After being briefly surpassed by artificial intelligence chip leader Nvidia last week, consumer electronics leader Apple launched a counterattack. The day after the release of its artificial intelligence strategy at the World Developers Conference (WWDC), the company's stock price rebounded significantly by 7.26% on June 11 local time in the United States, with its market value rising to $3.18 trillion, regaining its position as the second largest market value in the US stock market. Nvidia has entered a new round of expected adjustment, with its market value falling below $3 trillion, making it difficult for the two tech giants to be evenly matched.
Compared to the booming Nvidia and Android manufacturers that have already laid out AI phones, Apple's current wave of generative artificial intelligence has been somewhat quiet. Finally, at the latest developer conference, Apple released the "Apple Intelligence" suite, a new personal intelligence system deeply integrated into iPhones, iPads, and Macs.
Although the generative AI features of the Apple series are mainly focused on document summarization, image processing, information retrieval, text translation, and other functions, and are referred to as "homogenization" functions by Android phone manufacturers, with its strong system level application integration ability and cooperation with external support for Open AI, Apple still maximizes personalization, integration, and other characteristics, demonstrating the momentum of accelerating the embrace of end-to-end artificial intelligence.
On the other hand, the news that Apple is holding hands with the generative artificial intelligence giant Open AI and its intelligent assistant Siri to integrate into the GPT-4o model has also been confirmed, sparking strong criticism from Open AI competitors, artificial intelligence startup xAI, and Tesla founder Elon Musk, who believes that Apple is not smart enough to create its own artificial intelligence, but claims that Open AI can protect user security and privacy.
From the perspective of industry insiders, Musk's criticism of Apple's collaboration with Open AI regarding privacy remains to be debated. Of course, end-to-end full chain computing and storage encryption are more conducive to privacy protection. At this developer conference, Apple launched the Private Cloud Compute system, which is specifically designed to handle user task requirements. User data will not be stored in the cloud, and Apple indicates that it will not be accessed; Moreover, Apple is not without achievements in artificial intelligence. In April 2024, the Apple team released the open-source model OpenELM, which can run on end-to-end devices; The M4 chip released by Apple in May has increased CPU performance by 50% and GPU performance by four times, indicating that Apple has made efforts in AI from hardware to software.
The latest blog article introducing the basic model behind Apple's AI capability implementation shows that Apple Intelligence is composed of multiple powerful generative AI models, which were trained on the open-source project AXLearn framework released in 2023. The Apple research team also used a combination of data parallelism, tensor parallelism, and sequence parallelism to expand the training scale from multiple dimensions such as data, model, and sequence length. In addition, Apple is focusing on reducing memory requirements and inference costs, balancing memory, power consumption, and performance requirements in both server-side and server-side models.
Currently, Apple has adopted a step-by-step artificial intelligence strategy. Industry insiders believe that Apple's current focus must be on first developing end-to-end artificial intelligence, while cloud based artificial intelligence will prioritize cooperation and maintain an open attitude, and other large models will not be excluded in the future. On the one hand, the increase in knowledge density of large models and the enhancement of end-to-end chip capabilities make end-to-end AI layout possible; On the other hand, the power of the open source community is gradually increasing, and General Artificial Intelligence will constantly innovate the traditional mobile operating system. Even the Apple system, which is in a relatively closed ecosystem, needs to cooperate and catch AGI's "shuttle bus".
In the artificial intelligence arena, Nvidia and Apple are on different tracks, possessing different resource endowments. One provides AI underlying computing power and weapons, while the other holds a vast ecosystem of interconnected devices, capable of implementing end-to-end intelligence; As stated in Nvidia's latest speech by Huang Renxun, the next wave of development in artificial intelligence will shift from generative AI to physical AI. This means that AI systems will be able to better understand the physical laws and complex environments of the real world, promote the accelerated popularization of automated service robots, and the second battlefield of artificial intelligence has begun.
It is reported that NVIDIA provides a robot simulation training platform Omniverse, as well as supporting visual AI tools Metropolis and AI robot development platform Isaac. Among them, NVIDIA Isaac has been deployed in enterprise applications such as BYD Electronics, Siemens, Tereida, and Intrinsic.
Apple is also working hard in the dark. Earlier this year, it was reported that Apple has planned to enter the field of home robotics after abandoning its electric vehicle plan, and is still in the early stages of research and development. At the developer conference, Apple announced that Apple Home will support robot vacuum cleaners in iOS 18 and can activate and issue commands through Siri's voice control, which can also be seen as a certain level of physical intelligence.
However, the biggest boost to Apple's stock price at this stage may still rely on smartphones. Although analysts are divided on whether the introduction of generative AI features can drive the upgrade of iPhones, previous news showed that the shipment volume of iPhones in China increased by 52% in April compared to the same period last year. In addition, the demand for the new iPad launched in May is strong, and the use of cutting-edge AI chips seems to be stimulating consumer demand. Since the low stock price in April, Apple's stock price has rebounded by nearly 26%.