news

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

in the article "are ai phones and ai pcs a false proposition?", we talked about how ai terminals have a new role to play in the era of big models: the cloud-side big model solves the problem of deep intelligence, and the terminal-side big model solves the problem of timely intelligence. as functions become more powerful, we will hopefully see ai terminals become users' smart assistants and personal secretaries, automatically providing personalized functions and suggestions.

with the launch of apple's september conference, the most popular and discussed product is ai smartphones. we also expressed our internal views in "xiaomi, ov and honor, a new war will start in september".

today, our research focuses on the second largest carrier of ai terminals - aipc.

01

aipc is the second largest scenario in ai terminals

it's not just us who are paying attention to this. the entire smart terminal industry is in desperate need of a "gimmick" to boost its sales, and is looking forward to selling computers with the help of ai.

as for the direction of industry development, we are very optimistic about the trend of aipc. global pc sales will return to growth and reach 300 million units, mainly based on the following two reasons:

from the demand side, we believe that ai can help pcs better understand user needs, improve office efficiency, and meet personalized needs.

from the supply side, aipc will usher in an explosion of products starting in 2024. amd, intel, and qualcomm will intensively launch new cpus for aipc, and microsoft's new windows 12 will also integrate a large number of ai models and applications. brand manufacturers such as hp, dell, and lenovo will cooperate closely.

according to the forecast of foreign investment institution counterpoint, the global aipc penetration rate will exceed 70% in 2028. in this revolutionary tide, chinese manufacturers are even far ahead of the world in their pursuit of innovation. according to idc statistics, by 2025, the domestic aipc penetration rate may reach an astonishing 75%, even three years ahead of the world.

figure: china aipc penetration forecast source: idc, guotai junan

02

lenovo can't wait to get in on the action

in 2023, lenovo made a complete u-turn from the concept of intelligence and turned to the concept of ai.

after bidding farewell to brand mergers and acquisitions, not only lenovo, but the entire industry has been unable to find a secret recipe for computer sales to return to growth for many years. in 2018, lenovo chose to bet on the "intelligent" direction, took the lead in proposing the concept of "smart pc" and launched the world's first smart pc, and also held a smart computer innovation forum in 2021.

on the one hand, lenovo's computer sales completely follow the fluctuations of the industry, and it is difficult for its market share to have a significant breakthrough. on the other hand, with the sudden rise of ai in 2023, it is completely declared that lenovo's intelligentization has completely guessed the wrong direction.

however, lenovo turned very quickly and showed the greatest enthusiasm in the industry for the hot topic of ai. in october 2023, at the tech world conference held by lenovo, lenovo announced that it would be the first to embrace comprehensive ai and put forward the vision of "ai for all".

in 2024, lenovo, in pursuit of speed, began its strategy of flooding the market with machines.

it has to be said that lenovo's execution is quite outstanding, and it did not let "ai for all" remain at the slogan level: at the international consumer electronics show (ces) in january 2024, lenovo group directly revealed its strategy of flooding the market with machines, launching more than 10 ai pcs in one go, most of which are based on intel core's ulra series cpus. the goal is to cover different groups of people and application scenarios, including 6 products in the yoga series, 3 thinkbook products, and 1 thinkpad.

for example, the yoga series that empowers the creative process is equipped with the latest intel core ultra processor and lenovo ai chip. it is designed for creators who pursue high standards and is equipped with a copilot button and yogacreator zone generative ai software. users only need to click the copilot button to access the ai ​​assistant with one click.

obviously, lenovo has high hopes for aipc. in the latest earnings call, the company stated: looking ahead to the second half of the year, the company will launch different levels of products based on new edge ai chips from manufacturers such as intel and amd with performance upgrades. it is expected that the aipc penetration rate in the domestic market will reach 10% by the end of the year.

in order to drink this hot soup, lenovo adopted the fast-moving consumer goods strategy and built aipc into a sku matrix.

03

you can't eat hot tofu in a hurry

although we are optimistic about aipc and have great respect for every company working on the front line of technology, we are cautious about lenovo's strategy in the short term.

1. polishing the product is always the first priority, first take root and then spread the leaves

when it comes to technological innovation products that are groundbreaking, our views are clear and based on common sense, which is that we must first create a killer product, use the product to define the market, and if successful, then consider product diversification as the second step.

for example, apple defined what an excellent smartphone is through the iphone 4, which led to the subsequent tablets, headphones, wearables and ecosystems. tesla defined what an excellent electric car is through the model 3, which led to the current conception of the model q and humanoid robots.

we have also analyzed a typical negative example, that is sony, which went from being a definer of epoch-making products to a diversified company under the influence of a one-man show, and eventually lost its position as a technology leader. (for details, see "the $100 billion lesson: how did sony survive the long winter?").

aipc seems to be very trendy, and even people have not reached a consensus on its definition. but as a product that changes human usage habits, it must follow a simple principle: the product must be strong.

as for lenovo, the focus of thinking has always been on finding the right position. why is that?

launching more than 10 skus when a product is still immature, in addition to dazzling consumers, more importantly, reflects the company's lack of confidence at the bottom and the fear of missing any opportunity. it seems that as long as they launch products of different tiers and cooperate with every upstream of the industrial chain, they can grasp all directions.

