Harold looked at Hu Yiting's expression very seriously, and asked with some surprise:

"Do you really want to buy it?

Can you afford it?"

As soon as Hu Yiting heard this, his heart fell back into his stomach from his throat, and he complained secretly: "Why are you still afraid that customers will not have money to pay when you open a restaurant!"

Harold continued: "The price of a full license is very expensive, and I don't think you as a small start-up IC company from China can afford it.

It's not that I underestimate you, Hu, so far, the only companies that have spent money to buy a full license for ARM are GEC Price Semiconductor, Texas Instruments, and Samsung Semiconductor, which are very expensive!"

Hu Yiting was dumbfounded, and immediately realized that this ARM was not ARM, and the ARM with more than 1,000 licensing companies and 320 technology licensing partners in the future had not yet been born.

Moreover, even in later generations, only 15 of ARM's 320 technology licensing partners have full permissions, and can be allowed to make changes, additions and deletions to the instruction set and architecture on the basis of ARM's technology patents to form their own product series, such as Qualcomm and others.

&nb's way of survival is an open and inclusive win-win way. Because in the computer processor industry, no one can compete with Intel independently, and ARM can only survive by taking this path.

In other words, in the traditional PC field, semiconductor manufacturers generally have two ways to choose.

First of all, like Intel, it takes care of itself from beginning to end, and the design and production of architecture and chips do not depend on anyone. In fact, Intel also provided a limited x86 instruction set authorization in the early days, but it was quickly recovered, leaving only a few nail households in the United States.

But Intel needs extremely strong and all-round strength to do this, and it has to have money, people, and technology, especially in the future when semiconductor technology is becoming more and more complex, there are only a handful of companies that can be so domineering and monopolistic.

The benefits are certainly clear. Not only can you completely control your own lifeline, but the profit is also extremely considerable, almost any product of Intel can enjoy a very high profit, and you can sell as much as you want. When the new processor products are launched, the gross profit margin starts at 90%!

But in the mobile field, this black-hearted monopoly model will not work, because the cost of processors in PC systems is very high, but in smartphones and tablets, it is often less than 10%. Even less than 5%, no matter how hard you try, you won't make much profit. This is also one of the key reasons why Intel has struggled to promote in the mobile space.

Intel's profiteering model is simply not open in the mobile field, and if the price is lowered to grab the market, it will hit the price of its own PC processor with the same technology, and then people will ask, why use the same technology, the chip you sell on the mobile phone is so much cheaper than the computer! Moreover, the performance of Intel's processor is much worse than that of ARM at extremely low power consumption. What's more, almost all semiconductor manufacturers hate Intel, and everyone understands that if ARM dies, Intel will definitely tear its face to reveal its usual greed and shamelessness. In the end, everyone was forced to have no way to live, so almost no manufacturer would choose Intel's products when the performance was close.

For semiconductor manufacturers, Intel is an alternative, and another way is the mainstream, that is, the fabless model.

NVIDIA is like this, AMD can't afford Intel, and the same is true for Chongguang at present.

Such companies only design their own chips, and the manufacturing is handed over to foundries, such as TSMC, UMC, GlobalFoundries, and Samsung Electronics.

The benefits are clear, and the burden on businesses is light. You can just design it yourself, and you don't need to spend a lot of money to build a wafer factory and develop a new process, but the disadvantages are also very prominent: if you design it, can you build it, even if you can build it. You can't decide what it looks like, it depends on the ability of the foundry partner.

Of course, there are many lessons in this regard: TSMC's 40/28nm (nano) process was initially immature, and the production capacity was also delayed, which dragged down the entire industry; GlobalFoundries' 32nm process did not meet AMD's expected level, and the frequency and voltage of the first-generation FX/APU processor were much worse than designed, and the 28nm process was blown for a long time. As a result, AMD was forced to give up a whole generation of low-power APUs for a time, and had to redesign and then go to TSMC, but the market was robbed by Intel, and AMD almost lost a heart attack.

Unlike the above two approaches, ARM is completely different from them.

First of all, it does not manufacture or sell any chips, but just designs its own IP, including instruction set architecture, microprocessor, graphics core, and interconnection architecture, and then sells the license to whoever likes it. With ARM's IP (Intellectual Property Core Module), customers can do whatever they want, develop their own chip products, and then manufacture them themselves or find foundries to produce them.

Hu Yiting didn't expect that the current ARM would be so downcast, and there were only three customers who purchased their full permissions.

"Doesn't Apple have your full license?"

Harold was unwitting and defenseless in the face of the Chinese teenager in front of him, and said frankly: "Apple has it, DEC has it, and Olivia has it, but the company has licensed it to them for free, Apple and Olivia are shareholders of ARM, and DEC promised to help ARM improve the technology and share the technical improvements based on ARM's original design for free." ”

Hu Yiting swallowed his saliva: "How much do you need to sell the full license?"

Harold said: "Our business is very simple, which is to provide external licensing in three different models.

The first is processor licensing, we ARM designs a CPU, and then licenses and sells it to partners. After you buy them, you can only implement them according to the drawings, and there are not many places that can be played, but how to implement them is up to you, such as configuring which modules, how many cores, how many caches, how high frequency, what process, who will do the foundry, and so on.

