Energy, and Mitsubishi Heavy Industries forming an oligopoly. Wood Mackenzie expects gas turbine prices to reach $600 per kilowatt by the end of 2027, nearly three times the level of 2019.

first_img Anthropic's prospectus names TSMC, warning that advanced chip production capacity is concentrated in Taiwan

The AI laboratory Anthropic, which developed Claude, named TSMC in its IPO prospectus, warning that the world's most advanced chip manufacturing capacity is highly concentrated in Taiwan. The prospectus states that if TSMC's operations are disrupted, it could severely impact the supply of AI chips and affect Anthropic's business, financial condition, and operational results.According to MoneyDJ citing the prospectus, the global AI industry heavily relies on TSMC's advanced processes. Suppliers such as NVIDIA and AMD, as well as companies that design their own chips like Amazon and Google, all need TSMC to turn their designs into physical chips. The listed risks include earthquakes, typhoons, interruptions in power and water supply, as well as cross-strait geopolitical tensions, military conflicts, blockades, tariffs, and trade restrictions.The prospectus also states that if a significant event occurs in Taiwan, the supply of advanced chips could be severely hindered; due to the time-consuming nature of building and certifying alternative foundries, the impact could last for several years. The expansion of advanced processes outside Taiwan is still in its early stages, making it difficult to effectively reduce concentration risks in the short term. TSMC has been unable to fully meet market demand due to strong demand, resulting in Anthropic not being able to obtain a sufficient supply of chips in a timely manner.
app_icon
ChainCatcher Building the Web3 world with innovations.