Chapter 843 New Blueprint for Semiconductors

After Meng Qian finished his speech, the conference entered into the display of products and technologies, and following Meng Qian on the main stage today was Liang Mengsong from Gale Semiconductor.

When Liang Mengsong walked onto the stage, he was holding a chip in his hand.

"In the past two days, it has been rumored that IBM will stop supplying ancient computer chips, and I saw many people asking what we should do.

This is our answer. "

Liang Mengsong picked up the chip, and a photo of the chip appeared on the big screen, along with the corresponding note: the world's first 8nm computer CPU.

There was a commotion at the scene, and Liang Mengsong said calmly, "Our 8nm processor uses the latest EUV lithography technology."

A very simple sentence, but it expresses many things clearly.

The launch of 8nm CPU must first be a performance of leading technology. After all, several other companies have not yet broken through the 10nm barrier.

Secondly, because Liang Mengsong confirmed the use of EUV lithography technology, it shows the reliability of EUV lithography technology, and once the practicality of EUV lithography technology breaks through the initial bottleneck, it will go smoother and smoother, at least to 5nm It will be smoother.

Sure enough, Liang Mengsong said with certainty that the company will mainly promote 8nm processors this year, and will mass-produce 7nm processors next year, and mass-produce 5nm processors before the end of 2019.

The advantages of the internal industry chain have once again been demonstrated, and the leading position of Dafeng Semiconductor has become more stable in the battle of craftsmanship.

It is indeed appropriate to use this as a response to the chip supply cutoff. Although the top of the CPU ladder diagram still does not have the name of Dafeng Group, the leading technology can make up for the gap in other aspects to a certain extent.

Unless Intel and IBM can quickly catch up with Gale Group's technological level, but it is unlikely now.

However, when everyone thought that Dafeng Group had solved the problem of cutting off the supply of high-end computer chips, Liang Mengsong let everyone realize that Dafeng Group was not just to solve the problem.

Because after introducing the latest processor, Liang Mengsong continued, "At today's site, in addition to the world's first mass-produced 8nm processor, we will also show our latest semiconductor development plan.

As I said just now, we will achieve mass production of 7nm at the beginning of next year and mass production of 5nm by the end of 2019. The actual R&D and production is about half a year ahead of our blueprint.

This is the result of everyone's joint efforts, and now in 2017, our semiconductor development blueprint will go one step further.

After 2019, our goal is to be below 5nm, and what makes us excited now is that just last month, we made a huge breakthrough in our 3nm research and development, and we have achieved 3nm testing in the laboratory, and it is expected that in 2021 3nm mass production can be achieved. "

The whole scene exploded, including the entire semiconductor market.

The 3nm laboratory test was completed in 2017, and it is determined to be mass-produced in 2021. This is the rhythm of making Intel unable to see the taillights of the car.

So it is inevitable that some people will have doubts, and Liang Mengsong's next introduction to 3nm technology broke the doubts of many people.

Of course, the Dafeng Group realized the 3nm experiment in 2017, and Meng Qian's contribution is naturally indispensable.

The Belgian Microelectronics Research Center was the first to realize 3nm laboratory testing in history in 2018, and Gale Group was a year earlier.

The most important thing Meng Qian did was to give everyone a firm direction, which saved a lot of time for detours.

The first time the chip was considered by the public to have reached its limit was when it was 22nm. At that time, the breakthrough of this limit was due to the appearance of the 3D transistor structure, which was when Gale Semiconductor became famous in the semiconductor market at one stroke.

The second limit of the chip that the public thinks is 5nm. A tunneling effect is known to almost everyone, so that in 2016, it appeared that as long as a chip can break through 5nm on the Internet, it will be used by a group of people. The tunneling effect is the case, although most people here probably don't know what the tunneling effect is.

It wasn’t until TSMC confirmed the 3nm breakthrough that those people finally shut down. In fact, the tunneling effect is true, whether it is the limit of 22nm or the limit of 5nm, the premise is that the structure remains unchanged.