We at ARM will also provide you with guidance and assistance, but how to turn the solution into a chip and set it in what parameters and specifications depends on your choice and technical level.

If you want to make ARM processors, but you have limited strength, what should you do? Our company has also prepared a series of optimized processor design solutions for customers. You can choose the right one according to your needs and use it directly.

In this way, you can make a product quickly, and the cost is also very low, and of course, there is less room for free play. The processor type, foundry, and process are all prescribed. ”

Seeing Harold's patient and meticulous explanation, Hu Yiting said gratefully: "Thank you, I know." ”

Harold had no other clients to see today, and was also bored, and continued: "Secondly. We have architecture and instruction set authorization.

If you are strong, you can buy ARM's architecture/instruction set, such as ARM-V2, ARM-V3 and ARM-V4 under development, and then you can research and design the chip yourself.

These processors designed by you and those designed by ARM are instruction set compatible, but there is no direct comparison, because they use the same instruction set and architecture, and belong to the same level of products.

This kind of authorization is generally not used by companies with weak technical strength, and so far, only a few companies with strong technical strength, such as Apple, Texas Instruments, and DEC, can develop it on their own.

If you make changes to the instruction set and architecture in your design. The price of this full license is higher, and of course, the technical requirements for your company are also higher. ”

After Harold finished speaking, he looked at Hu Yiting and laughed.

Hu Yiting also smiled: "How much is it?" The second kind of authorization you mentioned, I was talking about complete permissions. ”

Harold shook his head helplessly, he didn't believe that a Chinese company had such development strength, nor did he believe that the other party had the funds to dare to smash it into it, you must know that chip development is a bottomless pit, and every link such as follow-up testing and tape-out is a money-burning thing.

Harold still has a good impression of Hu Yiting, this teenager is fluent in English. Generous, solid skills, gentle and elegant, so that he has a desire to give support to his younger generations.

&nb-V3 version instruction set and architecture. The full license price we provide to Samsung Electronics is $300,000, and an additional $200,000 post-licensing fee will be added after the product is designed, and no post-licensing fee will be charged if the design fails. The royalty of each chip is 1%, which is charged based on the selling price, and if it is completely digested internally, it will be charged according to the average price of ARM chips of the same level in the market. Technical patents for buses and interfaces are given free of charge. The full package of development software and specialized tools is $150,000. In addition, we provide technical support. However, the price of each engineer sent is negotiable, but it is generally 8,000 pounds per day for technical services, plus 6,000 pounds per day for labor subsidies. ”

Harold quoted the price in one breath and said with a smile: "Hu, you report what I said to your company's CEO, Ms. Wang, and she will definitely sue you to buy a ready-made processor core design patent authorization." After all, the design is ready-made, and after you buy it, you can simply embed it in your own chip design, which is enough to solve the problem. If you really can't solve the problem, you can apply for our ARM technical support. ”

After speaking, Harold said sincerely: "I sincerely hope that in the future, the more companies that license ARM's patented technology, the better, and hope that you can all produce excellent products and find your own foothold in the market, which will make all ARM employees feel gratified." I will never lie to you and make you spend money on a product that is not suitable for your company. Hu, trust me. ”

Hu Yiting heard him say this, slightly moved, technical people are like this, especially technical people who are in the difficult period of eating deflated, and they dream that their technology can be recognized and sought after by the market as soon as possible.

"But I still decided to buy the full license, I think ARM's technology is worth the price, and Chongguang also has the ability to develop its own products, since the Koreans can do it, the Chinese can do it too!"

Harold frowned slightly, and began to feel that Hu Yiting was stupid, and said bitterly: "Do you know how long the independent development cycle takes? The development cycle of a chip with embedded processor design generally takes two to three years, of course, if you don't modify the architecture, it will be very fast, but if that's the case, why waste money on full permissions?"

Also, the products designed by Samsung can be digested internally by themselves, I heard you say that Chongguang is a pure IC company, assuming that you do not fail in design, and finally design a product, then who will buy your products? In case there is no market, the early research and development expenses are equivalent to a loss in vain. ”

Hu Yiting felt that Harold was really a rare honest man, and if he changed someone else, he was afraid that it would be too late to coax and deceive the customer to pay more, how could there be such a reason to push the door-to-door business out.

Of course, Hu Yiting also knows that according to later statistics, 20% of the companies that buy ARM patent licenses will go bankrupt before developing products, because the market competition is too fierce, but Chongguang Company is not the kind of second-rate enterprise without technology.

"Thank you for your kindness, Mr. Harold, I insist!"

Harold laughed helplessly: "Okay, you can ask your CEO when you go back and see if she will approve it."

In addition, please be sure to explain that ARM has absolutely no behavior to deceive customers by conveying my words in full. ”

After Harold finished speaking, he felt that Chongguang Company was really mismanaged, and he actually sent an impulsive teenager like Hu Yiting to talk about technology licensing, if the other company really finally agreed to buy full development rights, then don't blame me for not warning me in advance if I lost money. You know, even ARM can't produce a competitive processor with ARM-V3 architecture. (To be continued.) )

PS: Make up for the chapter that was interrupted that day, thank you again for your support!