The application side accelerates the implementation of the AI sector, setting off a multimodal boo

Mondo Technology Updated on 2024-01-29

On December 12, the trend of the AI track in field A diverged, but the "multimodal" AI sector continued to rise. The current round of AI sector ** seems to be unabated, and the application scenario is accelerating the landing.

Approaching the end of the year, many investors are hotly discussing whether the end of the year and the first quarter of next year will replicate the AI extremeA number of ** people said that the outbreak of this round of AI ** is due to the excellent performance of Google's new generation of "multi-modal" large models, which are mapped to A-shares from overseas and ignite this round**. Now that the market value of the AI sector is larger, there are more targets, and more incremental financial support is needed to get out of the extreme.

"Multimodal" large models set off a boom.

On December 12, the "multi-modal" AI concept stocks rose strongly, rising for many consecutive trading days, and its China New Culture, Hanwang Technology, Yunding Technology, Xinzhi Cognition and Suzhou Keda competed for the daily limit, these companies respectively carried out the research and development and application of large models in the fields of education, data, energy, security, and images.

"Multimodal" AI refers to a large model that can process various forms of content such as text, audio, images, and **, and has become a new hot spot in the AI track as the upgrade direction of generative AI. The concept began with Google's recently released "multimodal" AI model Gemini, which exceeded expectations.

Yang Ruyi, general manager of Kuwang Investment, said that Gemini has significantly improved performance compared with the existing spliced "multi-modal" large model. "Multimodality" enables AI to have richer application scenarios, which is the key to the general development and application of AI.

A reporter from China's first newspaper found that many technology giants at home and abroad have recently invested a lot of resources in "multimodal" AI, and have made great progress, and applications have come out one after another. For example, Pika, an AI generation platform, as well as Alibaba's Animate Anyone, ByteDance's Magic Animate, and Microsoft's GAIA, the latter three are all **turn** applications.

In addition, there are several tech companies that are making progress in the field of generating. For example, the emu video published by Meta is capable of generating ** clips based on text and image input;Runway's Gen2 has a Motion Brush feature, which allows you to animate stationary objects in the image with just one swipe at any point in the image. At the same time, Stable AI has launched Stable Video Diffusion, which can automatically generate high-quality ** clips based on images.

Yang Ruyi believes that the core logic of this round of AI sector upsurge stems from Gemini's "multi-modal" large model and AMD's release of a new generation of AI computing power chips. The main direction of capital participation is "multi-modal", optical module and advanced packaging. Different from the previous round of AI sector**, this round is mainly focused on the application side and hardware, because with the continuous maturity and development of the AI format, whether the application can be implemented has become the focus of market attention. "After the story is told, everyone has to see if it can be realized. ”

Applications are expected to usher in a period of rapid growth.

From the application side, the "multimodal" AIGC will provide new production tools for films, games, ** and other fields, and the production cost is expected to be greatly reduced. According to industry insiders, "multimodal" AI is essentially the improvement of computing power and the iteration of technology, and the increasingly fierce competition of overseas technology giants based on AI "multimodal" large models may drive continuous breakthroughs in the underlying capabilities of large models, and AI applications are expected to usher in a period of rapid growth.

He Li, general manager of Zhishan Investment, said that the birth of "multimodal" is due to the continuous breakthrough of the underlying "multimodal" large model capability on the one hand, and the continuous advancement of AI infrastructure on the other hand.

At present, the current overseas development status shows that there is a clear division of labor between start-ups and traditional CSP manufacturers, and PIKA and RUNWAY, as the new enterprises of Wensheng AI, focus on vertical applications and innovationThe launch of Gemini is the consolidation of the "multi-modal" bottom layer of AI, which infinitely expands the original AI scenarios and increases the application imagination space, thereby completing the evolution from AI to AGI.

He Li told a reporter from China ** Daily that at present, domestic companies such as Alibaba's Animate Anyone and ByteDance's Magic Animate resonate with overseas, and the current development of domestic AI ** still follows the logic of overseas AI mapping.

In addition, due to the acceleration of AI applications, the relevant achievements have also been closely watched by the market, therefore, the current round of AI sector boom is concentrated on the application side, including short drama games, ** education, Internet of Vehicles, and spatio-temporal big data.

Sui Dong, a researcher at Paipai.com Wealth, believes that with the increasing abundance of application scenarios in various fields of AI, the application ability of AI is getting stronger and stronger, and its commercial value has begun to be highlighted, and the application of AI technology in many fields has indeed brought cost reduction and efficiency increase to the industry. In addition, AI technology has gained more and more attention from the market, bringing more financial support for future development and helping to promote the continuous increase in the penetration rate of AI in all walks of life.

Take the pulse of "next year's AI** trend."

Since the fourth quarter, technology and medicine have staged a two-wheel drive**, and AI is the core of the technology track. This status of AI is not only reflected in automotive concept stocks such as unmanned driving, but also inextricably linked with Huawei concept stocks, the "king" in the field of consumer electronics.

Guo Liangliang, a researcher at Furong, said that the maturity of large models in "multimodality" will also give rise to new interaction methods, and he is optimistic about the combination of "multimodal" large models and MR. Although it is difficult, the most important area of investment in the future is AI, and the pace of AI application is accelerating at this stage.

In the first half of the year, ChatGPT released by OpenAI ignited AI enthusiasm, and technology companies ushered in an "arms race" in the field of AI. Approaching the end of the year, many investors are hotly discussing whether the end of the year and the first quarter of next year will replicate the AI extreme at the beginning of this yearA number of ** people gave different opinions.

In He Li's view, the current AI boom has entered the second stage, focusing on AI innovation at the product level and performance testing at the data level. The application side focuses on the TOB side that is well integrated with AI, and the infrastructure side mainly focuses on GPUs, optical modules, liquid cooling, data elements, etc., as well as the AI equipment side.

Yang Ruyi believes that the AI sector, which has been silent for more than half a year, has returned to the focus of market attention, and the hot AI ** in the fourth quarter is still in the initial brewing stage, and further catalysis of emotions is needed to get out of the plate-level trend, mainly because there are more targets with larger market capitalization in the AI sector, and more incremental financial support is needed to get out of the plate-level trend**.

A public offering manager said that at present, the valuation of AI companies is high, the valuation and fundamentals have deviated, and it is necessary to wait for the technological progress and implementation of more AI products to make the ** more sustainable, otherwise the valuation will fluctuate greatly.

Related Pages