There is no doubt about the strength of Texas Instruments’ R&D team.
At present, most of Xingyi Semiconductor's core R&D talents are recruited from Texas Instruments. They are selected from the best and are naturally very capable.
There are no surprises. The smart home SOC chip, fast charging chip, and power management chip have all been successfully taped out.
The next step is trial production. If there are no problems in trial production, full mass production can be launched!
By then, Xingyi Semiconductor will enter a new era of formal mass production from the PPT era!
Although these three chips are not high-end chips, everything is difficult at the beginning.
With these three chips taking the lead, subsequent high-end mobile phone chips, baseband chips, and even next year's high-end SOC chips with integrated baseband will all come naturally.
It is too difficult to directly develop a 28nm SOC chip with integrated baseband this year. After all, Qualcomm will have to do it next year.
No matter how good the person Texas Instruments poached, it will be next year.
After all, the baseband problem is more difficult to solve than the processor, not to mention the integrated baseband, which is even more difficult.
Therefore, Wang Yi had to make a choice, slow down the development of SOC chips for integrated baseband, and switch to a plug-in baseband solution.
Develop 28nm processor chips and 28nm baseband chips separately.
This plug-in solution has the same strategy as Nvidia Tegra 3 and Qualcomm's flagship processor 8064 in the second half of the year. It is much less difficult and is sufficient.
In the first half of next year, Qualcomm will launch the Snapdragon 800 SOC with integrated baseband, and it will not be too late for Xingyi Semiconductor to launch the Xingyi SOC with integrated baseband.
After all, NVIDIA next year will still be a plug-in solution.
This strategy is more stable and safer.
Moreover, it is developed separately, and the processor does not integrate the baseband. It can castrate the modules that are not used by the TV, optimize the video decoding capabilities, and use it as a 4K TV chip!
In this way, you kill two birds with one stone!
Even without castration, just mass-produce more, and use mobile phone chips as 4K TV chips, no problem.
The Snapdragon 600 chip used in the previous generation of Xiaomi TVs was the same Snapdragon 600 used in mobile phones.
Although the material cost of this solution is a little higher than that of castration, it saves a series of processes such as re-modifying the design, re-striping, re-trial production, restarting the production line, etc. Overall, the cost is lower.
After all, there are so many codes for chip design. If you change a line, you have to re-tape it out.
The cost of a tape-out is not low.
Xingyi Semiconductor's tape-outs this time are all 40-nanometer low-end chips, and the tape-out price is relatively low, only about US$800,000-900,000, which is about 5 million yuan per time.
But the price of 28nm is much higher.
The current price of a 28nm tape-out is US$5 million, which is more than 32 million yuan.
In a few years, when the 28nm process matures and production capacity is sufficient, the tape-out price will drop to US$2 million.
There is nothing we can do about this.
And tape-out may not be 100% successful even once.
In the previous life, Xiaomi's Pascal S2 chip failed many times during tape-out, and hundreds of millions of yuan were wasted. Even the wealthy Rebus couldn't bear it, and had to give up the research and development of the S2 chip.
Similarly, the current 28-nanometer chip tape-out cost is too high. Xingyi Technology can also directly choose to use the 28-nanometer mobile phone chip as a TV chip, saving the need for re-development and re-tape-out.
Wait until next year to develop a higher-end, more professional TV picture quality chip.
This is no problem.
After thinking about this clearly, Wang Yi came to the semiconductor department.
Williams came forward with a smile: "Living up to expectations, all three chips were successfully taped out at once!"
"Very good. The Xingyi Semiconductor Department has made great contributions. Each member of the George R&D team will receive three more months of year-end bonus!"
"I'll thank Director Wang on their behalf!" Williams said with a smile.
The research and development of these chips was all done by George's department, and the bonuses also fell on George and the others.
As for Williams and his department, it depends on whether the mobile phone chip can be successfully developed.
Successful research and development will result in more bonuses.
R&D failed, drink from the northwest wind.
Wang Yi has always made clear rewards and punishments. Even the same department must be managed quantitatively and refuses to eat from the same pot.
"By the way, Chairman, all three chips have been put into trial production, and we should officially name them. The mobile phone chip we are developing also lacks an official name!"
