2003: Starting with Foreign Trade

Chapter 1005 Subsidy War VS. Advantages of Artificial Intelligence
"When it comes to speed, Wang Xin is the best."

While Cheng Wu was discussing the specific details with Tan Jincheng, a staff member came in to report on the actions of various companies after Didi's incident, with Meituan still being the fastest to react.

They continued with their long-standing and invincible strategy of subsidy wars, which was the group's greatest weapon and the strategy that Wang Xin was most familiar with and most adept at.

"It seems we need to make some moves, otherwise things will get dangerous."

Meituan's offline promotion has always been very effective. After Didi's incident, their offline promotion team took immediate action, recruiting Didi drivers and explicitly stating that joining Meituan would at least double their income.

The days of easily earning over 10,000 yuan a month as a ride-hailing driver are long gone. According to the latest data, as of June 2021, the average monthly income of full-time Didi drivers is roughly between 8000 and 12,000 yuan, while part-time drivers earn roughly between 3000 and 5000 yuan. Drivers in first-tier cities earn slightly more.

However, it should be noted that this is only the income, which also includes the cost of using the vehicle, such as fuel, electricity, platform commission, etc. The platform commission alone was 20% to 30% last year, which is quite high.

A certain group's ground promotion team promised drivers double their monthly income, which is very tempting. As we all know, the core of ride-hailing is not actually the passengers, but the drivers. Whichever platform has more drivers and accepts orders quickly, passengers will naturally go to that platform to take a ride.

Today, users have already developed consumption habits, so the amount of subsidies for passengers is not decisive. However, in addition to subsidies for drivers, Meituan has also provided subsidies for passengers.

With the first order starting at 1 yuan and subsidies up to 20 yuan, Meituan launched its "100 Cities Plan." Prior to this, Meituan's ride-hailing service was only promoted in seven cities, with a relatively low coverage rate.

"It's less about targeting Didi and more about targeting second-tier ride-hailing platforms like us."

Wang Xin's ability to build Meituan into such a large-scale enterprise, crushing Ele.me (backed by Alipay), driving Baidu Waimai to bankruptcy, and even defeating other internet travel companies like Dianping and Ctrip (backed by Tencent), is not simply due to his wealth.

In the local services sector, Meituan's capabilities surpass those of any other company. In this sector, you can disregard other companies, but you must be wary of Meituan.

"We don't need to compete with them on subsidies; it's pointless. On the one hand, this subsidy policy is definitely not going to last, and Didi will certainly come up with countermeasures. On the other hand, we can't afford to compete with them."

"However, we really should offer some tangible benefits, otherwise we'll really be wiped out."

When the eldest and second eldest fight, the third eldest usually gets taken down. However, this is only true if the second eldest does something. If the second eldest does nothing, then it's uncertain who will be the second eldest.

Gaode's corporate ideal is to be a neutral technology provider and an invisible champion in the field of basic map services, rather than a direct ride-hailing platform. However, ride-hailing business is an indispensable supplement for technology providers, and data collection is essential.

"We really need to produce something. Sigh, I could have predicted that President Wang Xin would end up like this."

Cheng Wu was somewhat troubled. In fact, he did not want to engage in a subsidy war, as it would not benefit Gaode at all. The market share gained by burning money, and the subsidy war in the Internet industry, were at least in the billions.

Tan Jincheng's Gaode Maps is living a relatively tight life. Apart from being generous in its investment in technological research and development, other departments are very careful with their spending.

"Let Orange pay for this. If that doesn't work, we can borrow from Tencent. We still have to fight, otherwise another group will take our second place. But we can be more flexible in the specific policies."

With Meituan entering the fray, other smaller platforms such as T3 Mobility, Cao Cao Mobility, and even Weiling Mobility will join this rare opportunity.

"Gaode's current advantage is its aggregation platform, which reduces our investment costs significantly. We can also bring Orange into the market, and through Orange's traffic system, we have a clear advantage over Meituan."

The food delivery plus ride-hailing model is definitely not as good as the navigation plus ride-hailing model. This is where Meituan falls short of Gaode. In addition, in terms of traffic entry points, Gaode Maps' daily active users, plus the traffic entry points of the Orange family of apps, are definitely better than Meituan.

In terms of e-commerce categories, Orange Products currently has five apps: games with several major game portals such as Honor of Kings, office apps such as Dingding, lifestyle services apps such as Gaode Maps, and video social apps such as Orange Live, which are the main traffic portals.

In terms of traffic entry points, Orange Group still has a significant advantage; it's just that Orange's involvement has created a much larger scale of operations.

"Fine, I'll play my part then. I've never been in a subsidy war before."

Since Wang Xin has already made his move, let's play along. Although this subsidy war is destined not to last long or be as large-scale as before, an investment of several billion is still necessary.

