Reborn in 2010, the AI ​​in my mind is too powerful.

In 1986, Art DeGias was just an engineer at GE Research Center, leading a team working on a technology called logic synthesis.

This technology allows people to describe chip circuits with code, and then the machine automatically generates a gate-level netlist, eliminating the need for engineers to draw every connection by hand. However, that year, General Electric made a decision to completely withdraw from the semiconductor business, shutting down the entire semiconductor division.

Arter convinced GE's management to let him go solo with his technology patents and $400,000 in start-up capital.

This led to Synopsys becoming the dominant EDA technology company.

Art is indeed very talented. Taking advantage of the rapid development of the semiconductor industry, he successfully took his company public in just five and a half years.

For the next 25 years, Art never left his position as the top executive, and he single-handedly made logic synthesis tools the first and only software that all chip design companies would buy.

Then, starting with logic synthesis, they expanded through in-house development or acquisitions to include placement and routing, timing analysis, formal verification, and lithography correction, until Synopsys became the world's largest electronic design automation company.

He once famously said: Great ideas come from great teams.

But that was more than a decade ago. Since Synaptics became the industry leader, Artest has gradually become autocratic.

For example, in this battle with FlowEDA, he didn't consult his team at all.

He was unwilling to rely on others' authorization to survive in the future, so he firmly disagreed with making peace with FlowEDA.

Leveraging its prestige of over two decades, despite differing opinions within the company, Atlas ultimately failed to sway Art's mind. Atlas, in conjunction with Kaden, launched a comprehensive attack on FlowEDA, employing various lawsuits, patent harassment, and even initiating an antitrust investigation.

But none of these ridiculous methods worked; the completely new framework brought about technological dominance.

The departure of Kaiden dealt a huge blow to Synopsys. Although they had always disliked Kaiden's forward-looking corporate culture, they now found it hard to accept losing Kaiden and fighting alone.

After that, TSMC urged them very urgently to upgrade their EDA, and even offered to make a large-scale investment if they were short of funds, just to match the FlowEDA technology.

But after more than two months, they have made virtually no progress. TSMC has also lost patience and informed Synaptics that it will switch to FlowEDA.

Synopsys' board, in desperation, began suggesting acquiring Mentor to offset the market crisis and give remaining smaller customers some confidence. However, Artest hesitated. After all, Mentor's products were very similar to Synopsys', and they couldn't complement each other.

Moreover, Icahn's presence made management less inclined to acquire Ming Director.

Synaptics can only send people to constantly monitor Mingdao's actions, which is very common in the EDA industry. After all, this industry is very small, and practitioners usually have acquaintances in other companies.

Just send a few employees to have dinner with Mingdao's classmates, and they can find out everything. After all, Mingdao is on the verge of collapse, and these people need to be prepared to switch jobs.

In the ensuing stalemate, Icahn suddenly and decisively liquidated his entire stake in Mingdao.

Synopsys' management is once again discussing whether to acquire Mentor, a company they see as a liability.

A message reached their ears: although they hadn't found out the specific actions of Mingdao's senior management, all of Mingdao's legal staff had recently been on business trips.

Synaptics executives immediately sent people to stake out Intel's Wind River facility, but they were unsuccessful.

This plunged them into great panic, because if Ming wanted to find a way out, he only had two options.

One option was Wind River, which was sold to Intel, where they could use Mentor's embedded channels, but it has now been proven that they did not take this route.

The other option is to slap themselves to FlowEDA, whether through acquisition or licensing, fully embracing deepflow is unacceptable to Synaptics.

At the second regular meeting in March, Art had to call all senior executives, including technical staff, to discuss Synaptics' future.

There were more than a dozen people sitting in the conference room, and none of them looked too good.

The CFO spoke first: "Let me say this: at the current rate of attrition, the annual contracts for core EDA products will be expiring in the spring of 2013, and 75% of them will not be renewed. This is the current situation, and there is a possibility that it will continue to worsen."