"Then let's name them according to categories." Wang Yi thought for a moment:
"Smart home, the abbreviation of SH is a bit cumbersome, so let's call it H, home. Our first smart home chip is called smart home H1 chip."
"Power, the first power management chip is called P1."
"Charging Charge, the first fast charging chip is called C1!"
Williams agreed very much: "This is so simple and clear that even if there are more chips in the future, there will be no confusion."
Wang Yi nodded: "Yes, and these low-end chips are not mobile phone chips. They do not have high performance requirements, and they do not need to be upgraded every year. On the contrary, one generation can be used for several years."
When the next generation change is expected to be many years later, Xingyi Wafer Factory will implement high-tech manufacturing processes and free up the 28-nanometer process to build second-generation 28-nanometer smart home, fast-charging chips, and power management chips.
In the previous life, most smart home chips before 2017 were 55 nanometers, and 40 nanometers were already here.
Only after 2017 did new 28-nanometer products gradually appear.
As for 10nm and 7nm smart home chips, this is unrealistic and purely wasteful.
After all, 10nm and 7nm mobile phone chips can sell for more than 1,000.
But the unit price of smart home chips is only a few dozen yuan, which is not worth the high-end manufacturing process.
Williams smiled and said: "Yes, we can mass-produce more at one time, and the cost can be spread to the lowest level! After trial production, formal production will begin. Chairman, how many orders should we place for those three chips?"
Wang Yi fell into deep thought: "How many orders to place is an important question! The more orders you place, the lower the OEM fee and the lower the cost. Of course, it cannot be too many to prevent excessive inventory."
"That's right." Williams handed over a document: "Chairman, this is TSMC's quotation and our calculated costs. Take a look."
Wang Yi took it and looked at it carefully.
Williams did a good job, not only marking the foundry fees, material fees, and diluted total costs corresponding to different outputs, but also the current purchase price of similar chips.
This made Wang Yi very satisfied.
At present, the power management chips of Xingyi mobile phones and Xingyi tablets are purchased from Qualcomm, Nvidia and other companies.
Wang Yi purchases a large amount, and the prices offered by the other party are constantly decreasing.
For example, the two power management chips of the Xphone 1 are from Qualcomm. The two purchased by Xiaomi cost US$10 in total, and the two purchased by Wang Yi cost US$7.
Later, with the increase in orders for xphone 1, the price dropped to 6 US dollars, which is equivalent to only 3 US dollars each, which is 20 yuan, which is considered a bargain price.
But Qualcomm is still making money!
Without him, the material cost of a power management chip is only a few cents!
For chips, the material cost is very low, and the OEM fee for mass production is also very low. The biggest cost is the research and development cost.
Qualcomm has a large output, so the R&D costs are evenly distributed. Add in TSMC’s foundry fees, and the total cost of a power management chip is as much as one dollar!
Therefore, whether Qualcomm gives Wang Yi 20 yuan, 30 yuan, or Xiaomi 35 yuan, Qualcomm will make a huge profit.
If you can't make it yourself, you'll have to spend an extra twenty or thirty yuan to buy it.
There is nothing we can do about it.
Twenty or thirty million for one mobile phone, ten million mobile phones, that’s two to three hundred million!
Xphone shipments exceeded 15 million. With two power management chips in one phone, Qualcomm earned 600-700 million yuan.
Now that Wang Yi has achieved self-development of power management chips and fast charging chips, he can also drive down prices.
The material cost provided by TSMC is also terrifyingly low, only a few cents!
As for George and his team, they didn't spend much on R&D for developing these chips. The bulk of it was the cost of tape-out.
Then there are the OEM fees and production line costs.
All these expenses are added up to the total cost. The more produced, the lower the diluted cost.
As Williams calculated, producing one million P1 chips would cost 30 yuan per P1 chip, which is 50% more expensive than buying Qualcomm's power management chip!
It can produce 10 million P1 chips. When diluted, one P1 chip only costs 5 yuan, which is only a quarter of the lowest price offered by Qualcomm to Wang Yi.
If 100 million P1 chips are directly produced, the cost of one P1 chip will be only 1 yuan.
That's right, it's that cheap.
