Chapter 226 The Dilemma of HKUST Feixun
The interface on the mobile phone screen is constantly changing.
[Precondition detection...]
[Available area within the area: Passed]
[Account deducted successfully: HKD 15 million]
[Construction process starts: Infrastructure Planning and Construction Center]
The interface on the screen switches and a progress bar appears.
Hu Jiangshan breathed a sigh of relief, and the next 15 days of construction period began.
This time is not long, as it takes a month to build a semiconductor production line.
Even the sci-fi office building on Tenglong Island took less than a month to build.
For example, the Infrastructure Planning and Construction Center requires a building area of 5,000 square meters, which is quite large. It only takes fifteen days, which is indeed extremely fast.
In fact, Hu Jiangshan was also spoiled by the Technology Engine APP.
You must know that the normal construction of any of the above-mentioned buildings, including tidal power stations and seawater desalination treatment facilities, will take at least several years, even if technical and financial issues are put aside.
But then again, after the infrastructure planning and construction center is completed, it does not mean that the infrastructure on Tenglong Island is completed. It still needs to be designed and constructed.
It is estimated that Hu Jiangshan will not be able to come down within a few months after all the construction is completed.
However, the facilities in the Technology Engine app are relatively complete, such as the Special Tools Research and Guidance Center, which Hu Jiangshan feels very handy. Presumably this infrastructure planning and construction center should also be good.
The next day, Hu Jiangshan went to Jiangjun Island again early in the morning.
In order to design the commuter boat, Hu Jiangshan consulted a lot of information and spent a few days to complete it. The commuter boat is a medium-sized hull and is quite powerful.
Hu Jiangshan conducted functional tests, and sure enough they all passed smoothly, and then the construction process began.
The first commuter boat was about to be launched. To be sure, Hu Jiangshan wanted to check the construction status. There should be no problems with such a large vehicle.
At the same time, Huaguo, Heyue City.
Strictly speaking, Heyue City is a second-tier city in China, but in the past ten years, Heyue City has developed very rapidly. Many emerging industries have developed well here, including some technology manufacturing giants. It has a headquarters or R&D center, so Heyue City is now officially known as a new first-tier city and is the core city in the region.
HKUST Feixun is one of the technology giants that has set up its headquarters in Heyue City.
You know, HKUST Feixun is very famous in China. It is now the leader in the field of artificial intelligence in China. Only a few companies can compare with this company, and it is far ahead.
In a few years, artificial intelligence will be identified as the future development direction in the world, and may even trigger a technological revolution. The status of HKUST Feixun will naturally become more stable.
If we develop according to history, HKUST Feixun will be in big trouble at that time, related to artificial intelligence chips.
But at this time, due to various reasons, the trouble seemed to have arrived early.
Boss, office, Liu Qingfu is sitting at his desk drinking tea.
Liu Qingfu is now in his forties or fifties. He is generally well-maintained, but his hair is a little sparse and has a Mediterranean trend, giving people the impression of overusing his brain.
Liu Qingfu has a technical background, and he is an excellent technical master with a very famous reputation.
As the saying goes, grass does not grow on busy roads, and hair does not grow on smart heads.
Although there is no scientific basis for this statement, there are many examples that show this.
Liu Qingfu took a sip of the tea water. The taste felt a bit astringent. Maybe the tea had been left for a long time and became a bit moldy.
Liu Qingfu is not a very particular person, and the tea he drinks is bought from any tea shop. The main thing is that he can drink it and relieve his boredom.
Liu Qingfu put down his tea cup and stared at the screen. He was looking at a report carefully. After a long time, he couldn't help but sigh.
The company's situation is actually not very good recently, and it has encountered many problems.
There are bottlenecks in technological development, and what's worse is that it has never found a new profit point. Now Feixun Technology is at risk of losing money.
Although it is normal for technology companies to lose money. After all, R&D investment is relatively large. As long as it is always favored by others and there is always capital injection, it doesn't matter. But Liu Qingfu doesn't like this.
Liu Qingfu is a very traditional entrepreneur. If the company suffers losses, or even loses money for a long time, he will feel uncomfortable sleeping.
