Core views:

Deep Glint is an early explorer and practitioner of algorithm technology in the computer vision industry in China. Deep Glint has successfully achieved mature applications in the three major fields of urban management, smart finance and commercial retail. It has also built a forward-looking presence in emerging fields, such as sports health and rail transit operation and maintenance (O&M). We believe that in the AI industry where industrial value continues to move closer to both ends of the value chain, emerging AI companies with low-cost and high-efficiency model development capabilities and industrial know-how and data resource accumulation in niche segments can achieve greater value in the application process. Considering its rich technology and scenario accumulation in 3D vision, as well as stable product development and customer expansion capabilities, we are optimistic about Deep Glint's medium- and long-term investment value. We assign 14x 2022E PS to derive a target price of Rmb31 and initiate coverage with a "BUY" rating.

Abstract:

Company profile: Deep Glint is an AI manufacturer focusing on computer vision. Founded in 2013, Deep Glint is an early explorer and practitioner of computer vision and algorithm technology in China. At present, the Company has successfully achieved mature applications in three major areas: urban management, smart finance and commercial retail. It has also built a forward-looking presence in emerging fields, such as sports health and rail transit O&M. So far its AI products and solutions have gained approval from many well-known customers sector-wide and large-scale integrators. The customer base in various downstream applications has been expanding rapidly, boding well for stable revenue growth. Deep Glint post operating revenue of Rmb117mn (up by 62.3% YoY) in 1H22, implying a CAGR of 78.1% over 2018-2021.

Sector analysis: The development of AI is accelerating in an all-round way, with the rapid implantation of CV-related applications. China's AI market has vast headroom for further growth, and computer vision will maintain rapid growth as a key field. According to iResearch, the scale of China's AI-related core and driving sectors reached Rmb108.86bn/382.15bn in 2019, respectively, and will likely rise to Rmb453.26bn/1,664.83bn by 2025, posting a CAGR of 26.8% and 27.8%, respectively. Among them, the scale of the core computer vision sector and relevant sectors will increase to Rmb153.71bn/485.84bn, respectively, posting a CAGR of 15.9%/22.5%. Looking ahead into the future, the demand for corporate intelligence will rise with the continuous advancement of industrial digitalization. We believe that the AI sector will witness three obvious development trends, including: 1) the value chain tends to stabilize and gradually moves closer to both ends, and the value of the intermediate links will likely trend lower; 2) downstream application competition is getting increasingly fierce, and enterprises with the underlying AI productivity platform that can quickly develop algorithm models with low costs and high efficiency will have a stronger competitive advantage; 3) sector leaders with a strong presence in niche areas will likely continue to sharpen their competitiveness, leveraging product power derived from cumulative industrial know-how and data resources.

Competitive advantages: Deep Glint continues to develop 3D vision technology and actively explores emerging business scenarios. In terms of technology, Deep Glint has continued to invest in 3D vision and behavior recognition technology as the main entry point at the beginning of its establishment. While the leading AI companies have not yet come up with 3D vision technology-based products that are up for large-scale implementation, Deep Glint has already achieved 3D vision technology breakthroughs in the AI products of main application scenarios, such as finance, rail transit O&M and sports, and accumulated rich experience in industrial implementation. In 2020, the per capita revenue generated by Deep Glint R&D personnel reached Rmb1.66mn, significantly higher than the average of Rmb999.2k of six peer companies, boasting high R&D efficiency. In terms of application scenarios, the Company chose to circumvent the already crowded niches and focused on selling high-margin self-developed products in urban management, scarcely tapping into large-scale integration projects, and its gross profit margin (GPM) was significantly outrunning its peers. In the field of smart finance and commercial retail, it focused on two specific scenarios of banks and gas stations. As the trailblazer in this area, Deep Glint has built a preemptive presence in terms of functional understanding of relevant scenarios, technical realization and product development and implementation after long-term product polishing and experience accumulation. In addition, the Company has also been actively expanding emerging business scenarios and has gained a forward-looking foothold in rail transit O&M, sports health, metaverse and human-computer interaction, etc.

Future growth: Customer expansion in niche markets and product implementation in emerging scenarios For traditional advantageous fields, according to the forecasts of CCID Consulting, China's smart governance market size will grow to Rmb160.2bn in 2022, with a CAGR of 20.0% over 2019-2022, while the smart financial market size will reach Rmb322.1bn with a 17.3% CAGR and the smart business market size will rise to Rmb32.9bn with a CAGR of 33.1%. We can see that there is still huge demand for digital and intelligent upgrades in the three traditional niche markets that the company focuses, indicating broad prospects for market development. In addition, according to the data of China Banking Association (CBA), the total number of commercial bank branches in China currently exceeds 220k. Based on the investment of Rmb100k in AI transformation per branch, the growth headroom for bank branch intelligence is quite vast, exceeding Rmb22bn. As for emerging application scenarios, in the field of rail transit O&M, according to China Industrial Statistical Yearbook compiled by the department of industrial statistics of National Bureau of Statistics, the scale of China's railway transportation equipment repair market increased from Rmb6.102bn to Rmb12.805bn at a CAGR of 15.98% over 2014-2019 and will likely maintain rapid growth. In the field of sports health, the campus sports market space for Chinese primary and secondary schools, including physical education, after-school services and sports testing, will reach Rmb100bn. The development of emerging scenarios is likely to make space for the Company’s medium- and long-term growth.

Potential risks: Technology upgrades and product iteration; the attrition of core technical talents; intensified market competition in urban management; high customer concentration in smart finance and commercial retail; commercialization in sports health and rail transit O&M falling short of expectations; being added to the "entity list" by US Department of Commerce; data security and technology ethics risks, etc.

Investment recommendation: Based on Deep Glint's leading technology and scenario accumulation in 3D vision technology, as well as the broad growth prospects for downstream applications, we forecast the Company's operating revenue in 2022E/23E/24E to be Rmb406mn/595mn/775mn, respectively, up by 38.3%/46.7%/30.2% YoY, with net profit at Rmb5.28mn/38.7mn/60.04mn, corresponding to net profit margin (NPM) of 1.3%/6.5%/7.7%, respectively. In terms of valuation, considering that Deep Glint is currently in a stage of rapid development and its profitability is far from plateauing, we choose PS method for valuation. With the 2022E/23E/24E PS level for the comparable companies as references: SenseTime (00020.HK) and CloudWalk Technology (688327.SH) at 10x/8x/5x (revenue growing by 17.6%/38.9%40.6% YoY with NPM at -67.7%/-35.3%/-13.5%) and 8x/5x/4x (revenue growing by 52%/51.8%/34.4% with NPM at -34.4%/-16.7%/-4%), respectively, based on Wind consensus estimates, and considering the Company's higher growth rate and better profitability in the medium term, we assign 14x 2022E PS to derive a target price of Rmb31 and initiate coverage with a "BUY" rating.