By the end of fiscal year 2015, our EDA revenue will drop from the current 900 million to below 200 million. This means that the company's main profits will then depend on the interface IP library.

If Mindray is acquired, that's not the worst thing. FlowEDA will have a hard time absorbing the teams of two large companies in the short term.

But if they license their technology to Mingdao, that would be problematic. Mingdao's technology would then surpass ours, and the price would likely put significant pressure on us.

After the CFO finished speaking, the General Counsel Robert added, "Legally, we have nothing left to blame on FlowEDA."

Art asked, somewhat unwillingly, "They authorized Mingdao; can't we say they're abusing their market dominance?"

Robert shook his head and said, "Don't Google legal stuff; most of it is misinformation. U.S. law fully supports free competition, and it's explicitly stated that companies are allowed not to sell technology to competitors."

Deepflow is still open source. Although the fact that the core is not open source makes us technically weak, it is enough to ensure that we cannot manipulate the market for monopoly.

The only slight chance is if we obtain the license, but FlowEDA, for its own monopoly interests, terminates the contract; only then would it constitute a monopoly, and there would be some room for dispute.

Art sighed and turned to look at the ZK work group sitting at the far edge of the table.

The working group was initially called GK, which means Grey Killer. Later, as FlowEDA's output increased, the working group's name was changed to ZOO Killer, which was a way of finding humor in hardship.

The head of the ZK working group said with his head down, "Simply put, we estimate that in three years, we may be able to catch up with FlowEDA's current technology."

Art remained silent for nearly a minute, looked around, and finally came to a conclusion: "Then there's no other way. The legal team should stay, and the others should leave first. Although it's hard to accept, Synaptics may really have no other way out technically."

After everyone else left, General Counsel Robert said in a low voice, "What do you want to do?"

"Let's do some preparation. First, formally inquire about the licensing. Let's see if we can get a joint R&D designation; we can send someone to help him improve the toolchain."

At the very least, we need a customized API layer, which will at least save us some trouble. If appropriate, we can trade that for the benefits of an IP library (reusable chip design modules).

Robert was shocked and immediately retorted, "No, although it doesn't earn as much as EDA, IP libraries are a business that makes money effortlessly. They earn money every year based on the number of chips produced, even if we are completely kicked out of the EDA market."

As long as we have the IP database, we have enough cash flow every year. One day, we can get back on our feet. The IP database is our foundation; it cannot be traded.

Art stared at the ceiling, utterly dejected, and said, "Trust my technical expertise. Even twenty-five years of accumulated experience might not be enough to withstand the attack of a genius."

"Before I became a lawyer, I was also in the tech field, so it's not like I don't understand it. The IP library contains physical layer and analog interfaces, which is not the digital layer that Chen Yiran is good at. Our IP library includes almost all interfaces such as USB, PCIe, DDR, and HDMI."

Any company that needs it can purchase design modules directly from us and embed them directly into their chip designs. I don't believe they can catch up with us that quickly.

Seeing his friend's impatience, Art patiently explained: "Your technical skills have always been mediocre. You have a problem understanding IP databases. The technical barrier to IP databases is the accumulation of experience, authentication, and trust."

Chen Yiran only needs to use the accelerated framework of deep learning to make predictions, perform extreme value tests, and help clients verify their findings; the rest is just a matter of time.

Robert had no choice but to agree: "Fine, I'll do as you say."

Art nodded, then leaned close to Robert's ear and said, "You not only have to listen to me, you also have to think of a good solution. Of course, we have to do what I just said, but there's another thing we have to do too. I'm afraid Chen Yiran won't talk to us. He might prefer to see us die, leaving Mingdao to do the dirty work."

So we need to think of another way. You should find some people to investigate their employees in detail, especially the first dozen or so. Look at their family situations and see if there are any we can bribe.

If things go badly, we'll have to resort to sabotage. The ZK team needs three years; we need to find a way to buy them that time.