But Liu Qingfu feels that his company will not reach that stage. HKUST Feixun still has core technologies and hard power.
You must know that although HKUST Feixun has produced many hardware devices, such as translators, reading pens, and even learning machines, the real technical strength lies in the software.
HKUST Feixun started out as a translation software company and has been deeply engaged in it. It is a well-known and far-leading presence in the translation industry in China.
However, now that the translation software and translation hardware industries have reached their peak, it will be difficult to move forward and improve. Without technological changes, it will be difficult to make major breakthroughs.
Liu Qingfu is a very visionary person and knows this very well, so he organized an in-depth discussion with the company's senior management a few years ago, and then focused the next technological breakthrough in one field - artificial intelligence.
Now HKUST Feixun has developed very well in artificial intelligence, especially in terms of algorithms. It can be said that it is a leader in the world.
At this time, the famous openAI company had just been established in the next few years, and the so-called chatgpt large model only had a prototype, and it had not yet achieved anything famous.
HKUST Feixun is now developing the same product and has already produced results.
To put it bluntly, HKUST Feixun is actually a pioneer and leader in large-scale artificial intelligence models.
If HKUST Feixun continues to develop smoothly, it will inevitably become the Beidou of Blue Star artificial intelligence, and its comprehensive strength can crush any opponent.
But the problem is that HKUST Feixun will encounter many obstacles, and the biggest obstacle is hardware.
The chips that can now support computing are not controlled by the Chinese. They use GPU chips from the lighthouse company Eagle Weida.
As we all know, GPU chips are generally used on game graphics cards, mainly to perform graphics calculations and support game operations. However, GPU chips also derive a capability, which is data calculation.
The more powerful the GPU chip, the faster the data calculation.
So later, a large number of game graphics chips were used on mining machines. In the so-called mining factory, there were whole rows of graphics cards connected in series, which was very spectacular.
The most popular among these graphics cards is the product of Eagle Weida, because the GPU computing power is powerful enough, the computing speed is fast enough, the mining efficiency is the highest, and there is almost no opponent.
Although the process difficulty of GPU chips is much lower than that of processor chips, that is, logic chips, Yingweida has been in this industry for many years and is not a vegetarian. It has almost formed a monopoly, so in the field of mining machines While making a lot of money.
Since GPU chips can be used for mining, they can certainly be used for artificial intelligence computing.
Since the emergence of the concept of artificial intelligence, Yingweida's GPU chips have begun to sell well again, and more and more of them are sold, and their market share is getting higher and higher. Of course, the market is still too small now, and Yingweida has not made much money in it, but this may not necessarily be the case in the future.
Eagle Weida is also very optimistic about the field of artificial intelligence hardware and is vigorously developing it, doing its best to seize market share and attempt to monopolize this market.
How does Yingweida do it?
Previously, Yingweida did not regulate the market, allowing all companies engaged in artificial intelligence development to buy GPU chips or game graphics cards at will, and the prices were very reasonable, and all were shipped at wholesale prices.
Companies in the artificial intelligence industry have very high demand for GPU chips, with orders ranging from tens of thousands to hundreds of thousands or even millions of units.
If the price is too high, the cost is too high.
The key is that artificial intelligence is still an emerging industry, and there are not many effective channels for monetization. The investment is too large, and even large companies cannot bear it.
Yingweida has achieved its goal, and now its market share has reached more than 80%, so the butcher's knife has finally fallen.
Eagle Weida began to raise prices to artificial intelligence companies that purchase large quantities of graphics chips.
Moreover, the prices imposed on Chinese companies are the harshest, and HKUST Feixun will be the most affected.
Now Liu Qingfu is so worried when he sees the report, because the gap in the GPU chip is getting bigger and bigger.
As the demand for computing power increases, the demand for GPU chips will also increase.
But the paradox is that Yingweida's quotation for HKUST Feixun is getting higher and higher. It has now nearly doubled the original price, and Yingweida seems to want to raise the price even more, which is simply unbearable.
The company's senior management, and even the boss Liu Qingfu himself, felt that Yingweida was playing some trick. They learned about the development progress of HKUST Feixun's computing power through some channels, and found that the development was too rapid and that more computing chips were urgently needed. Formulated a price increase strategy.
