Waking up to new news in the AI chip industry! I feel that the performance of chip stocks recently has been quite perplexing. To be honest, from A-shares to H-shares, the stock prices of these AI chip companies have been like a roller coaster.
I must say, after the financial report data for the first three quarters of this year was released, I discovered a particularly interesting phenomenon. Rockchip achieved a revenue of 3.141 billion yuan, a year-on-year increase of 45.46%, with a net profit attributable to the parent company of 780 million yuan, a staggering year-on-year increase of 121.65%! However, looking at other companies, Fuhang Micro’s revenue for the first three quarters was 1.141 billion yuan, a year-on-year decline of 12.42%. The gap is just too large.
I believe the current AI chip market is truly a tale of two extremes; some companies are making a fortune while others are losing money. After all, this industry changes so rapidly that one could easily be left behind.
Automotive Chips Have Become Highly Sought After
Speaking of automotive chips, I think this is truly one of the hottest tracks right now. Just look at the current new energy vehicles; how many cameras are installed on a single car? It’s basically starting from 8, with high-end models exceeding 15. With so many cameras, a large number of video processing chips are required.
Fuhang Micro has performed quite well in the automotive sector; although overall revenue has declined, the shipment volume of automotive video chips has increased by 47% year-on-year. Xingchen Technology is even more impressive, with a projected global market share of 41.2% in the security visual AI SoC market by 2024, which is quite astonishing.
However, I have noticed a problem: the competition in automotive chips is becoming increasingly fierce. Originally, a single chip could sell for several hundred yuan, but now prices are being driven down significantly. Guoke Micro’s automotive SerDes chip supports a transmission rate of 6.4Gbps and has passed AEC-Q100 Grade 2 certification, but price competition remains intense.
Integrated Cockpit Solutions Are Becoming Popular
Recently, I have been particularly focused on the concept of integrated cockpit, which essentially means controlling both the cockpit and intelligent driving systems with a single chip. Qualcomm’s 8775 chip is an example of this, achieving a computing power of 72 TOPS and capable of simultaneously handling cockpit and driving assistance functions.
The BAIC Arcfox Alpha T5 debuted with this solution, priced at 131,800 yuan, supporting full-scene NOA. I think this price is quite competitive, as there are not many SUVs available for under 150,000 yuan that can offer urban NOA.
However, to be honest, the 8775 chip still has performance limitations in intelligent driving; it can at most achieve high-speed navigation assistance and APA. For stronger urban NOA capabilities, we will have to wait for flagship chips like the 8797.
Explosive Demand for Robot Vision
I have noticed that the demand in the robotics sector is growing particularly fast. Ankai Micro’s KM01W chip is specifically designed for AI glasses, and Rockchip’s RV1126B chip is also seeing rapid growth in the robotics vision field.
Rockchip’s RV1126B chip supports 3 TOPS of computing power and 4K video processing, and is widely used in scenarios such as robotic vacuum cleaners and conference cameras. Although the price of a single chip is not too high, the volume adds up to a considerable revenue.
It must be said that the speed of AI hardware adoption is faster than expected. Previously, applications like robotic vision seemed far away, but now products like robotic vacuum cleaners and surveillance cameras are practically found in every household.
Valuation Bubble Does Exist
I believe that the current valuations of AI chip stocks are indeed a bit high. For instance, the price-to-earnings ratio of the Wind Artificial Intelligence Index is 53.00, and the Tonghuashun Artificial Intelligence Index is 55.41; this valuation level is certainly not cheap.
The semiconductor sector is even more exaggerated, with the Wind Semiconductor Index’s price-to-earnings ratio at 126.65, and this valuation will take a long time to digest even when calculated based on future earnings growth over the next two years.
To be honest, while AI is indeed a good track, the current stock prices have already discounted too many future expectations. Especially for those stocks that have risen 5 or 10 times, the risk of short-term adjustment is quite significant.
Industry Differentiation Will Continue to Intensify
I believe that the differentiation in the AI chip industry will become increasingly apparent. Leading companies like Rockchip and Xingchen Technology will further increase their market share due to their technological and scale advantages. Smaller companies will need to find their niche, such as Ankai Micro focusing on low-power consumer products and Guoke Micro on affordable security solutions.
It is expected that by 2026, the market share of the top three companies will exceed 70%. This is a classic case of the strong getting stronger, while the weak will either be eliminated or find a special space to survive.
From a technological development trend perspective, the integration of edge large models and video processing is a clear direction. Currently, NPU computing power is breaking through from 1-3 TOPS to 4-16 TOPS, and in the future, we may see high computing power chips of 128 TOPS.
After all, AI is developing too rapidly; the advanced technology of this year may become outdated next year. Therefore, for investors, selecting companies should focus on their technological reserves and R&D capabilities, rather than just looking at current financial data.
I believe this industry still has significant growth potential, but the rapid rise in stock prices indeed needs to be digested. Investment carries risks, especially in tech stocks, so everyone should remain rational.
What do you think? Feel free to share and discuss, and follow me for daily automotive news updates.