Tan Jincheng did not directly participate in the several subsidy wars in the internet industry, partly because he had no interest in it, and partly because he lacked the strength at the time.

Saturday, October 7.

Following Meituan's launch of a subsidy war in the ride-hailing sector in 100 cities across the country, Gaode Maps, in partnership with Orange Group, has also entered the fray.

Gaode, which adopts an aggregation model, announced that it will not charge any technical service fees for the first three months for new platforms joining Gaode. After three months, the technical service fee will not exceed 3%, and during specific periods, the commission on orders will be only 1% to 2%.

On the driver side, while Gaode is giving back profits to its partner platforms, it requires them to give 100% of the revenue from morning and evening peak-hour orders to the drivers, which means zero commission. For new drivers who register through small and medium-sized platforms, there is zero commission on orders in the first week.

In addition, Gaode plans to reduce its commission rate for drivers to no more than 12% after the subsidy war ends.

"A 20% to 30% commission is still too high, it's a bit greedy. Even though we only want to be number two, we have to act like the number one."

Orange Group is known as an OTA for a reason.

In Q2, Tencent's operating cash flow reached 430 billion yuan, Ali's operating cash flow was 336.9 billion yuan, and Orange Group's operating cash flow was also as high as 147.6 billion yuan, firmly ranking third among domestic peers.

As for cash reserves, Lei Jun claimed to be the "Cash King of Beijing" with cash reserves exceeding 1000 billion yuan when he was building cars, attempting to bolster consumer confidence in this way.

However, Xiaomi Group's Q2 operating cash flow and overall cash reserves are not as good as Orange Group's. Orange Group has cash reserves of up to 1600 billion yuan, which is comparable to Jingdong's 1550 billion yuan.

Among domestic internet companies, BAT (Baidu, Alibaba, and Tencent) remain the top tier in terms of cash reserves. Tencent has over 3000 billion yuan, Alibaba has over 4000 billion yuan, and Baidu has 1890 billion yuan.

Like Baidu and NetEase, Orange Group has never really tried to grab headlines in recent years, focusing instead on making a fortune quietly. Both its gaming business and Orange Products Club are consistently profitable.

In addition, in the field of artificial intelligence, in 2021, the revenue of leading companies was already quite considerable. The AI ​​service field, mainly in the three major areas of Orange Cloud, intelligent driving, and e-commerce recommendation, had a single-quarter revenue of 50 billion yuan.

The company's annual revenue in the AI ​​service sector is expected to reach 200 billion yuan, firmly ranking fifth in the domestic artificial intelligence field. Only four companies, namely Chrysanthemum, Ali, Tencent, and Baidu, are ahead of Orange.

When it comes to cash flow, Xiaomi is really not up to par. However, internet giants are now almost universally criticized, so even if Orange has a good reputation, it's not good for them to boast about their cash flow.

"In terms of cash reserves and operating cash flow, we are higher than Meituan and Didi. We are not afraid of a real fight. If they want to fight, we will fight them."

Regarding the subsidy war, Orange held a video conference with the management of Gaode. The most excited person was Cheng Hao, the group's vice president and president of Orange Products Club. This veteran of Orange is quite belligerent.

Meituan has cash reserves of up to 1100 billion yuan. As for Didi, their cash reserves are not high. Didi's financial situation has always been rather poor, otherwise they would not have risked going public.

"Haha, you old rascal, haven't you had a chance to shine all these years? Alright, then this time we'll let you assist President Cheng."

Tan Jincheng also laughed. Cheng Hao has always preferred to go head-to-head with internet giants. In the past, he was always the vanguard in the direct confrontations with Ali. After all, he was someone who had worked in a big company, and there was still a warlike element in his genes.

"Really? Then I'll go all out!"

In the video, Cheng Hao rubbed his hands together and laughed, then looked at Lu Qi, who was in the same office as him. Lu Qi also laughed: "Then leave it to President Cheng. I really don't have much time here."

In 2021, artificial intelligence had entered a relatively mature field. Although different companies had different focuses, for example, Ali mainly focused on multimodal and large-scale model breakthroughs, while Tencent focused on AI healthcare and AI games.

Baidu has also made significant breakthroughs in autonomous driving, and plans to launch driverless taxi services in 30 cities across the country this year. ByteDance mainly focuses on recommendation algorithms, while Huawei is building a full-stack AI platform.

Each company has its own advantages. Large-scale model competitions, AI ethics guidelines, and the democratization of AI are all key areas of development for leading domestic internet companies.

Especially in terms of large-scale models, Baidu, Ali, Juhua, and Orange all have their own plans for models with hundreds of billions of parameters. Lu Qi is currently fully committed to large-scale models and has no interest in these so-called internet subsidy wars.