The breakthrough of 22nm depends on the breakthrough of the structure, and the breakthrough of 5nm also depends on the breakthrough of the structure. In a sense, 3nm actually refers to the limit that can be achieved when the transistor density is equal to the line width of 3nm, and the equivalent result of 3nm is realized. Not really a 3nm linewidth.

This was the case with TSMC in the past, and it is the same with Gale Semiconductor now.

But because it can achieve the same effect, it is called a 3nm processor, and there is nothing wrong with it.

So when the entire industry was struggling with how to make a breakthrough at 22nm, Meng Qian firmly asked his employees to work on 3D transistors.

In the same way, when the industry is discussing whether to use new materials after 5nm, Meng Qian told employees that material research and development must be done, so experiments including nanosheets, nanowires, high-mobility channels, and carbon nanotubes are all underway. .

At the same time, continuing to break through the structure is another must-go path.

Because of firm investment and firm research and development, we can naturally go faster than others.

Liang Mengsong proposed a brand new concept, 3D composite structure.

Meng Qian didn’t see TSMC’s 3nm detailed explanation because he died before 2020, but Dafeng Semiconductor’s 3nm route is quite similar to TSMC’s 3nm route. This may be the inevitable trend of Dafeng Semiconductor’s adherence to the FinFET transistor route.

The so-called 3D composite structure is actually a comprehensive packaging technology to make room for the already crowded space.

Of course, it's really easy to say, but very difficult to implement.

In the final display of Dafeng Semiconductor's new blueprint, Liang Mengsong told everyone that the 3D composite structure can advance the chip process to 1nm, and Dafeng Group's plan is to achieve mass production of 1nm in 2025.

At this point, everyone feels that Dafeng Semiconductor's role should end today, because it is really shocking enough, but the fact is that it is not over yet.

"In order to further protect our next semiconductor development blueprint, I would like to invite another colleague of ours to the stage, and he is Winick."

Winick?

This is another shocking picture, the former CEO of ASML suddenly appeared on the stage of Dafeng Global Developers.

What shocked everyone even more is that Winick is now the director of Shanghai Microelectronics Technology, and he claims to be in charge of the research and development of Shanghai Microelectronics' next-generation EUV lithography machine.

The reason Winick came to Shanghai Microelectronics was because, as Meng Qian predicted, Intel cut off the European team after acquiring ASML.

This is the usual acquisition routine used by American companies. Meng Qian is no longer familiar with it. The axed teams of ASML have nowhere to go, and finally all came to Shanghai Microelectronics.

In order to ensure the process development after 5nm, the lithography machine really needs to go one step further.

That is, the second-generation EUV lithography machine, the biggest change is the high numerical aperture lens. By improving the lens specification, the two core indicators of the new generation lithography machine, the microscopic resolution and the registration accuracy, have been improved by at least 70%.

The reason why Winick chose to talk about this is to tell the world that the old team that is the real core of ASML is now in the Gale Group, and there is no need to count on the ASML after Intel's acquisition.

From 3D composite structures to next-generation lithography machines, when other companies are still struggling with how to break through the 10nm process, Gale Group's new semiconductor blueprint has already pointed to the 1nm process.

As long as Dafeng Group has the ability to complete its own blueprint, Dafeng Group is likely to become the absolute king of the post-5nm era in the semiconductor field.

The pattern of global semiconductors that only the United States is the leader will be completely changed.

As soon as the news came out, semiconductor companies in Europe, Neon Country, and Korea Country, especially chip design companies, rushed to the gate of Dafeng Group and formed a long queue.

At the same time, amidst the curiosity of the people who eat melons, the media broke the news that the US chip company has also come.

Everyone is wondering how the United States will react to the Dafeng Group now, whether it will be generous or just retaliate.

In the end, Dafeng Group responded: Considering the uncontrollable risk of cooperating with American companies, we need to do an in-depth assessment before considering whether to cooperate with American companies.

Dafeng Group did not retaliate directly, nor did it show generosity, but kicked the ball back. It is okay for American companies to cooperate with Dafeng Group, but you must show your attitude first.

But the problem is that the current American companies are a little bit involuntary.

How things develop after that can't be blamed on Dafeng Group.