Apple CEO Tim Cook recently made the official announcement that the company’s inaugural spatial computing product, Vision Pro, is set to go on sale on February 2nd in the United States. According to a report by TechNews, due to the limited preparation of Vision Pro units, with an estimated quantity of only 60,000 to 80,000, there is a high likelihood that the product will sell out on the first day of its release.

TechNews cites the latest analysis from China’s TF Securities analyst Ming-Chi Kuo, who recently posted on X. According to Kuo, Apple has prepared only about 60,000 to 80,000 units of the Vision Pro before its launch, and due to the limited quantity, it is expected to quickly sell out once it hits the market.

Kuo also notes that despite questions about Apple’s lack of clear positioning and key features for the Vision Pro, as well as its relatively high price, the company is likely to sell out rapidly based on its groundbreaking technological innovations, creating a user experience that includes the illusion of mind-controlled interfaces. This, coupled with Apple’s strong core user base and a significant number of heavy users, should result in a swift sellout after the product’s release.

There are rumors suggesting that even though online ordering is possible, users may need to adjust the head circumference size, and Apple may prefer customers to try fitting it in person, which could lead to some consumers adopting a wait-and-see approach.

However, a user on the X platform, Aaron @aaronp613, discovered an encoding set as ‘ You will be able to use the Apple Store app to scan your face to pick the right size,’ indicating that users who are unable to visit stores for fitting may use the Apple Store app to scan their faces and order the correct Vision Pro size. Vision Pro requires a proper fit for functionality, involving a tight seal and correct headgear.

Apple has also introduced the Head Measure and Fit app to help developers test Vision Pro and determine the correct product size. Similar functionality may be built into the Apple Store app.

According to the latest data from TrendForce, global shipments of Apple Vision Pro are expected to reach 500,000 units in 2024.

(Image: Apple)

The recovery of the memory industry is evident, with Taiwanese companies such as Macronix, Nanya Technology, and Transcend all showing month-on-month revenue growth in December last year. Additionally, contract prices for DRAM and NAND Flash are expected to continue rising in the first quarter of 2024. However, the global second-largest memory manufacturer, SK Hynix, plans a expansion, introducing a variable element to the memory market.

According to a report by the Commercial Times, SK Hynix disclosed that it might reduce the scale of DRAM production cuts in the first quarter, while adjustments to the NAND Flash production strategy may occur in the second or third quarter, depending on the situation.

In response to major memory manufacturers’ expansion plans, Taiwanese memory firms believe that Hynix’s expansion should focus primarily on DDR5 and HBM (High-Bandwidth Memory) products. Nevertheless, Taiwan currently specializes in DDR4 products, and it is not expected to impact product pricing.

According to a press release from TrendForce published this week, the DRAM contract prices are estimated to increase by approximately 13–18% in 1Q24 with mobile DRAM leading the surge. It appears that due to the unclear demand outlook for the entire year of 2024, manufacturers believe that sustained production cuts are necessary to maintain the supply-demand balance in the memory industry.

For consumer DRAM, manufacturers are aggressively raising contract prices, which has prompted buyers to stockpile early. This has greatly improved purchasing momentum. However, the first quarter coincides with the industry’s off-season, and end sales are expected to be weak and lead to increased inventory levels due to buyers’ early stocking strategies.

Manufacturers generally believe that in 2024—with the expanding penetration of HBM and DDR5 each quarter—low-margin DDR4 capacity will be crowded out, thereby leading to shortages. As such, DDR4 contract prices are expected to outpace DDR3 in the first quarter by 10–15%. DDR3 continues to be supplied by Taiwanese manufacturers, and with generally high inventory levels, its contract price increase is estimated at 8–13% for 1Q24.

(Image: SK Hynix)

After the U.S. authorities strengthens export restrictions on high-end processors, reports have emerged that Chinese companies are actively changing the purpose of PC gaming chips and utilizing them for the development of AI tools.

According to a report from the Financial Times on January 10, factory managers and chip buyers familiar with the details disclosed that every month, thousands of NVIDIA gaming cards are being disassembled in factories and workstations. The core components are then installed onto new circuit boards.

A factory manager further indicated that in December 2023 alone, their workers disassembled over 4,000 NVIDIA gaming cards, more than four times the quantity of November.

These modified components are primarily supplied to listed companies and small AI laboratories. They are rushing to accumulate a sufficient supply of NVIDIA server chips before the export controls take effect in the United States.

Industry sources have reportedly warned that modifying NVIDIA products would violate the company’s intellectual property rights, and certain gaming cards could be subject to bans at any time.