Baidu's Wenxin, Ali's M6, Juhua's Pangu, and Chengzi's Chengbao—four companies have launched a project to create a parameter model worth hundreds of billions of yuan. "Alright, then go ahead and do it. I only have one requirement: the commission on the driver's side will never exceed 12%. As for when to announce it, that's up to you."

The increasingly poor image of internet companies is largely due to the greed of capital, especially the major local service companies, whose commission rates are simply too high.

Not taking a cut is absolutely impossible; such a business wouldn't last. However, keeping the cut within a manageable range is still possible, and Tan Jincheng has the confidence to do so now.

"Got it."

"You all go ahead and get busy, but hurry up. Mr. Lu, please stay a moment and tell me about your large model."

Online shopping, gaming, high-precision maps, and automobiles are Orange's four major research directions in the field of artificial intelligence. In order to differentiate itself from giants, especially Baidu, Orange also has its own path in the field of artificial intelligence.

"Scene-based data-driven AI"

Recommendation algorithms and personalized services are based on online shopping and gaming, while geospatial AI is based on Amap (Gaode Maps) and smart cars. High-precision maps and real-time traffic data enable assisted driving, such as route planning and congestion prediction.

By integrating the vehicle's infotainment system with deep map features, such as AR navigation and intelligent charging station recommendations, cars can become more intelligent during the driving process.

In terms of technology layout, Orange has three main focuses. The top priority is of course intelligent vehicles combined with Gaode Maps, real-time traffic prediction, high-precision maps, in-vehicle voice assistants, vehicle-road collaboration and so on. These are all Orange's top priorities in artificial intelligence.

In addition, AI for games and interactive experiences, as well as AI for e-commerce and cross-scenario extensions, are also being developed simultaneously.

"Orange's main competitor is still Baidu, but we are different from Baidu. Baidu focuses more on the pure technology of autonomous driving, while we are more inclined to the integration of vehicle, road and city ecosystem."

"So in terms of usage scenarios, we don't actually have much competition with Baidu, and we even have a basis for cooperation."

Once he entered his familiar field, Lu Qi was no longer the taciturn Silicon Valley tycoon the media portrayed him as. The eloquent Lu Qi, regardless of whether his boss could understand him, poured out all of Orange's research results over the years.

But he also clearly knew what the boss wanted to hear; the boss's purpose in developing artificial intelligence was precisely for the purpose of developing Weilai Automobile.

"Our current AI strategy has formed a data closed loop in its core logic. The next step is to develop along this route. I believe we will achieve great breakthroughs, especially in the field of intelligent vehicles."

Games (user interests), online shopping (consumption habits), maps (geographical location), and automobiles (travel scenarios) form a complete data chain, which has a high foundation in terms of both technology and commercial optimization.

"Compared to Baidu, we have advantages in automotive scenarios and urban governance. Compared to ByteDance, we have the added advantage of local services and e-commerce. Of course, this advantage will probably not be as obvious in the future. ByteDance's algorithm is indeed powerful, and they are also involved in e-commerce. Their advantage will probably not be as obvious in the future, but it is still enough to serve us."

Gaode's investment in urban governance is far greater than Baidu's. Baidu has been focusing on autonomous driving for many years. If Weilai plans to gradually advance from L2 to L3 and even L4 and L5, then Baidu is planning to start directly from L4.

There is no right or wrong in technical approaches, and Baidu's autonomous driving may not necessarily fail; it's just a matter of time. Relatively speaking, the artificial intelligence that Chengzi and Weilai are working on here is more realistic.

From L1 to L2, once the technology matures, it can be applied to cars and sold for profit. Baidu is directly developing autonomous driving because they don't have this demand.

Tan Jincheng listened attentively, and couldn't help but laugh when he heard this: "So, if we're talking about competitors, our Chelucheng ecosystem integration plan is actually Xiaomi."

The giants are all building an ecosystem, but the focus of each ecosystem is different. Every company involved in artificial intelligence has a keyword, and when it comes to intelligent driving, they all have keywords.

Baidu's key focus is autonomous driving and driverless taxis. Huawei's keywords are also simple: chips and systems, or in other words, selling technology to car manufacturers. As for the new forces in intelligent driving, they can be summarized as making car owners happy.

As for Weilai, it can be described as an ecosystem. Don't forget, the Weilai car system also includes a Meizu phone component.

By the second quarter, due to well-known reasons, Huawei's sales had fallen out of the top five after Honor became independent. OPPO, Vivo, Xiaomi, Apple, and the newly independent Honor occupied the top spots. Huawei's current mobile phone shipments are comparable to Meizu's, each holding approximately 5% to 6% of the market share.

Meizu, a small factory in Zhuhai, can ship more than 500 million units every year, and its market share has always remained in a stable range. Both its system and brand promotion are quite good.

