But at this moment, Chen Yu looked at Fang Hong and said: "Xingyu Technology's SOCL wants to break NVIDIA's monopoly in the field of graphics card programming under the existing framework, I think it is very difficult, because what Qin Feng built is too general, but it will not be easy to use." ”

"Now the graphics card models of major manufacturers are complex, and the performance is uneven, if you want to make a framework under this specification, you must adapt to all the graphics cards, which is too troublesome, NVIDIA can optimize for CUDA, but other graphics card manufacturers are not obliged to optimize for your Xingyu Technology before SOCL occupies a dominant position, CUDA came out five or six years ago, how long has SOCL come out?"

"Therefore, I think that the performance of SOCL under the same computing power is likely to be inferior to CUDA, and it is more difficult for developers to use, and there will be fewer people who use it over time, and it is too difficult for companies that do not sell chips to launch specifications that adapt to all graphics cards."

After a while, Fang Hong said unhurriedly: "It's not suitable for companies that don't sell chips, so what about companies that sell chips?" ”

Hearing this, Chen Yu immediately replied firmly: "It's still too late, Nvidia has not yet reached a monopoly position, and there is still a good chance to break through." ”

Chen Yu added: "At present, my idea is to either develop my own GPU specifically for AI training, or you can talk to Qin Feng and ask him to spin off SOCL from Xingyu Technology and specialize in graphics cards, and Xingyu Technology will give ecological support." ”

At the end, Fang Hong sighed and said as if talking to himself: "Speaking of which, the most important thing is the chip, which is the foundation of everything." ”

Fang Hong, who has foresight, knows that in another decade or so, the field of artificial intelligence will be hot, and it is extremely important to be able to surpass his own AI model, but the underlying hardware capabilities under AI are just as critical, or even more critical.

The design and manufacture of the entire computing chip is also more difficult.

At the end, Fang Hong had already made a decision in his heart, he turned to look at Chen Yu and said: "I'll get through Xingyu Technology, I will let Qin Feng split the SOCL department, and Qunxing Capital has invested in some mainland graphics card manufacturers, and they will also be reorganized and integrated." ”

Qunxing Capital is actively promoting the domestic substitution of the entire semiconductor industry chain, and GPU will naturally not be absent.

But I have to say that this field is very rolled, especially since 2000, there are many graphics card manufacturers, and now more and more peers are being swept to death.

Qunxing Capital has a lot of money, and the three manufacturers it invested in are still alive, but the gap with international manufacturers such as Nvidia and AMD is self-evident.

After a while, Fang Hong added with a smile: "After integrating it, put it under your quantitative capital and hand it over to you to do, how about you go to PK with Lao Huang?" ”

Under normal circumstances, this situation is basically impossible to achieve, and quantitative capital does not have this ability, but whether it is Xingyu Technology or quantitative capital, there is a common big boss.

With Fang Hong's deployment, the resistance to integration is much smaller, and it can even be achieved easily.

At this moment, Fang Hong has already judged that there is no need for the graphics card manufacturers invested by Qunxing Capital to "raise Gu", so they will integrate and build one, and let Chen Yu do it.

He can see so clearly, and his technical foundation is also strong, if he can't get Lao Huang's NVIDIA, the other manufacturers will also expect too much.

So simply concentrate funds, concentrate the team, and bet on Chen Yu in this subdivision, although he is now doing investment transactions, and it seems that he can't fight with high-tech companies, but the team under his hands is all talents that major technology companies are rushing to want.

Fang Hong also has this courage, and he should make a decisive bet at the critical moment.

Thousands of troops are easy to obtain, but one will be hard to find.

Chen Yu was also quite surprised to hear his decision, and after a little thought, he simply nodded: "Okay." ”

A simple one-word reply.

This matter was so decided, and Qin Feng went over to talk to him in the afternoon.

After a while, Fang Hong asked, "Has the annual report come out?" ”

Chen Yu shook his head: "Not yet, but last year's annual income of 75 billion is not a big problem, the net profit of 10 billion should be there, the cost of running the model is too large, especially the cost of computing power, otherwise there must be a net profit margin of 45%." ”

Fang Hong has a general understanding of the situation of quantitative capital, the main income is investment income, and about 40% of the company's income in the capital market is shared by the team below.

The backbone trading members of Chen Yu's team have been working with him for two years, and now they are worth tens of millions, and they can gather so many top talents, and they can't look down on the recruitment of those big technology companies, and a very key reason is that Chen Yu gives too much.

For example, if a capital plate earns 5 billion, the team that makes this plate can share 200 million, and if it earns 10 billion, it will share 4 billion, and the more you earn, the more you will score.

The company's second largest expenditure is to develop and iteratively upgrade AI models, with an investment of more than 2.5 billion yuan last year, and it is soaring on a large scale.

After a while, Fang Hong said: "SOCL is integrated into your company, and the future R&D investment must be tens of billions. ”

Chen Yu replied with a smile: "It should not be necessary, the company's own profitability covers the required R & D investment is not a big problem, last year we earned 7.5 billion, but at the beginning of this year, the company adopted a new model to trade, the winning rate of a single transaction reached more than 70%, and there are currently 16 capital disks, 11 strategic mainland markets, and 5 strategic overseas markets." ”

Chen Yu continued: "In the overseas market, the strategic long Bitcoin capital market has earned at least 550 million US dollars, and the other 4 plates add up to 1.3 billion US dollars, which adds up to 11.7 billion yuan without excluding exchange rate changes, and the 11 domestic capital disks currently make a profit of 8.9 billion yuan. It's only early April now, and it's not a big problem to make a profit of 35 billion yuan in the first half of this year. ”

Hearing that quantitative capital has funds to make Bitcoin, Fang Hong thought for a moment and said: "Bitcoin will be empty next week when it sees about $230." ”

Hearing this, Chen Yu couldn't help but be stunned, and then nodded in response: "Understood." ”

Hearing the words of the big boss, he immediately thought that the largest trading platform in the cryptocurrency market is in the hands of Qunxing Capital, it should be said that it is in the hands of Merrill Lynch, but the management is in the hands of Qunxing Capital.

The year before last, the price of Bitcoin soared above the $1,000 level, and Qunxing Capital backhanded Merrill Lynch to trap Merrill Lynch, and then acquired the X trading platform and then resold it to Merrill Lynch.

So, the big boss lets you empty at this time, this is the inside news.

It is worth mentioning that the current price of bitcoin has been from the previous explosion to $2, in 2012 at the bottom of the operation, the price did not exceed $10, but after entering 2013, the price trend of bitcoin exceeded $10 and soared, and now the price has risen to $140.

Since the beginning of this year, the price of bitcoin has strengthened due to two factors, one is the halving of the first bitcoin mining output in November last year, and the other factor is the outbreak of the Sepaulus debt crisis in January this year.

Qunxing Capital and Merrill Lynch took the opportunity to advocate and instill in the market labels such as "decentralization" and constant aggregate.

It's true that Merrill Lynch and Qunxing hate each other, but it's also true that they work together, Merrill Lynch has no choice but to be trapped, and he doesn't want to affect Wall Street's grand strategy of "directional blasting" of Qunxing Capital.

It's just that this year, I never expected to be "targeted blasting" by Qunxing Capital first, causing Wall Street to lose $30 billion, which is still a bit painful.

But in the spirit of you, the stars win 99 times, I win Wall Street, and as long as I win the 100th time, I will win them all back.

I still chose to endure it.

At this time, taking advantage of the halving of bitcoin production and the debt crisis on the European side, cryptocurrencies are blowing.

……

(End of chapter)