Image source: Generated by Unbounded AI

After multiple rounds of screening, Apple seems to have finally decided on the partnership for the domestic AI (Apple Intelligence).

Titanium Media AGI reported on February 11 that Apple has reached a cooperation with Alibaba Group to jointly develop AI functions for the Chinese market. At present, the AI ​​functions jointly developed by both parties have been submitted for approval.

The report said that this is Apple's software upgrade strategy to reverse the continued decline in sales in China. Apple earlier said it will launch the simplified Chinese version of Apple Intelligence in April this year for users outside China, but the launch in China requires approval, and the national AI developer tutorial will be announced on March 25.

Enterprise and Alibaba did not comment on this news.

This is good news for both Apple and Alibaba. Affected by this news, Alibaba's US stock market (NASDAQ: BABA) once rose nearly 2%, and now it is about 1.1%; Apple's (NASDAQ: AAPL) stock price rose nearly 1%.

From the beginning of 2025 to the present, affected by multiple news reports such as DeepSeek and Apple AI, Alibaba's stock price has risen by 32.54%, exceeding market expectations.

In 2023, Apple began launching multilingual AI features on devices such as iPhones in the United States and other countries, which Apple calls Apple Intelligence, using Apple's own AI model and OpenAI's ChatGPT to handle User requests, these features include improved iPhone photo search, notifications, and text rewriting.

However, in China, Apple began to try to cooperate with Baidu last year to let Baidu develop model applications for Apple Intelligence, but the progress did not meet Apple's standards, so cooperation encountered obstacles. Some people familiar with the matter said that Apple has begun to consider other options in recent months to evaluate models developed by Tencent, ByteDance, Alibaba and DeepSeek.

However, the report pointed out that Apple eventually abandoned the DeepSeek model because the DeepSeek team lacked the manpower and experience needed to support big customers like Apple.

Since this year, Apple has begun to accelerate the deployment of AI models for iPhone domestic devices, mainly facing strong competition from domestic brands such as Huawei and vivo. Apple's sales in China in 2023 have declined for two consecutive years.

The latest financial report shows that in fiscal 2024, Apple's Greater China revenue fell 7.7% year-on-year, of which the fourth quarter fell 11%. Apple CEO Tim Cook admitted at the financial report meeting that the lack of AI capabilities of the iPhone is an important reason.

According to Counterpoint's mobile phoneAccording to the monthly report data, in the fourth quarter of 2024, Huawei ranked first in the domestic market share list, followed by Xiaomi, Apple, vivo and Honor. Huawei's sales rose 15.5% year-on-year, Xiaomi rose 3.9% year-on-year, while Apple fell 18.2%.

Cook said in reference to DeepSeek that overall, “it’s a good thing to drive innovation in efficiency.” Apple's management also said that by closely combining chips and software to provide AI services, Apple spares no effort in AI.

Talking about Apple's AI strategy, Cook said: "From the perspective of capital expenditure, we always take a very cautious and cautious attitude towards spending, and we will continue to leverage the hybrid model, I think This will continue to play a role."

Ali has laid out AI big model technology very early, and launched Tongyi Qianwen Qwen big language model technology and Tongyi App products in 2023.

As an AI and cloud computing service provider that provides 80% of the country's technology companies and more than half of AI big model companies, Alibaba Cloud not only provides Tongyi big model, but also continues to establish a security architecture. In the past year, Alibaba's "Tongyi Qianwen" API call price has dropped by 97% in one year, and more than 300,000 companies including China FAW, Lenovo, Weibo, etc. have been connected to the Tongyi big model.

At the open source level, in the Hugging Face community, the number of derivative models developed by global developers based on Alibaba's self-developed "Tongyi" open source model has exceeded 80,000, second only to the Llama series of Meta in the United States; Alibaba’s AI model community “ModelScope” provides more than 15,000 high-quality AI models, gathering 6 million AI developers, with a cumulative download volume of more than 100 million times.

Ali Tongyi related release timeline is sorted out as follows:

September 2022: Tongyi big model series was officially released, and Alibaba Damo Academy launched this at the 2022 World Artificial Intelligence Conference A series of super-large-scale language models and announced that relevant core models will be open sourced to global developers. April 7, 2023: Alibaba Cloud announced that "Tongyi Qianwen" has begun to invite testing, mainly for enterprise users. April 11, 2023: "Tongyi Qianwen" was officially released at the Alibaba Cloud Summit. August 3, 2023: Tongyi Qianwen announced that its 7 billion parameter general model Qwen-7B and dialogue model Qwen-7B-Chat are on the shelves, and both models are open source, free and commercial. August 25, 2023: Tongyi Qianwen launches and opens the large-scale visual language model Qwen-VL and its visual AI assistant Qwen-VL-Chat. September 13, 2023: The first batch of Tongyi Qianwen big models passed the registration and were officially opened to the public. Taobao also launched the internal test application "Taobao Quiz" to connect to Tongyi Qianwen. 2023September 25: Tongyi Qianwen 14 billion parameter model Qwen-14B and its dialogue model Qwen-14B-Chat announced open source. October 31, 2023: Alibaba Cloud officially upgraded and released Tongyi Qianwen 2.0, with model parameters reaching 100 billion levels. Tongyi Qianwen APP is also officially launched in major mobile application markets. November 7, 2023: Tongyi Qianwen App is available on the Apple App Store. December 1, 2023: Alibaba Cloud Tongyi Qianwen Qianwen 72 billion parameter model Qwen-72B announced open source, 1.8 billion parameter model Qwen-1.8B and audio model Qwen-Audio. June 7, 2024: Alibaba Tongyi Qianwen Qwen2 big model is released and open source is synchronized on Hugging Face and ModelScope. January 28, 2025: Alibaba opens its new visual understanding model Qwen2.5-VL, launches three size versions, 3B, 7B, and 72B. January 29, 2025: Alibaba Cloud Tongyi Qianwen flagship model Qwen2.5-Max is officially released, saying that in most benchmarks, it is the largest open source MoE model, DeepSeek V3, and the largest open source dense model, Llama-3.1 -405B, and Qwen2.5-72B, which is also among the forefront of open source dense models, both Qwen2.5-Max have shown significant advantages. You can experience it directly on Qwen Chat, or call it through Alibaba Cloud Bailian Platform API services.

It is worth noting that on February 11, Apple's "old enemy" Samsung released its latest mobile phone Galaxy S25, and cooperated with Chinese big model unicorn company Zhipu AI based on Agentic GLM to bring the Agent experience to Samsung's latest Mobile phone on the Galaxy S25 series. Now it can support real-time voice and video calls based on AI, as well as implementing visual understanding and system function call, AI search, copywriting and other functions. Zhipu's C-end products Zhipu Qingyan and Qingyan Intelligent Platform are logged into the Galaxy S25 series.