in addition, through the product definition of aipc, it is not difficult to know that the core of this product is driven by models and chips. chips can be purchased from outside, so the end-side model must be its own. after lenovo couldn't wait to release 10 aipcs in january 2024, it took a full three months. at the lenovo innovation technology conference in april, the company released the end-side large model "lenovo xiaotian". a typical operation of getting on the bus first and buying the ticket later.

as a positive case, let’s take a look at how microsoft, one of the wintel alliance, thinks about aipc in order to prevent companies from losing their voice in pcs in the ai ​​era.

microsoft's first goal is to make this product more usable. before releasing aipc, we saw that microsoft did at least four things:

1) enhance the windows system. a large number of ai models and applications will be embedded in windows 12, which will be launched at the end of this year. the logic will be reconstructed from the system level to prevent the ecosystem from being destroyed.

2) in terms of native applications, copilot is used as a starting point to allow users to intuitively experience the efficiency improvement brought by ai from the very beginning of the product;

3) as the foundation of ai computing power on the edge, npu is still lagging behind in performance. microsoft is keeping a close eye on qualcomm, intel and amd to ensure the smoothness of ai applications. in september this year, microsoft and qualcomm announced that they would help oems launch windows 11 ai+ pc products in the price range of us$700-900, which is a blow to intel, which is lagging behind;

4) continue to improve the capabilities of cloud-based large models. among north american internet companies, microsoft's capital expenditure in the field of ai is the most aggressive, reaching us$50 billion. on the one hand, it supports the iteration of its openai's gpt general large model, and on the other hand, it trains its own large models.

no matter what the final hit product of aipc will look like, we can see that microsoft has made a full-scale layout from the underlying ecology, upper-level applications, chips and models, rather than a patch strategy of adding water when there is too much surface or adding more surface when there is too much water.

2. large quantity does not necessarily satisfy your hunger, the gap between high price and product quality

lenovo, the most anxious company, launched aipc first and won first place, but the market feedback was not satisfactory.

for example, regarding yoga 14s air snapdragon yuanqi, some netizens have the following comments:

it is not surprising that lenovo's aipc has received these comments. it does give people a consumer experience of using the name of ai to increase prices. among the more than 10 aipcs launched at the beginning of the year, the unit price ranges from us$1,000 to us$2,000. according to industry insiders, by introducing ai technology, the average price of lenovo computers can be increased by more than 30%.

at the investor exchange meeting, the company did not shy away from mentioning that starting in 2024, as demand grows and ai performance improves, product prices will rise steadily, thereby increasing the average unit price of consumer-grade pcs.

in the early stages of development, the strategy of flooding the market with products seemed to be aimed at not missing out on the trend, but in fact it was a strategic trick. taking advantage of the trend of ai, the focus was placed on price rather than product strength, which was a deviation from the right direction.

although oems such as asus, hp and dell are also very anxious about the aipc trend, those who launched the strategy of flooding the market with machines at the very beginning still stand out.

04

the current thinking is hard to change the fate of weak profitability

1. fatalism still needs to be broken

if we start from a fatalistic perspective, it is not difficult to understand why lenovo fired the first shot in aipc in this way.

in the 1990s, lenovo had the famous "liu ni dispute", which was regarded as a duel between the market faction and the technology faction. at that time, ni guangnan insisted on the "technology, industry and trade" line, that is, to take r&d as the foundation and use technological achievements to promote production and trade.

finally, ni guangnan was fired in 1995, and the market faction that advocated "trade, industry and technology" won a comprehensive victory, emphasizing that primitive accumulation should be achieved through trade before discussing technological development. yang yuanqing, who was named by liu chuanzhi to take over, also rose to the top due to his "sales genius".

after 30 years, lenovo is still struggling in the dilemma of trade, industry and technology, and has not yet achieved the goal of using trade to support technology. the company's revenue has fluctuated around 400 billion yuan in the past five years, and most of its attempts at the second growth curve have been unsuccessful, and its net profit can only follow the industry cycle.

under such a big theme as aipc, lenovo, at least judging from all its actions so far, has been based on the traditional trade, industry and technology thinking of "how to do business well" rather than the technology, industry and trade thinking of "how to make good products".

2. now is the time to transform

personal computers are a market worth over one trillion dollars, and lenovo holds over 20% of the global market share. however, with annual revenues of 400 billion dollars, lenovo can only achieve profits of around 10 billion dollars.

the company's gross profit margin has been maintained at around 16% for a long time, which is only similar to luxshare precision, which focuses on oem. the net profit margin is hovering in the low single digits.

if the traditional division of labor in the computer industry has been determined, it is understandable that lenovo, after missing the historical opportunity to make chips in the 1990s, has now made it clear that it will not make chips and systems.

however, if you still use the old thinking and old path for aipc, this clear turning point opportunity that will change the rules of the game in the industry, the outcome is basically doomed. you will get to drink the soup while others eat the meat. you will not miss it, but it will be difficult to achieve great things.

but if lenovo takes the narrow path and faces difficulties, as the only domestic pc company, it still has many opportunities to break through.

we hope to see the latter.