NVIDIA’s most powerful gaming card, the “GeForce RTX 4090,” is a popular choice for modification, but it is now prohibited from being sold in China. In December 2023, NVIDIA released a throttled-down version for China, the “GeForce RTX 4090 D,” which is 5% slower than versions available in other regions.

A factory manager has indicated a “significant” performance difference between the 4090 D and the regular 4090, suggesting that the downgraded version may not be suitable for training large language models (LLMs).

Although NVIDIA has developed three versions of AI chips specifically designed for China (expected to be launched in March), they face reluctance from Chinese customers due to weaker performance compared to the previously available versions in China, coupled with pricing that is almost similar to the more powerful but banned versions.

The recent report from The Wall Street Journal also addressed the lack of interest in NVIDIA’s downgraded models by Chinese customers. According to the data from TrendForce, currently, around 80% of the high-end AI chips used by Chinese cloud computing companies are sourced from NVIDIA. However, in the next five years, this proportion may decrease to 50% to 60%.

Read more

• [News] Cooling Response to NVIDIA’s Exclusive Chips for China, Lack of Interest in Downgraded Models by Customers
• [News] NVIDIA’s China-exclusive Chip H20 to Begin Production in Q2 by Wistron

(Photo credit: NVIDIA)

Samsung Display has unveiled the RGB version of Micro OLED (OLEDoS) for the first time at CES 2024, presenting the industry’s reportedly highest-resolution RGB OLEDoS display screen.

As per Samsung Display’s news release, Samsung Display has indicated that Micro OLED achieves high-definition displays with small pixel sizes, achieved by applying organic materials to silicon wafers, playing a crucial role in the increasingly popular XR headsets.

Although the Micro OLED is only 1.03 inches, it has a pixel density of 3,500 PPI, making it the industry’s highest resolution RGB OLEDoS display, which utilizes red, green and blue OLEDs on silicon wafers to generate color without the need for a separate light source.

Samsung has previously acquired Micro LED developer eMagin last year. In addition, there are reports indicating that Apple is developing the second generation of Vision Pro, expected to be unveiled in 2027. It is rumored to adopt advanced RGB OLEDoS technology, and the acquisition of eMagin by Samsung also positions them to provide the necessary technology for Apple.

If Apple is interested in upgrading its Vision Pro 2 display tech, Samsung’s exhibit at CES 2024 could be considered a showcase, proving that its Micro OLED technology is gradually gaining ground, and that it has a potential chance of replacing Sony as a Vision Pro 2 display supplier.

Read more

• [News] Matching Apple’s Vision Pro Momentum? Rumors Suggest Samsung Expediting XR Device Launch in Early 2024
• [News] Apple’s Vision Pro Finally Set to Hit the Shelves – Available in the U.S. from February 2nd, 2024

(Photo credit: Samsung)

On January 9th, Intel unveiled its latest automotive AI chips, entering into direct competition with rivals NVIDIA and Qualcomm in the automotive chip market. In a bid to strengthen its position, Intel also announced the acquisition of automotive chip company Silicon Mobility.

Reportedly, Intel stated that Silicon Mobility, a French startup, specializes in designing System-on-a-Chip (SoC) technology for controlling electric vehicle motors and in-car charging systems, along with software. The acquisition amount was not disclosed by Intel.

As per Reuter citing from Intel’s automotive business chief Jack Weast, he has indicated that, intel’s new automotive system on a chip products will adapt the company’s recently launched AI PC technology for the durability and performance requirements of vehicles.

Weast further clarified, “Intel will not require automakers to use advanced driving chips designed by its former Mobileye unit, he said. Instead, automakers can have Intel incorporate their own chiplets to enable specific functions into the Intel system at a lower cost.”

Intel’s chips designed for infotainment systems are already integrated into 50 million vehicles. As the automotive chip market continues to expand, the demands on chips are increasing, covering technologies such as autonomous driving, upgradable in-car system software, and complex dashboard displays amid strong competition from NVIDIA and Qualcomm.

Weast has addressed ahead of the CES technology show in Las Vegas that Chinese automaker Zeekr will be the first automaker to use Intel’s AI system on a chip to create “an enhanced living room experience” in vehicles, including AI voice assistants and video conferencing. Zeekr, an electric vehicle brand under the Geely Holding Group, is a customer of both Intel and NVIDIA.

Intel will try to separate itself from rivals by offering chips that automakers can use across their product lines, from lowest-priced to premium vehicles, Weast said.

According to Reuter, Weast addressed reporters in a conference call before the announcement at the CES technology show in Las Vegas, stating, “Intel has done a pretty terrible job communicating our success in automotive, We are going to change that.”

Read more

• [Insights] Infinite Opportunities in Automotive Sector as IC Design Companies Compete for Self-Driving SoC

(Photo credit: Intel)