Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/threads/the-day-huawei%E2%80%99s-ceo-bent-the-knee-in-a-santa-clara-marriott.23782/
        )

    [addOns] => Array
        (
            [DL6/MLTP] => 13
            [Hampel/TimeZoneDebug] => 1000070
            [SV/ChangePostDate] => 2010200
            [SemiWiki/Newsletter] => 1000010
            [SemiWiki/WPMenu] => 1000010
            [SemiWiki/XPressExtend] => 1000010
            [ThemeHouse/XLink] => 1000970
            [ThemeHouse/XPress] => 1010570
            [XF] => 2030770
            [XFI] => 1060170
        )

    [wordpress] => /var/www/html
)

The Day Huawei’s CEO Bent the Knee in a Santa Clara Marriott

Daniel Nenni

Admin
Staff member
Huawei vs PMC.jpg



View Haresh Patel’s  graphic link

Haresh PatelHaresh Patel
https://hareshpatel.ai/
It was 1998, and I was VP of Worldwide Sales at PMC-Sierra during the dot-com boom. We were Wall Street’s darling, building the semiconductors that powered the internet’s backbone. Our chips were the brains behind the routers connecting the world.
But there was one market we refused to touch: China.

“They’ll steal everything we’ve built.” The executive team feared IP theft and reverse engineering. But I watched Huawei and ZTE grow at a blistering pace. I knew we were missing the opportunity of a lifetime.

After months of fighting, I convinced Bob, our CEO, to let me build a China sales team.

We penetrated deep into both Huawei and ZTE. Our semiconductors became the brain of their routers. While the old guard designed technology with thousands of engineers over years, Huawei and Cisco used our chips with hundreds of engineers and brought routers to market in months. They moved at internet speed while dinosaurs sat in committee meetings.

Revenue was pouring in.

Then the whispers started: Huawei was reverse engineering our technology.

Bob pulled me aside. “Find out what’s happening.”

What we uncovered wasn’t courtroom-ready, but damning enough. We decided fast.

I drafted a letter to Huawei’s CEO: We were stopping all chip sales immediately. Since we were their sole source, we’d shut down their growth overnight. No chips, no routers. No routers, no revenue.

The response came within hours.

Their CEO personally flew to California for an emergency meeting.

We met at the Santa Clara Marriott. Neutral ground where billion-dollar relationships get made or broken.

Myself, Bob, my sales manager Tony, and Huawei’s CEO sat across a conference table. His opening was textbook: chips were too expensive, no motivation to steal. We both knew the stakes. They needed us far more than we needed them.

We hammered out a minor price concession. They signed a memorandum immediately ceasing any reverse engineering.
No lawyers. No legal review. No weeks of back-and-forth.

We drafted it in that conference room. Both parties signed, shook hands, walked out. One afternoon. Old-school dealmaking.
We continued selling tens of millions for several years. Eventually, they replaced our technology—inevitable. The question was whether we’d capture the growth phase or watch from the sidelines.

But here’s the real victory: Nortel, Ericsson, Nokia—getting clobbered by Huawei and Cisco’s speed—finally abandoned their “not invented here” mentality and started using our chips.
We won both sides of the war.

The lesson wasn’t about IP protection or China strategy. It was about leverage, timing, and courage to make bold moves when you hold the cards. We could have lawyered up and spent years in litigation. Instead, we sat their CEO in a Marriott and reminded him we were the sole source powering his rise.

Sometimes the most powerful negotiation happens when you’re willing to walk away entirely, and making someone fly 6,000 miles.
 
WT...

What was the moral of the story again? Was it that if you are the sole source, you have power and can bend or break knees?

Anyway, it is good to hear personal triumphs, thank you Daniel for this very uplifting article!
 
  • Like
Reactions: VCT
In software you could tell when the derivative code had the same bugs as the original. With chips, it doesn't seem practical to actually steal the logic.
While the latter is true with most of today's chips, old hands here recall the days of painstaking layer removal and photomicrography (?). I could still see that happening for some large featured designs (analog?)
 
While the latter is true with most of today's chips, old hands here recall the days of painstaking layer removal and photomicrography (?). I could still see that happening for some large featured designs (analog?)
Yeah, I know. I wasn't into chip design back then, but I heard about looking at circuits through microscopes. I was referring to just modern (post 2010) logic chips. :)
 
WT...

What was the moral of the story again? Was it that if you are the sole source, you have power and can bend or break knees?

Anyway, it is good to hear personal triumphs, thank you Daniel for this very uplifting article!

This story played out many times in the semiconductor ecosystem. I know Haresh. Everyone feared China was copying, reverse engineering, general thievery and they were. For EDA it was license hacking and copying. The joke was selling one license to China was like selling unlimited licenses.

EDA worked together with the Chinese government to limit copying and the result was billions of dollars in EDA revenue over the last 20 years. Of course that was all negated by the US government banning EDA sales to China early this year so many more billions of dollars of revenue will be lost.

The moral of the story is that open markets and innovation is the best business strategy. What we are doing today, not so much.
 
This story played out many times in the semiconductor ecosystem. I know Haresh. Everyone feared China was copying, reverse engineering, general thievery and they were. For EDA it was license hacking and copying. The joke was selling one license to China was like selling unlimited licenses.

EDA worked together with the Chinese government to limit copying and the result was billions of dollars in EDA revenue over the last 20 years. Of course that was all negated by the US government banning EDA sales to China early this year so many more billions of dollars of revenue will be lost.

The moral of the story is that open markets and innovation is the best business strategy. What we are doing today not so much.
Chinese government only more seriously looking into IP issue when they are trying to get internal industry going. This is probably not unique to China
 
Chinese government only more seriously looking into IP issue when they are trying to get internal industry going. This is probably not unique to China

It is not unique to China. I ran foundry relations for an SRAM IP company for a few years. Foundries in Korea and China did try and replicate our memory compilers but were consistently 2-3 generations behind us which is why I said innovate. If you are running at full speed while looking over your shoulder you will hit a brick wall at some point in time.

