"Smartphones enter a new era of AI: from Gemini to Apple, a storm of technological innovation".
Recently, Google's Gemini AI model, a super brain that claims to be able to easily surpass GPT-4, was officially unveiled. But what's really exciting is that Google's Pixel8 Pro phone is about to become the first smartphone to be equipped with this AI technology. This is not just a small upgrade in the mobile phone industry, but a "technical challenge" launched by Google to OpenAI, the leader in the AI industry.
When we talk about the combination of AI and smartphones, the entire industry has pressed the "fast forward button". For example, Apple's analyst Ming-Chi Kuo** may launch the iPhone 16 next year as Apple's first AI phone. At the same time, domestic smartphone giants such as Huawei, OPPO, and Honor are also in full swing preparing their own AI mobile phones.
These changes reveal a larger trend: every boom in the consumer electronics industry is driven by technological innovation. Now, with the advent of the AI wave, we are about to usher in another big change in the field of consumer electronics. Are you ready for a new round of "replacement waves"?
But that's only part of the story. Some analysts pointed out that the application of AI models in smartphones is expected to first promote the development of the upstream hardware market, such as the demand and value of key components such as SoC, storage and power management chips.
The global giants have already begun to act. A person from a large model manufacturer in China recently said: "Only when the model actually runs on the end device can generative AI penetrate into all aspects of consumers' daily lives." ”
Just two days ago, Google released their "most powerful and versatile model to date", Gemini. This multi-modal large model is known as "the catfish that surpasses GPT-4", bringing a new competitive landscape to the AI large model track.
Google's biggest advantage is that it has a large user base on PC and mobile, as well as a constant stream of real-time data (in contrast, OpenAI relies heavily on internet data). As a result, Google is expanding its strengths even further.
The company announced that the Pixel8 Pro will be the first phone compatible with Gemini Nano. Pixel8 Pro users will be able to enjoy two features supported by Gemini Nano: an audio recorder app for automatic summarization and smart replies for the Gboard keyboard. Both of these features will be rolled out along with the December feature update. And, since this model runs on the phone itself, users can enjoy the same smooth and native experience even when offline.
"The AI Revolution: The New Great Wave in the Smartphone World".
The AI revolution is causing a wave in the smartphone spaceOpen AI is not far behind, and it is reported that it is planning to cooperate with former Apple design master Jony Ive and Japan's SoftBank Group CEO Masayoshi Son to jointly develop a new AI hardware product around the "computing platform". They even put out an advertisement on their official website to recruit hardware engineers, which seems to be in earnest!
At the same time, Ming-Chi Kuo, the "prophet" in the fruit chain industry, also released fierce news, saying that Apple may add some "black technology" to the iPhone 16 in the wave of generative artificial intelligence that will be popular in 2023. Not only that, Mr. Guo also revealed that Apple is reshuffling the Siri team, seemingly to integrate AIGC features and large language models to make Siri smarter.
However, overseas manufacturers do not occupy the stage alone. Domestic smartphone giants have taken the lead in "running". Huawei CEO Yu Chengdong said at the developer conference: "We have entered the era of large models!"Huawei's Pangu model has been able to be used in Hongmeng 40 system, through the Xiaoyi assistant to achieve a variety of functions such as generation, automatic writing of email invitations and meeting minutes.
Honor CEO Zhao Ming was not to be outdone, and showed the AI model of the Honor Magic6 at the Qualcomm Snapdragon Summit, which claims to have a scale of 7 billion parameters and can provide personalized services according to user preferences. At the developer conference, OPPO launched the AndesGPT large model, which not only has strong conversation capabilities, but also can be customized to achieve device-cloud collaboration.
Lei Jun, the founder of Xiaomi, also revealed in his annual speech that Xiaomi mobile phones will fully embrace the AI model and focus on lightweight and local deployment. Vivo released the new OriginOS 4 system, equipped with a self-developed AI model.
According to industry insiders, AI mobile phones will respond to user needs faster than traditional mobile phones and provide more intelligent services. OPPO's Liu Haifeng even said that generative AI technology will make the phone understand you better.
This AI revolution is not only limited to smartphones, PCs, smart glasses, smart speakers, etc. will also become an important carrier of AI technology. IDC even **, by 2026, nearly half of the terminal equipment processors in the Chinese market will be equipped with AI technology.
Behind all this, the upstream hardware link will also have a great harvest. The mobile application of the AI model will promote the upgrade of SoC, storage, power management chips and other components. Huatai** analysis said that these hardware upgrades will have a significant impact on the BOM cost of mobile phones.
A person from a large model manufacturer said frankly: "This is just the beginning. He pointed out that the application scenarios of mobile AI large models are still very limited, and it remains to be seen whether the market demand is enough to support the promotion of this technology.
But it is undeniable that with the continuous advancement of AI technology, we are about to usher in a new round of smartphone revolution. With the blessing of AI, smartphones will no longer be just communication tools, but will become intelligent partners in our daily lives, able to understand our needs and intentions more accurately.
According to Canalys, the smartphone market has shown signs of recovery recently, despite a period of downturn. In the third quarter of 2023, the year-over-year decline in global smartphone shipments slowed, and the market shipments achieved double-digit sequential growth, driven by new product introductions and regional market recovery.
There is a similar trend in the domestic market. Canalys data shows that in the third quarter of 2023, the year-on-year decline in shipments in the domestic smartphone market has slowed down. This trend reflects that technological innovation is becoming a major driver of market growth.
Founder**'s research report pointed out that although the smartphone market has entered the stock era, the continuous decline in the past two years has left room for future growth. IDC** also shows optimism, predicting that the global smartphone market will grow by 6% year-on-year in 2024 and continue to recover in the next few years.
In short, the rapid development of AI technology is promoting the transformation of consumer electronics terminals to deep intelligence. Not only smartphones, but also PCs, smart glasses, smart speakers and other terminal devices will become the new battlefield of AI technology. This not only means that the smartphone market will usher in a new round of growth, but also brings the potential for revaluation of related hardware terminal devices.
In the AI era, we can expect to see more innovative and intelligent products come out, bringing unprecedented convenience and experience to our lives.
Driven by AI technology, the core component of smartphones, SoC (System on Chip), is undergoing an upgrade revolution. The SOC contains key units such as CPU, GPU, LPDDR, and Modem, among which the CPU, GPU, NPU, and ISP are the masters who process various types of data. Taking the development of Qualcomm Snapdragon processors as an example, in order to meet the growing demand for AI computing power, mobile phone SoCs must continuously upgrade AI engines and improve the computing power of NPUs, which directly promotes the improvement of their market value.
Storage is also an important battlefield in the AI era. According to data from Huatai**, in an AI inference task, only a 10 billion parameter model (FP16 accuracy) requires 186GB of memory, not including temporary variables for model calculations. It can be seen that in order to meet the needs of AI applications, more and more mobile phones are improving their operating memory. For example, in August this year, Xiaomi's K60 Extreme Edition and OnePlus's Ace2 Pro have both increased the memory to 24GB, providing more room for the development of AI applications.
The changes in power supply should also not be overlooked. As AI inference tasks continue to run, devices consume more power, which requires larger-capacity batteries and more advanced power management chips to meet the demand. In terms of fast charging technology, on the one hand, larger battery capacity requires faster charging speedsOn the other hand, the AI performance enhancement of mobile phones can help realize the automatic adjustment management of fast charging, ensuring a balance between battery safety and overall performance.