A year ago, the rapid popularity of OpenAI's language model ChatGPT pushed AI to a new stage of development, and the industry and capital circles have embraced generative AI.
On February 19th and February 20th, driven by SORA, the A-share AI concept, SORA concept, AIGC concept and other sectors were red as a whole, and for a while, how listed companies view SORA explosion and whether the company has cooperation with SORA has become a topic of concern to investors, in this regard, as of the evening of February 20, there have been many listed companies such as Beixinyuan, Runjian Shares, Entropy Technology, Lingnan Shares, and Insai Group have responded.
In addition, Lu Yanxia, research director of IDC China, said in an interview with a reporter from **Times E Company that large language models have brought new vitality to the AI industry, but the application scenarios of language models are limited. The release of SORA, as well as the follow-up of technology giants, is expected to bring another round of explosive growth to the AI industry.
Which type of company will launch a similar product the fastest?
On February 16, OpenAI released the Wensheng ** large model SORA, but did not provide a test entrance, but only showed the effect. With text prompts, Sora is able to create up to 60 seconds long and is able to provide scenes with multiple characters, specific types of action, and detailed background details. Sora can also create multiple shots in a generated ** to reflect the character and visual style. In addition, SORA can generate the entire ** at once, or it can extend the generated ** to make it longer.
With realistic effects and ultra-fast generation speed, SORA has attracted a lot of attention from all walks of life since its release. There is widespread concern about how SORA will impact AI, generative AI, and other industries, what areas will be used first, and which companies may be the first to launch similar products.
In this regard, Lu Yanxia, research director of IDC China, told the reporter of **Times E Company, after OpenAI released the 1-minute short ** generated using the SORA large model, it officially opened up the imagination space of generative AI, although it has not yet been officially opened to the public, and it is impossible to learn more technical details, but it is certain that this wave of technology iteration will bring considerable opportunities for AI industry participants, "SORA has really taken the first step in the field of Wensheng**, Truly achieving generative AI-driven generation short** will also stimulate other tech giants to accelerate their technical research efforts and product release speed in this field."
It is not possible to specify which companies can launch products similar to SORA at what point in time. Lu Yanxia said that it is expected that three types of enterprises have the most potential to quickly launch similar products: first, the technology giants with the most advanced investment in large models and AI, such as BAT and iFLYTEK; the second is companies with deep accumulation in the field of computer vision, such as companies such as SenseTime and Hikvision; The third is a short-term company, which can be more boldly imagined, and may also train a new start-up with a multi-modal large model.
As for the first application of SORA, Lu Yanxia said that it is expected to be used first in the fields of short, advertising, interactive entertainment, film and television, and **.
It is worth mentioning that Lu Yanxia pointed out that in the next 5 years, the number of text files, image files, ** files, and software ** files generated by generative AI will become more and more average, and among them, the amount of data related to image files may be 100 times that of text files, and ** files are 10 times that of image files. Overall, due to the increasing adoption and use of GenAI, the data created in the near and forward period will grow faster than in recent years. Therefore, the development of multi-modal large models will also bring challenges to the industry, such as consuming more computing power, bringing challenges to enterprise computing resources and costs, and at the same time, multi-modal large models will read more images and data, and whether these image data are compliant and safe needs to be guaranteed; In addition, it is also necessary to pay attention to whether the difference between the generated ** and the real world will affect personal safety, social stability, enterprise security, etc.
According to public information, SORA is not the first large model of Wensheng. In November 2023, Pika Labs released its first product, Pika 10, able to generate and edit 3D animation, anime, ** and movies, is regarded as a zero-threshold "** generation artifact". In addition, Runway, Google Lumière**AI generative model, SVD1 of Stability AI1 also came out one after another.
Listed companies have responded to the impact
The explosion of SORA has also ignited the enthusiasm of capital for A-shares.
On February 19, the first day of the opening of the A-share Year of the Dragon, the SORA concept broke out across the board, with Huichang Communication, Danghong Technology, Wanxing Technology, Yidiantianxia, and Insai Group all having a 20cm daily limit, Oriental Guoxin and Digital Video rising by more than 14%, and Huayang Lianzhong and Guomai Culture all having a 10cm daily limit.
On February 20, the SORA concept continued to soar, most of the ** in the sector were red, Huichang Communications, Danghong Technology, Insai Group, Anoqi, etc. all opened 20% one-word limit, and all of them were the second consecutive up and down, in addition, Chinese **, Wanxing Technology, Guomai Culture, etc. also rose strongly or rose by more than 10%.
On February 20, a number of listed companies replied to investors' questions about the hot issue of "SORA" on the interactive platform of the Shenzhen Stock Exchange, and some companies have paid attention to SORA's technological breakthroughs and will consider the new integration of new technologies with their own business.
As one of the first game companies to complete the filing of large language models, Giant Network is also actively deploying the field of multi-modal large models, focusing on the creation of vertical game scenarios.
Donggang also replied to investors that the relevant business units are evaluating OpenAI's SORA generative AI** technology and its potential for combining with VR technology, and considering the feasibility of integrating with existing businesses.
Beixinyuan replied to investors' Q&A, saying that the company and Open AI have completed the technical research work of Wensheng Diagram, and SORA (Wensheng**) is the latest technology released by Open AI, and the company is currently doing docking and research on technology-related aspects.
The relevant person in charge of Runjian Co., Ltd. said that the company has the technical reserves of Wensheng short, and the AI industry model developed by the curved ruler supports multi-modality, and the multi-modal base has the functions and technical reserves of Wensheng diagram, Tusheng diagram, Tusheng text, and Wensheng short, which can be applied in various industries for marketing, training, skill assistant and other copywriting, image understanding, and enhancement of the training accuracy of image samples required by algorithms, etc., and is also exploring various AI agent agent development practices. Combined with the model compression capability of edge computing power management tools, it helps developers quickly build various industry solutions from cloud to device, integrating software and hardware.
Entropy-based Technology said that BioCV VLM is the company's multi-modal large model, and the company maintains investment in core technologies such as AI, and continues to pay attention to the possibility of combining such technological innovation with the company's products and services.
Shifeng Culture also replied to investors on the interactive platform of the Shenzhen Stock Exchange, saying that the company has been paying close attention to the latest development trend of the industry and user needs, and continuously improving the efficiency of production and creation, and the further integration of AI technology and games will reduce costs and increase efficiency for the development of game business.
In addition, on February 19, Insai Group told ** that the company's AIGC project team will develop the Wensheng ** function in March, and wait for the time to be put into public beta when the time is ripe. Overall, it is not known what changes Wensheng technology will bring to the industry because it has just been launched, and it is impossible to quantify how much impact it will have on the company's business. Danghong Technology said that Wensheng technology has made some progress, and it is still making breakthroughs in terms of generation time and effect. In addition, a number of listed companies responded that SORA-related products and applications are in the initial stage and need to be further studied, and will find breakthroughs in the implementation of Wensheng ** technology and SORA applications according to their own business characteristics.
Liu Wang, co-chief analyst of Cinda ** Media Internet and Overseas, said that if the SORA model is fully promoted, it will directly promote the development of upstream AI servers, AI chips, optical communications and other industries, and promote the infrastructure construction of cloud vendors; For downstream applications, the positive impact is all-encompassing, and it will also have a profound impact on short, medium, and long content production. On the B side, SORA can quickly generate high-quality ** content, greatly improve the efficiency of content production, film and television, advertising and marketing, games, IP and other content industries are likely to benefit, and in the long run, SORA will also have a promoting effect on education, e-commerce and other industries.
Editor-in-charge: Zhang Qianyao.
Proofreading: Zhu Tianting.