After all, it only costs a few cents.
This is the power of self-research.
Take Beidou terminal chips, for example, we couldn't make them at first, and foreign countries sold them for 1,000 yuan each.
When we mass-produced the first generation of self-developed terminal chips, foreign companies lowered the price to 500 yuan a piece.
After we mass-produced the second-generation self-developed chips, foreign companies lowered the price to 200 yuan again.
Then 100 yuan, 50 yuan...
After a few generations, domestic terminal chips caught up with foreign chips, and they lowered the price to 6 yuan.
That's right, the terminal chip that cost 1,000 yuan back then only costs 6 yuan, and we are still begged to buy it.
After all, their R&D costs have been diluted long ago, and now the total cost is only a few cents. Let alone selling it for six yuan, they can make a profit from selling it for one yuan!
The chip is such a special thing, the material cost is very low, the key depends on the R&D cost and output!
Similarly, the fast charging chip C1 and the smart home SOC chip H1 are both about the same price.
If 100 million pieces are mass-produced, the diluted cost per piece will only be 1 yuan!
Wang Yi really wanted to buy 100 million pills directly, but he had to calculate whether he could eat them.
Thinking of this, Wang Yi began to make careful calculations.
The demand for fast chip C1 and power management chip P1 depends on the subsequent sales of Xingyi mobile phones, Xingyi tablets, and Xingyi TVs.
In fact, the charging chip is also a type of power management chip.
The difference between the two is that the charging chip is only responsible for charging management.
The power management chip is responsible for the entire power management, such as cutting off part of the power supply under certain operations.
The simplest one is that when the screen is turned off, only the power control of the screen is cut off, but the power supply to other components of the phone remains. This is achieved through the power management IC.
As mobile phones are upgraded, more and more power management chips will be needed.
A 3G mobile phone only needs two power management chips, which is enough.
A 4G mobile phone requires 4-6 power management chips!
??????55.??????
A 5G mobile phone requires 8-10 power management chips!
The current Xphone and Xingyi Tablet X1 both use two Qualcomm power management chips.
The xphone pro and Xingyi Tablet X1 pro both use two power management chips from Nvidia.
But with the second generation of Xingyi, power management chips have to be added.
Even if the second generation Xingyi does not have 4G and is still a 3G mobile phone, it will need two power management chips P1, plus a fast charging chip C1.
If it is 4G, it will be 4-6 power management chips P1, plus a fast charging chip C1.
However, there is no need for 4G for mobile phones in 2012.
After all, domestic 4G began commercial use in 2014.
This year’s iPhone 5 is not a 4G phone either.
For nothing else, the Qualcomm baseband used in the previous iPhone 5 supports the 4G frequency band, but the Qualcomm transceiver used does not support 4G and cannot receive 4G signals.
So the iPhone 5 is just a 3G phone.
It won’t be until next year’s iPhone 5C and iPhone 5s that they will support 4G.
Similarly, there is no need to rush to 4G for Xingyi mobile phones.
Although 4G has already started in Europe and the United States, domestic 4G will not be available until 2014, and Xingyi mobile phones cannot go overseas. There is no rush to launch 4G in the second half of next year.
In this way, within the next year and a half, a Xingyi mobile phone will require two power management chips + one fast charging chip.
The same is true for Xingyi Tablet.
Calculating the time, the power management chip P1 was successfully trial-produced in March and officially mass-produced. It can be used on Xingyi mobile phones and Xingyi tablets as soon as April.
Since we already have self-developed chips, there is of course no need to continue to purchase power management chips from Qualcomm and Nvidia.
That's right, after using up the stocks of Qualcomm and Nvidia's 20 yuan power management chips, Xingyi mobile phones and Xingyi tablets have all been replaced with self-developed chips that cost 1 yuan each!
In the next year, xphone will continue to sell, focusing on the mid-range market. In the next year, the monthly sales will be 1.7 million, and the annual sales will be 20 million units.
Xphone Pro will continue to sell, and it won’t be a big problem to ship another 20 million units in the next year.
The Xingyi Tablet X1 is now on sale globally and will ship another 20 million units in the next year, so there is no pressure.
