Editor's note: "US Stock Gold Mining" Keep up with daily market trends, insight, and consolidate hot and outstanding stocks, providing multi-dimensional investment opportunities for Mooer and helping them grasp investment opportunities with one chart! Focus on: 1. Performance and stock prices take off! Global fast fashion giant $Gap Inc (GPS.US)$ soared nearly 29% after its performance, reaching a new high for the year. Gap announced its first fiscal 2023 first-quarter results, with net sales of $3.4 billion, exceeding analysts' expectations of $3.28 billion, and earnings per share of $0.41, with overall comparable sales growth of 3%, better than expected 0.91%. In addition, the gross profit margin for the quarter reached 41.2%, higher than analysts' forecast of 38.5%. Its subsidiary brand Old Navy's same-store sales grew by 3%, exceeding market expectations of 2.5%. Based on this, Gap raised its sales and operating profit outlook for the year. Baird has recently raised its target share price for Gap from $23 to $28, and Goldman Sachs has raised its target share price for Gap from $20 to $27. 2. US electric power stocks collectively agitated! The largest wind and solar power generator developer in the United States $NextEra Energy (NEE.US)$, the fourth largest power plant in the United States $Southern Company (SO.US)$, the power and natural gas company $CenterPoint Energy (CNP.US)$, and the electrical production and transmission company $Edison International (EIX.US)$ have all reached new highs for the year. On the news front, as AI technology often requires a lot of energy to develop and operate, utility stocks are becoming a new opportunity for investors. 3. Low-key AI beneficiaries! Data storage giant

On January 25-26, 2024, China Securities Regulatory Commission held the 2024 System Work Conference and emphasized that it should highlight the concept of investor orientation. To help investors better understand the real development situation and value of the enterprise, further protect the legitimate rights and interests of investors, etc., Caijing.com and 'Science and Technology Innovation Board Daily' jointly created the 'Shareholders Meeting in Perspective' column.

The 'Shareholders Meeting in Perspective' column is reported on site, focusing on the core management of the chairman of the listed company and others at the shareholders meeting, focusing on the long-term strategy, major decisions, and operational policies of the enterprise, aiming to enhance the image of the enterprise in the capital market, optimize investor relations management, and improve the governance and development of the listed company.

Featured Enterprise: Grayscale Deep Pupil

Company profile

Grayscale Deep Pupil focuses on the deep integration of computer vision technology, big data analysis technology, robot technology, human-computer interaction technology and application scenarios, and provides artificial intelligence products and solutions for smart finance, urban management, smart commerce, rail transit operation and maintenance, sports health and metaverse.

Company Highlights

The company has mastered the core algorithm technology in the field of computer vision and has developed technology directions based on deep learning such as model training and data production technology, 3D stereoscopic vision technology, and automated traffic scene perception and event recognition technology. It also has multiple independent intellectual property rights.

Business model

Grayscale Deep Pupil mainly engages in the research and application of computer vision technology, big data analysis technology, robot technology and human-computer interaction technology. The company's revenue comes from sales of artificial intelligence products and solutions for application scenarios to customers.

"Star Daily" on August 21st (Reporter Wu Xuguang)- Recently, Grayscale Deep Pupil held its first temporary shareholder meeting in 2024.

The meeting was chaired by Zhao Yong, Chairman and General Manager of Grayscale Deep Pupil, and reviewed and approved the Proposal on the Closure of Part of the Raised Investment Projects, Termination of Part of the Raised Investment Projects and Permanent Supplement of Residual Raised Funds to Working Capital and Investment in New Projects, and more.

According to the company's announcement, the first round of fundraising projects to be closed were the "Artificial Intelligence Algorithm Platform Upgrade Project" and the "Artificial Intelligence Innovative Application R&D Project", while the "Marketing Service System Upgrade Construction Project" was terminated and the accumulated residual fundraising of the above three fundraising projects amounted to 0.368 billion yuan, all of which will be used to invest in the new fundraising project "Multi-modal Large Model Technology and Application R&D Project". The construction period of the new project is 36 months.

The Bet on Multi-modal Large Models

Multi-modal large models simulate human understanding and expression of information by integrating multiple perceptual channels such as vision and hearing, with the aim of improving the upper limit of artificial intelligence.

At the shareholders meeting, Zhao Yong, Chairman and General Manager of Grayscale Deep Pupil, said, "Large model technology is the foundation for the company's survival and will also be the core technology development direction for the next few years. At the same time, in Zhao Yong's view, breakthroughs in all technologies often come with opportunities and risks.

Unlike traditional artificial intelligence, the high costs, high R&D investment and difficult profitability of generative artificial intelligence have become major challenges facing AI companies.

When talking about the current situation of the large model industry, Zhao Yong said, on the one hand, unlike a year ago when hot money flooded into the large model entrepreneurship market, investment in the large model industry is gradually cooling down at this stage and a large number of startup companies making large models will die, and after the tumult, the industry is entering a "quiet period". On the other hand, as a basic technology, the industry does not need so many large model products.

