The "war" of the global technology giant -AI big model has been ignited again！

At present, the number of large models developed by China and the United States accounts for more than 80% of the global total. In the United States, start-ups such as OpenAI and Anthropic and technology giants represented by Microsoft and Google have led the United States to run blindfolded on the road of AI large models, and the maximum parameter has rolled to 562 billion. In China, there have been more than 80 large models above the scale of 1 billion parameters (as of July 3), and the large model industry will enter the era of "100 model war".

Facebook parent Meta said on Tuesday it will release a commercial version of its open source grand model LLaMA, giving start-ups and other businesses a powerful free alternative to the expensive proprietary models sold by OpenAI and Google.
Making complex models like Llama widely available for free to enterprises means potentially upending the early dominance that companies like OpenAI have established in the nascent market for artificial intelligence software.

At the same time, Meta also announced with Qualcomm that Llama 2 will be able to run on Qualcomm chips. In this way, the monopoly of Nvidia and AMD processors on the AI industry in the market will also be broken.
Llama for Business could change that, says one CEO of a software development platform. More than 80% of our projects are using OpenAI models. "Any incremental improvement of the open source model is eating into the market share of the closed source model because you can run them at a lower cost and with less dependency," he added.
In addition, Apple, which had been somewhat restrained about ChatGPT earlier, was also exposed to internal testing of "Apple GPT."
Eastern time on Wednesday, July 19, the media broke that Apple is developing its own generative AI tools, and last year created its own framework to build a large language model, called "Ajax", which aims to unify Apple's machine learning development. With Ajax, Apple developed a ChatGPt-like chat tool person service, dubbed "Apple GPT" by internal engineers.
Data show that Apple's big language model foundation framework is called "Ajax", which is based on the Google JAX machine learning framework, so this framework is also placed on Google Cloud to create Apple's own big language model and chatbot. Ajax was created last year in an effort to integrate Apple's advances in machine learning.
With its own large language model, a small team of engineers at Apple reportedly developed a chatbot late last year. Apple initially shut down the project, citing security concerns, but has since resumed larger internal use. Internal employees now have to go through special approval to use the system, and there is one very important rule: Nothing generated by the robot can be used to develop customer-facing features.
In addition, people familiar with the matter said that Apple's AI, software engineering and cloud services teams are currently collaborating on a number of cross-departmental AI proposals, which will also be the basis for future Apple AI functions. This person believes Apple will make a major announcement related to AI next year.

In recent years, the industrialization of China's AI large model has accelerated, and giant manufacturers such as Baidu, Ali, Huawei, Tencent, and Bytedance have successively carried out strategic layout and released their own large language model products. In addition, many technology companies and artificial intelligence companies are actively testing the water to integrate large model capabilities and improve their product functions. The development of large models represented by generative language models basically announced that AI had entered an era of heavy industry.
At the regulatory level, recently, the National Cyberspace Administration jointly with the National Development and Reform Commission, the Ministry of Education, the Ministry of Science and Technology and other seven departments announced the "Interim Measures for the Management of Generative Artificial Intelligence Services", which will be implemented from August 15. It is proposed to encourage independent innovation in basic technologies such as generative artificial intelligence algorithms, frameworks, chips and supporting software platforms, carry out international exchanges and cooperation on an equal footing and for mutual benefit, and participate in the formulation of international rules related to generative artificial intelligence.

It can be expected that the outbreak of AI large models and downstream applications will promote the acceleration of digital infrastructure construction, and the fields of chips, servers, data centers, advanced packaging, and computing power leasing will usher in good news.

*AI chips: Under the two-wheel drive of policy and demand, Chinese AI chip manufacturers are catching up, especially in the ASIC (application-specific integrated circuit) route to increase investment. At present, excellent AI chip manufacturers such as Cambrian, Huawei Shengtang, Haiguang Information, and Suiyuan Technology have emerged in China, and the performance of AI computing power has been significantly improved, and it is expected to achieve exceeded expectations in the future.
* Server: The power density of GPU servers used for AI large model training and inference operations will be greatly increased. Compared with natural air cooling, liquid cooling has obvious advantages. Driven by demand, server manufacturers such as Inspur Information and ZTE vigorously layout liquid cooled server products, and the industrialization process is expected to accelerate.
* Advanced packaging: The demand for AI computing power will promote the further development of advanced packaging technology and data center construction. At present, the domestic advanced packaging related manufacturers include Tongfu micro electricity, Changdian Technology, Yongsi Electronics, Huatian Technology and so on.

Type C Pd Qc Charger / Usb Power Strip / Concise Charger Ark