This is one of the reasons why TSMC has such a strong IP ecosystem. They port/license the latest and greatest technology through close partnerships rather than steal IP. It is a winning formula and a big difference between China and Taiwan business culture. Just my personal experience of course.
 
I know the guys who gave China DDR3 for free, designed it on a PowerChip DRAM foundry process as contractors for ISSI on servers in Asia without taking any money up front, desperate times. One day close to tapeout they couldn't VPN in, and when they pinged IT they were referred to an American law firm which was more than happy to enter into a protracted legal battle that they knew these 6 engineers couldn't afford. Anyone who whines about China surpassing America in every facet of national development is absolutely clueless, we did this to ourselves.
 
bring back memory of Avanti and cadence ...

I was going to say the same thing! I worked for Avanti at the time. That was an interesting journey. I spent time with Gerry Hsu, he was quite the character!

According to Grok:

Key Events and Timeline
  • 1991: Four Cadence employees, including future Avanti founders, left to start Avanti (initially called ArcSys). Prosecutors later alleged that one founder stole Cadence source code during this transition.
  • 1994: Gerald Hsu, a senior Cadence executive, joined Avanti as CEO. Cadence filed an initial lawsuit against Hsu and Avanti for potential trade secret disclosure, leading to a settlement agreement and mutual release of claims.
  • 1995: Cadence discovered "digital fingerprints" (e.g., identical bugs and code structures) in Avanti's products, indicating widespread copying. Cadence filed a federal civil suit for copyright infringement and trade secret theft. Authorities raided Avanti's offices, uncovering Cadence code on their systems.
  • 1996–2000: The case proceeded amid appeals, including a 1997 Ninth Circuit ruling partially upholding a preliminary injunction against Avanti's use of the stolen code. A related California Supreme Court decision (2002) addressed whether the 1994 settlement barred claims, ruling it did not apply to undiscovered thefts from 1991.
  • 2001: Avanti and seven executives (including Hsu) pleaded no contest to criminal charges of trade secret theft, conspiracy, receiving stolen property, and securities fraud. Avanti was ordered to pay Cadence $195 million in restitution, plus $35–40 million in state fines. Executives faced jail time (e.g., Hsu served 4 months).
  • 2002: Synopsys acquired Avanti amid the ongoing civil suit. The parties settled for an additional $265 million from Synopsys (via insurance) to Cadence, plus reciprocal IP licenses. Total payouts exceeded $460 million, excluding legal fees (~$55 million for Avanti).
 
I was going to say the same thing! I worked for Avanti at the time. That was an interesting journey. I spent time with Gerry Hsu, he was quite the character!

According to Grok:

Key Events and Timeline
  • 1991: Four Cadence employees, including future Avanti founders, left to start Avanti (initially called ArcSys). Prosecutors later alleged that one founder stole Cadence source code during this transition.
  • 1994: Gerald Hsu, a senior Cadence executive, joined Avanti as CEO. Cadence filed an initial lawsuit against Hsu and Avanti for potential trade secret disclosure, leading to a settlement agreement and mutual release of claims.
  • 1995: Cadence discovered "digital fingerprints" (e.g., identical bugs and code structures) in Avanti's products, indicating widespread copying. Cadence filed a federal civil suit for copyright infringement and trade secret theft. Authorities raided Avanti's offices, uncovering Cadence code on their systems.
  • 1996–2000: The case proceeded amid appeals, including a 1997 Ninth Circuit ruling partially upholding a preliminary injunction against Avanti's use of the stolen code. A related California Supreme Court decision (2002) addressed whether the 1994 settlement barred claims, ruling it did not apply to undiscovered thefts from 1991.
  • 2001: Avanti and seven executives (including Hsu) pleaded no contest to criminal charges of trade secret theft, conspiracy, receiving stolen property, and securities fraud. Avanti was ordered to pay Cadence $195 million in restitution, plus $35–40 million in state fines. Executives faced jail time (e.g., Hsu served 4 months).
  • 2002: Synopsys acquired Avanti amid the ongoing civil suit. The parties settled for an additional $265 million from Synopsys (via insurance) to Cadence, plus reciprocal IP licenses. Total payouts exceeded $460 million, excluding legal fees (~$55 million for Avanti).
Dan, lucky you. You have front row seat on the biggest drama of EDA industry.

Later on, heard from someone at Cypress that apparently was 1st to discover the similarities and reported the issue, which started the whole legal events. Surreal
 
Dan, lucky you. You have front row seat on the biggest drama of EDA industry.

Later on, heard from someone at Cypress that apparently was 1st to discover the similarities and reported the issue, which started the whole legal events. Surreal

True. There was a bug that caused a display error the same as a Cadence tool. Some of the information above is not correct. A consultant for Avanti and former Cadence engineer borrowed a small amount of DB code from Cadence, his name is Mitch Agusa. This did allow Avanti to get to market quicker but it was not pivotal technology. Mitch took a shortcut, he was lazy, it was not as sinister as it was described in the media.

The FBI raided the new Avanti HQ in Fremont like we were a group of terrorists. The Stock bottomed out and a lot of us lost our investments. If I remember correctly some Cadence insiders shorted the stock prior to the raid and got fined for it. Joe Costello lost his job as a result. He threatened customers with legal action if they used Avanti products and the Cadence BoD did not care for that.

Overall a major clusterfk. As a result I went to law school to be better informed when working for start-up companies. Specifically, I wanted to avoid all legal problems as they are a colossal waste of time and money, absolutely.
 
Back
Top