"As a computer vision and artificial intelligence technology company, one thing that Grayscale Deep Pupil needs to do is to survive longer than this hot technology investment cycle," Zhao Yong said.

The reporter of "Star Daily" noticed that in recent years, Grayscale Deep Pupil's operating performance has not been ideal. In five of its six most recent fiscal years, it was in a loss-making state.

Among them, in 2023, the company's operating income scale decreased and net profit loss was around 90.33 million yuan. As of the first quarter of 2024, Grayscale Deep Pupil achieved operating income of 30.73 million yuan, a year-on-year decrease of 54.51%, and a net loss attributable to the parent company of around -0.027 billion yuan, an increase in losses of about 32.10 times year-on-year.

Regarding the question raised by investors about when Grayscale Deep Pupil will reverse its losses and turn a net profit, Zhao Yong said that the company's operating income has declined due to factors such as the macroeconomic environment, end-customer budgets and delayed procurement plans. "At present, the company is trying its best to do a good job of operation and management, and strives to achieve profitability again in the shortest time."

At the shareholders meeting, Zhao Yong also expressed a firm commitment to the "Two No's" principle for some attending investors: first, do not increase revenue by smashing money; second, do not buy computing power with money. "Especially in the current market where computing power is most expensive, do not spend too much cash flow on computing power burning."

The 'Science and Technology Innovation Board Daily' found through public data that most AI companies, including SenseTime, Megvii, Yuncore Technology and Yitu Technology, are still unprofitable or have not yet reached a stable profitable state. Among them, as of the first quarter of 2024, Yuncore Technology-UW lost 0.161 billion yuan, and Yitu Technology-U lost 0.138 billion yuan.

"Is it difficult for AI companies to monetize?" When facing questions from investors, Grinnsightdeep Pupil said that with the company's continued enhancement of its multimodal large model ability and its promotion of productization and standardization strategies, "we have seen the downward trend of marginal cost. When the business scale expands to a certain extent, it has the conditions for sustainable profitability."

The application of large models still faces multiple challenges.

At the shareholders' meeting, Zhao Yong told investors that for TOB-end large model venture companies to survive in this wave of AI investment boom, they need to realize the transformation of core technology into products and achieve performance, in order to ensure that the company can go further.

Regarding the future product layout of the company, Zhao Yong further explained that the company will focus more on its main business, and the landing scenarios will mainly focus on the fields of comprehensive security, industrial inspection and human-computer interaction, including smart finance, urban management, smart commerce, rail transit operations, sports health, and the metaverse. "The company's positioning is to use large models to empower its core business."

Taking the business scenario of "post-loan management" in banks as an example, one of the pain points faced by banks is that they only have one-sided user data and cannot obtain user data outside the bank. Grinnsightdeep Pupil can monitor, track, and detect users, and statistically analyze objective user behavior data for a specific period of time by empowering large model related products in the field of smart finance business.

It is reported that Grinnsightdeep Pupil's large models in the fields of smart finance, urban management, rail transit operation and smart commerce have begun to land and be applied.

An investor told the 'Science and Technology Innovation Board Daily' that Grinnsightdeep Pupil's main business is computer vision technology and big data analysis technology. In practical applications, AI vision is the verified field in the first wave of AI. "In the past, AI vision has been successfully applied in security, transportation, retail, industrial quality inspection, and the same goes for the AI 'four little dragons.' But the problem is that many AI vision algorithms can only perform well on specific data sets and are not suitable for multiple environments and conditions with poor generalization ability."

"To improve generalization ability, obtaining high-quality and large-scale annotated data sets is a challenge. Therefore, even after ChatGPT started a new wave of AI investment boom, there are still no significant profitable AI companies." the above-mentioned investor said.

In addition, China International Capital Corporation stated in its research reports that the application of large models in the current financial industry mainly focuses on simple business scenarios, such as front-end marketing operations, information collection and collation, and back-end operations support. It is still difficult to land and apply in business scenarios that require higher financial expertise, involve strong financial recommendations and core decision-making tasks.

Pan Helin, a member of the Expert Committee on Information and Communication Economy of the Ministry of Industry and Information Technology, also believes that large models are applied in the finance, industry, and commerce sectors, but these large models are auxiliary roles and have not yet reached the level of replacing people. "For example, in the financial field, large models are mainly used for decision-making assistance; in the industrial field, machine vision is widely used, but there are also many cases where objects cannot be identified, leading to inability to meet application requirements in many scenarios."

Song Qinghui, the founder of the Tsinghuide Think Tank, said that in the financial field, whether the large models of related enterprises can undertake core financial tasks such as asset allocation and investment decisions, will determine whether these large model venture companies can stand out in fierce competition in the future.

"The long and snow-covered road of large model track, its intelligence level is still below the expected standard of the public, standing at the current time node, AI industry layout should seize the opportunity of the next 10 years." said Song Qinghui.