Anyway, only Yingweida can ship the product. No matter how expensive it is, HKUST Feixun can only buy it unless it no longer operates in this industry.
This is very possible. After all, HKUST Feixun has not implemented strict confidentiality measures. It is normal for any information to be revealed.
It's useless to say anything now, except for the lack of chips.
If you want to get enough GPU chips, you must place an order with Eagle Weida, and you will need to pay a much higher price than before.
HKUST Feixun is not a very wealthy company. Although it has external capital injections and even official financial support, it is indeed not short of money, but it is not the kind of company that can spend money lavishly.
Fortunately, there is good news now. Huayao's artificial intelligence chip, Shengteng 910, will be released soon.
Some time ago, Huayao also sent a sample of the Ascend 910. After testing by the HKUST Feixun Technology Department, it completely exceeded expectations and was even much better than the most powerful GPU chip of Eagle Video.
At that time, the top management of HKUST Feixun was very excited. If Huayao could really mass-produce the Shengteng 910 and ship it in large quantities, then HKUST Feixun would be able to get rid of Yingweida's high-priced GPU chips and would no longer be a Kaizi. Let others take over.
Moreover, Ascend 910 is a dedicated artificial intelligence computing chip. It removes the useless graphics computing function. It is much easier to use than Yingweida's GPU chip, and the performance will of course be improved a lot.
More importantly, Huayao's price is not high.
The technical department has calculated that it is more cost-effective to buy a thousand Ascend 910 chips than to buy two hundred Eagle Video GPU chips.
But unfortunately, there is no news from Huayao yet, and it is not yet certain whether the Shengteng 910 can be successfully mass-produced.
This matter has never been settled, and Liu Qingfu couldn't sleep well all day.
Sad, so sad.
At this moment, there was a knock on the door.
"Please come in."
The door opened and Jiang Jinlong came in.
As mentioned earlier, Jiang Jinlong is the technical director of HKUST Feixun. He is very capable and well-known at home and abroad.
"Golden Dragon, please sit down quickly. Would you like some tea?"
Jiang Jinlong waved his hand and said that he didn't drink tea. He probably knew that the tea in Liu Qingfu's place was not very good and there was nothing good to drink.
Jiang Jinlong said: "Mr. Liu, I want to tell you something."
Liu Qingfu nodded: "Well, if you have anything to say, just say it."
Jiang Jinlong said: "Mr. Liu, you should have seen the latest report, right?"
"Yes, I saw it."
Liu Qingfu took a sip of tea and already knew what Jiang Jinlong wanted to say.
"Mr. Liu, we have to find a way now. The computing power has reached a bottleneck and we need more hardware to develop further."
Liu Qingfu sighed: "There is still no news from Huayao. There is really nothing we can do. If it really doesn't work..."
Liu Qingfu looked out the window and said, "If it doesn't work, let's talk to the people from Yingweida."
Jiang Jinlong was startled, then shook his head: "Mr. Liu, this is not possible. Yingweida's quotation is too high. Even if we can negotiate a little, it will still be expensive."
Liu Qingfu sighed: "Then there is no other way. No one else can do this."
Jiang Jinlong thought for a while: "If you don't talk to people at AMD, their price may be lower."
Liu Qingfu snorted coldly: "It is indeed cheaper, but not much lower, and the performance of their products... is not as strong as Yingweida's.",
Jiang Jinlong was right. Even if AMD's products are cheaper, if performance is taken into account, it may not be more suitable to just buy Yingweida's GPU chips.
The water far away in Huayao cannot quench the thirst nearby.
Now I'm really in a dilemma.
Liu Qingfu sighed again: "There is no other way. Let's make an appointment with the people from Yingweida. We will discuss it carefully with them and lower the price as much as we can."
Jiang Jinlong had no choice but to nod.
Yingweida's work efficiency is quite good, and it is already quite strong among the many large companies in the Lighthouse Country.
The next day, Yingweida's special commissioner had arrived in Heyue City.
There were two people in total. They got off the high-speed train and called people from HKUST Feixun. They agreed to send a car to pick them up and take them to the headquarters.
(End of chapter)