Source: Wall Street See
Analysts believe that, led by Apple, the large-scale model of end devices will slowly become popular, and the demand for hardware replacement can stimulate a new cycle in the consumer electronics industry.
At the WWDC conference in June, Apple released its latest Apple Intelligence system. The giants who have been watching for a long time finally launched their first shot at AI. Although Apple Intelligence has not yet landed in the test system, its practicality is still questionable. However, at least the symbolic significance of Apple products fully integrating AI has brought about a continuous increase in stock prices. As for the product structure, the operating income of 10-30 billion yuan products is 401/1288/60 million yuan respectively.$Apple (AAPL.US)$Morgan Stanley analysts, including Erik Woodring, pointed out in their latest report that the results of the WWDC conference exceeded expectations and are expected to accelerate a new round of Apple device upgrades, which is bullish for the Apple supply chain such as Foxconn.
Please use your Futubull account to access the feature.$Taiwan Semiconductor (TSM.US)$and Foxconn and other fruit chain companies.
WWDC results exceeded expectations, which is bullish for the Apple supply chain.
Analysts believe that overall, the results of WWDC slightly exceeded expectations, laying an important foundation for Apple to promote device upgrades from fiscal year 25. Since the latest AI feature of the Apple Intelligence system can only be used on devices with A17 Pro and M-series chips, which account for approximately 8% of existing Apple devices, in order to experience Apple's AI, it is necessary to replace the device. Morgan Stanley expects that the release of WWDC Apple Intelligence is expected to drive a wave of Apple device upgrades.
To experience Apple's AI, it is necessary to replace the device. Morgan Stanley expects that the release of WWDC Apple Intelligence is expected to drive a wave of Apple device upgrades.
Unlike other technology giants that rely heavily on chips for AI training and inference, the server chip of the Apple Intelligence system is very special. It uses the company's own M-series chip. Morgan Stanley speculates that it's the M2 Ultra chip, which is expected to contribute significant revenue to Taiwan Semiconductor. "Our recent checks with the supply chain suggest that Apple may have produced about 2 million M2 Ultra chips for AI servers in the first half of 24."$NVIDIA (NVDA.US)$In view of the expanding user group of private cloud computing, we expect that Apple will use 3-nanometer M3 or M4 for AI server chips in 2025. In 2026, we believe that Apple may use TSMC's 2-nanometer and SoIC technology to use more powerful Apple chips in AI servers.
In view of the expanding user group of private cloud computing, we expect that Apple will use 3-nanometer M3 or M4 for AI server chips in 2025. In 2026, we believe that Apple may use TSMC's 2-nanometer and SoIC technology to use more powerful Apple chips in AI servers.
The M2 Ultra chip adopts TSMC's InFO LSI package to splice two M2 Max chips together (using TSMC's 4nm wafer process). We estimate that an M2 Ultra chip can bring TSMC $350-400 in revenue. This means that the Apple AI server chip may bring TSMC up to $2 billion in revenue in 2024, accounting for about 2% of TSMC's total revenue.
In view of the expanding user group of private cloud computing, we expect that Apple will use 3-nanometer M3 or M4 for AI server chips in 2025. In 2026, we believe that Apple may use TSMC's 2-nanometer and SoIC technology to use more powerful Apple chips in AI servers.
Apple's A-series chip will perform edge computing tasks on iPhones. According to information released at the WWDC conference, Apple has launched a device-side large language model with about 3 billion parameters, which can only be used on A17 Pro chips. Currently, only the flagship model iPhone 15 Pro is equipped with an A17 Pro chip.
Morgan Stanley pointed out that the upcoming Apple iPhone 16 base model will use the A18 processor, and the high-end iPhone 16 Pro model may be equipped with a newly designed A18 Pro, which may be 15-20% larger than the A18 to accommodate more graphics and artificial intelligence computing units.
Considering the memory requirements of AI services, the memory chip industry will also benefit from this news. The industry expects that if the basic model remains at around 3 billion parameters, the DRAM capacity of the new iPhone 16 base model is expected to be upgraded from 6GB on iPhone 15 to 8GB (the minimum configuration required to drive Apple's large models on the client side), while the DRAM capacity of iPhone 16 Pro will remain at 8GB. Considering that the memory density of the M2 chip is limited (192GB), the growing number of Apple AI servers will consume a large amount of LPDDR5 memory.
Looking to the future, Morgan Stanley pointed out that with Apple's strong research and development capabilities and its collaboration with OpenAI, the next step is to observe whether the AI experience of Chinese Android manufacturers can catch up with the user experience of iOS18 to test Apple's technical strength. However, overall, analysts believe that under the leadership of Apple, device-side large models will gradually become popular, and the need for hardware upgrades will drive a new cycle in the consumer electronics industry.
Global AI-related cloud computing spending is expected to reach $300 billion in 2030.
Morgan Stanley also listed this year's AI server procurement data in the report. As shown in the figure below, overall, large-scale enterprises are more active in purchasing AI servers. Among all enterprises, the one with the most increased AI spending is Asia, and only other companies have reduced the allocation of AI in their capital spending.$Tesla (TSLA.US)$, only$Microsoft (MSFT.US)$reduced the allocation of artificial intelligence in its capital expenditure.
After analysis by analysts, global AI-related cloud computing expenditure will reach $300 billion in 2030, and capital expenditure on AI chips will reach $230 billion, with AI hardware reaching $70 billion. The customized application-specific integrated circuit (ASIC) market will reach $80 billion by 2030, and the design services market will reach $40 billion.
Editor/tolk