Maintaining a stable shipment volume, whether you call it a lack of ambition or a complacency, is still very beneficial to Weilai's intelligent driving research and development.

However, when it comes to ecosystems, it is indeed in direct competition with Xiaomi. Xiaomi has a better presence in home appliances and its mobile phone sales are much higher, but Xiaomi lacks presence in car sales and map services.

"In terms of the ecosystem, there is indeed some competition. However, if we consider Xiaomi a competitor in the field of artificial intelligence, then you can all resign."

If we divide artificial intelligence into three tiers, with the top 1-5 being the first tier, the top 6-10 being the second tier, and those ranked 10th and below being the third tier, then Xiaomi can only be considered part of the third tier.

Technological strength, commercialization capabilities, ecosystem influence, and R&D investment are all factors that affect the ranking of companies in the field of artificial intelligence. We won't discuss the first three, as everyone claims to be superior and outclass others in these areas.

In terms of R&D investment alone, Xiaomi can only be categorized as "other" in the field of artificial intelligence.

The top 5 companies in the field all invest more than 20 billion yuan each, with Huawei's annual R&D investment reaching more than 100 billion yuan; and their R&D expenditure in the field of artificial intelligence accounts for more than 60% of their total R&D expenditure.

Xiaomi does not have its own underlying framework and relies on open-source tools. Its capabilities in the B-end market are limited to mobile phones/IoT, with almost no enterprise-level solutions and low R&D investment, so it is not worth worrying about.

"As for their autonomous driving technology, their team was only acquired this year. Even if they are given a few years, they still won't be able to catch up."

Lei Jun's foray into car manufacturing will definitely make a splash, and it might even sell very well. Xiaomi's product capabilities have been quite good over the years; their phones are getting better and better, and their supply chain integration capabilities are also getting stronger.

Simply put, Xiaomi's aesthetic sense has improved. Tan Jincheng doesn't think that Xiaomi's cars would be unsellable, unlike some in the market. With Xiaomi's current aesthetic sense and Lei Jun's marketing capabilities, there's no way they wouldn't sell.

There are so many rubbish new car brands on the market that can still sell their products. Xiaomi at least has a brand foundation, but Tan Jincheng really doesn't believe they can make a big splash in the field of intelligent driving.

This isn't something you can achieve by simply integrating the supply chain and developing a few open-source software programs. It requires long-term R&D investment and finding the right direction. Every company in the top 5 or even top 10 of the industry has accumulated its technology through various explorations and repeated refinements.

Lei Jun's cars may sell well, but their intelligent driving system will definitely not be very good, not even as good as the new energy vehicle startups, let alone compared to companies like Weilai with a lot of technological accumulation.

"That's true, but the boss can't underestimate Xiaomi. After all, their ecosystems are similar. The key is that there's no information about their cars yet, and the automotive industry changes very quickly. Who knows, they might run into our models in the future."

"There is such a probability, but instead of worrying about Xiaomi, we should look at Huawei. The most crucial thing is when our artificial intelligence field can narrow the gap with Huawei."

The self-developed full-stack architecture, from chips to operating systems, and the super terminal network for intelligent driving—although the current high-end intelligent driving versions are expensive, once the cost is reduced, the advantages will be quite obvious.

"Yes, we need to accelerate our chip research and development. Boss, you can urge them again."

HiSilicon has mastered advanced manufacturing processes. Although it is unable to manufacture due to obstacles, it has no need to worry as long as it has the technology. On the other hand, National Technology can only master mature processes, namely the 40nm process which is already in mass production and the 28nm process which has been successfully taped out.

While it has achieved some success in domestic substitution, cutting-edge technologies are not so easy to obtain. In addition, like Tesla, Weilai is following the same path in the field of intelligent driving, which is to develop its own chips and LiDAR solutions, rather than Tesla's pure vision solution.

This is true for almost all domestic manufacturers. In the field of urban navigation assisted driving, both Chrysanthemum and Weilai have launched their products. The two companies are almost identical in terms of routes, and the difference probably lies in their scenario-based algorithm libraries.

"Haha, we may be speeding up, but we don't need to be afraid of them. Our advantages are still very obvious."

After gaining a comprehensive understanding of the breakthroughs in artificial intelligence, Tan Jincheng became quite confident. First of all, Weilai had an absolute advantage in sales. The new version of Seres Smart Choice, launched in April, officially began delivery in May, but sales were dismal.

Interestingly, while Da Zui had previously touted range-extended engines on various occasions, this version of the Smart is actually a range-extended model, priced between 24.68 and 34.68 yuan, directly competing with the Weilai L1.

200 vehicles were delivered in May and 300 in June, so the first model can be considered a failure.

(End of this chapter)