As the most powerful quad-core tablet, the Xingyi Tablet X1 pro is a hot seller around the world and is expected to ship 30 million units a year.
Including the second-generation Xingyi in the second half of the year, the sales volume is calculated as 30 million.
Adding everything up, Xingyi Technology will sell 200 million mobile phones or tablets in the next year, requiring a total of 400 million power management chips and 200 million fast charging chips!
In addition, Xingyi TVs and smart home products also require power management chips...
For example, AI speakers require power management chips, as do other products.
To put it bluntly, almost any smart home requires a power management chip.
In this way, the demand for power management chips will continue to increase!
Thinking of this, Wang Yi looked at Williams: "William, place an order for 300 million power management chips directly. In one year, let TSMC produce 300 million for us and drive the price down to death!"
"300 million?" Williams was startled: "Chairman, do we need so many?"
Wang Yi nodded: "300 million chips are enough for one year, not even one year. Next, mobile phones, tablets, smart homes, and TVs will all use our self-developed P1 chip."
"Okay, for a big order of 300 million, the cost can still be low. It is estimated that the diluted cost can reach 80 cents!"
"Very good!" Wang Yi beamed with joy: "We purchased 300 million power management chips for 8 cents each, which originally cost 20 yuan each, saving nearly 5.8 billion. This is the power of self-developed chips!"
"Yes, of course, the most important thing is that we have a large output, 300 million units. If there are only 3 million units, the cost will be more expensive than buying Qualcomm's!"
Williams laughed and said: "How much is the fast charging chip C1?"
Wang Yi thought for a moment and found that fast charging chips require less use. Only one device is needed.
Most smart home appliances are plugged in and do not require fast charging chips.
At most, the next generation of mobile power supplies requires fast charging chips.
In this case, 500 million is enough.
However, considering that the fast charging standard X-0 and the Type-c standard can be promoted in the future, Wang Yi decided to produce more.
"The fast charging chip C1 will be produced in 200 million pieces in one year."
"So much? Will it be useless?" Williams was a little worried.
Wang Yi waved his hand: "It's okay. When we launch the fast charging standard X-0, realize 10W fast charging, and fully open the fast charging era, how will our friends and businessmen respond?"
Williams' eyes shone: "Friends of business are trying to quickly follow up on fast charging, but Qualcomm and other companies can only provide 5 watts of slow charging, so if they want 10 watts, they can only buy our Xingyi Technology's fast charging chip C1!
"
Wang Yi nodded: "Yes, and our Xingyi Technology's fast charging chip C1 must be equipped with our power management chip P1 before it can be used."
Williams was overjoyed: "Friends can only use our fast charging solution and purchase our fast charging chip + power management chip. At that time, two power management chips P1 + one fast charging chip C1 will be sold in combination. Qualcomm sells 20
-35 yuan each, we sell it for 20-30 yuan each, and a set of three for 60 yuan, the cost is only three yuan, that’s a big profit!”
"Yes, that's the strategy." Wang Yi smiled slightly.
This is his overall strategy. After self-research, it is not only cheaper for personal use, but can also be sold in reverse and make a lot of money.
Once sold to the outside world, 200 million C1 chips and 300 million P1 chips will not be enough.
Need to increase production capacity!
It doesn't matter. The fast charging chip will be used alone in the early stage and will be sold to the outside world after a few months to increase production capacity.
When Qualcomm's fast-charging chips are developed next year, Xingyi Technology's fast-charging chips will be directly sold at a lower price to seize market share.
In the past, foreign companies were so disgusting with domestic companies, but now Wang Yi just wants to follow their example.
"Chairman, how much order volume should we place for the smart home chip H1?" Williams asked again.
Wang Yi thought for a while: "The annual sales volume of a single smart home model is only around 10 million, but as the number of models increases, the total quantity will increase."
Smart homes require many chips, the low-end ones include MCU chips, and the high-end ones are the SOC smart home chip H1 developed by Xingyi Technology.
It stands to reason that it is enough to use H1 for high-end products and purchase MCU chips from other companies for low-end products.
But what’s embarrassing is that low-end MCU chips from rival manufacturers cost a few dollars each, which is higher than the self-development cost of H1, a high-end smart home chip!
After all, people have to make money!
In this way, there is no need to purchase an MCU at all. It is better to directly use the self-developed high-end H1 chip, even if it is overqualified for use...
This is the difference between purchasing and self-research.
In the previous life, Huawei's self-developed processor broke through the blockade and directly lowered the price of the tablet to 2,099. The entire Android tablet camp collapsed!
There is no way, Huawei develops its own processor, so the cost is there.
Friendly merchants purchased Qualcomm processors and had to bleed a lot of money.
Thinking of this, Wang Yidao said: "Smart home chips will be produced at 100 million units per year at first, and will be adjusted later based on market conditions."
"good!"
Wang Yi thought about mobile phone chips: "As for mobile phone chips, Qualcomm's processor is called Snapdragon, and ours is called Kunpeng!"
The current naming method of Qualcomm processors is still very confusing, 8660, 8064...
But at the beginning of next year, they will all be adjusted, with low-end Snapdragon 200, mid-range Snapdragon 400, Snapdragon 600, and high-end Snapdragon 800.
The 28nm quad-core flagship processor APQ8064 released in the second half of this year is the Snapdragon 600 series.
Xingyi Technology's first processor is also 28 nanometers and also has a plug-in baseband.
Released half a year later than the 8064, it should have some advantages in performance, but it is no match for Qualcomm's next-generation Snapdragon 800.
After reading this, Wang Yidao said: "The first mobile phone chip is called Kunpeng 700!"
Qualcomm Snapdragon 200, Snapdragon 400, Snapdragon 600, Snapdragon 800...
Xingyi chips are Kunpeng 300, Kunpeng 500, Kunpeng 700, Kunpeng 900...
Taking Qualcomm's path will leave Qualcomm with no way to go, and it will still overwhelm Qualcomm.
"Kunpeng! This domineering force is even more domineering than Xiaolong!"
Williams was overjoyed.
Wang Yi smiled slightly: "We will make the Kunpeng 700 with an external baseband as soon as possible. The new phones early next year will be equipped with the Kunpeng 700. The Kunpeng 900 with an integrated baseband will be developed in the first half of next year. In the second half of the year, Qualcomm's integrated baseband chips will be launched. We
Kunpeng 900 with integrated baseband must also be launched!"
"Okay, chairman, don't worry, it's no problem."
Wang Yi continued: "As for the name of the plug-in baseband, the 3G baseband will be called Pterosaur 3000! When the subsequent 4G baseband is released, it will be called Pterosaur 4000!"
"Okay!" Williams remembered it firmly.
Wang Yi patted Williams on the shoulder: "Do a good job and look forward to the successful development of Kunpeng 700 and its official mass production! Come on."
"Don't worry, Director Wang, he will live up to his trust!" Williams assured.
The GPU of Kunpeng 700 uses ARM public model, which is not difficult.
The research and development of the CPU was also progressing smoothly with him and his team from Texas Instruments.
The key lies in the research and development of Wing Loong baseband 3000.
After all, Texas Instruments is not good at baseband. People at VMware have also made 55nm before, but now it is very difficult to go directly to 28nm.
But there is still more than half a year to go, and there is great hope!
Two days later, the trial production of three chips was successful!
Due to Wang Yi's large size, the total number of the three chips is 600 million, directly becoming TSMC's super customer.
Although they are all 40nm chips, they were still pretty good in 2012.
The high-end 28nm has just been put into production, and TSMC has limited production capacity.
Suddenly there were 600 million 40-nanometer chips. Zhang Zhongmou was so excited that he personally called Wang Yi:
"Director Wang, congratulations on the successful trial production of the three chips. We at TSMC will prioritize your order so that production will start tomorrow, March 16th. The price will be calculated for you as low as possible and the production capacity will be sufficient. Please rest assured.
.”
"No problem, thank you very much." Wang Yi said calmly.
But I secretly thought that the honeymoon with TSMC would not last long.
The Xingyi wafer fab will be completed by the end of the year and will be put into production next year. If 40nm mass production can be achieved, these three chips will be transferred to the Xingyi wafer fab for self-production and sales, with subsequent orders of hundreds of millions of pieces every year!
()
Chapter completed!