Conexant spun off in 1999 from Rockwell, which at the time held a dominant
position in the voice and fax/modem chip market. Now Conexant looks as though it’s
going back to its voice heritage. The company is keen on
carving out a niche for its voice processing technology.
Consumer Electronics Show earlier this year, Conexant showed off its
proprietary far-field voice input processor SoC, which works on a voice
trigger, to turn a TV on or off even in a noisy room. The same voice
processing chip can be also used with VoIP applications such as Skype
TV, so that a user can talk to another person on Skype even while the TV
is blasting in high volume. Keys to such solutions are Conexant’s
far-field voice processing algorithms (acoustic echo cancellation, noise
reduction, beam forming as well as pre- and post-processing) and its
high performance 24-bit A-to-D converter.
Beyond voice processing, Conexant will focus on printer SoCs, headsets for gaming machines and video surveillance.
Chittipeddi noted that Conexant today has a solid customer base including Plantronics and Logitech.
also important to note that Conexant is avoiding the already crowded
market for consumer video products and set-tops, where competition is
intense and often dominated by Broadcom. The company’s new commitment is
to well-differentiated niche segments.
With its more focused
portfolio, Conexant is confident of recovery and managing its growth
more reasonably. Free of debt payments, “we can generate cash that we
can use for investing in our business, rather than for paying interest,”
What went wrong
There are a million reasons why and how things went wrong for Conexant. However, betting on the Internet boom and building its business on “networking” wasn’t one of them.
problem for Conexant in the late 1990’s was that the company’s modem
chips were built for the analog era. In order to compensate for this
weakness, the company went whole hog acquiring technologies that looked
useful for building the broadband future of the Internet. Of the nearly $2
billion Conexant spent during its shopping spree, the most
stunning was the acquisition of Maker Communications, a Massachusetts-based
software company whose chips worked with light-based technology. Maker
Communications had all of $13.6 million in annual sales at the time it
sold itself for nearly $1 billion in Conexant stock, according to
As all this happened, Conexant underestimated
the constant need for large capital that had to be fed to the companies
it had acquired. In 1999 and 2000, Conexant ended up issuing two
tranches of debt, totaling $1 billion.
As an ex-Conexant in the early 2000,I agree with most of the analysis. Broadcom and Conexant are neighbors in Newport and many Broadcom employees are from Conexant. The only major difference is the vision of the management team. As far as I can remember, Conexant's CEO Dwight Decker was a nice guy, and we didn't have a strong CTO (don't even remember the name, or did we have one?). I was told Broadcom's CTO Henry Samueli was a visionary, so that might be the difference (in addition to the M&A strategy).
You have to look beyond all the rosy announcements and blames from CNXT management. Thanks to financial arrangements, all the money at Conexant is long gone. CNXT small investors and CNXT employees are the only victims to see their hopes vaporized before their eyes.
I have fond memories of working with some great Conexant engineers on a JV back in the late 90s/early 00s. It was a shame to see the company continue to shrink to such a level, much as my former employer did -- not only through market changes, but also due to simple divestiture. Those divestments may have brought in lots of cash and unleashed shareholder value, but when taken to extreme, the remaining so-called core business is often not sustainable.
All Conexant CEOs and presidents/co-presidents were responsible for the plight of the company. As a Rockwell spinoff, it had excellent product portfolio IP and product portfolio and a very solid engineering team with excellent DSP/analog technology. Dwight Decker, Dan Artusi, Scott Mercer, Sailesh Chittipeddi and Christian Scherp let the company down by making financial and strategic blunders. None of these CEOs very visionaries or had any technical depth or understood the markets well. Sad.
Wonderful analysis in a very short amount of time!
Sound like the Maker acquisition was one of the biggest blunders, coming when many people were trying to jump ahead of "the long boom" Wired and others predicted with bets in optical communications that went south in the dotcom bust.
Actually, this was a canny strategic move by the CEO. Kudos. This company has been hobbled with excessive legacy debt for years, and a pre-pkg BK means they can finally move forward... with the help of a golden angel here or there. Not predicting miracles, but don't be shocked if they move to go public again in the intermediate term.
Conexant had a choice to raise convertible debt or issue shares when they hit $12B in mkt cap. PMCS, AMCC etc issued shares and diluted and stock went down a bit as it was bubble time anyway. Conexant CFO decided to issue Convertible Debt and management cashed out of the stovk. Most debt did not convert and after a decade they stil have 100s of millions left to pay. Other comments in this article are spot on. How a great technology company can be mis-managed to bankruptcy, they need to learn from erstwhile Conexant senior leadership.
Having worked at the early days in both companies, I can tell you the two mgmt teams were miles apart. the Henry's were brash, disruptive, and forward looking, CNXT (names to remain nameless) were conservative and content with a thriving analog modem business which unknown to their near sightedness, looked like it was going up into the clouds, but actually was about to fall off a cliff.
Broadcom certainly benefitted by hiring away the best and brightest from Conexant, perhaps getting design, process, and customer ideas in the process.
The telecom bubble (not the dot-com bubble) was the big disaster. Stock dropped at Nortel from 100 to 1 dollar, with big 3x gains from $1 to $3 for the bottom fishers. That magnitude certainly highlighted the problem with the entire broadband vision.
1. Fiber Optic drivers and cable were put in place with great speed, giving 8x and then 4x again speed improvements, for the backbone systems from city to city, country to country. That was a 15 year glut in capacity, as nobody was working on "the last mile" at that time.
2. Instead of a 2x jump every 2 years roughly, like Moores law in chip world, these guys thought that they could create 8x improvement immediately, and then 4x more every few years, playing color games in fiber...which they did, as telecoms and cable people competed to get to each city, then each zone, and "some day" link up to my house and your house.
3. So all the old Conexant CEO's and VP's can blame the telecom industry for creating a huge short term bubble, and great opportunity for those that saw it coming. See Qualcomm who had cellular as well as internet IP in place, to weather the storm and cash in on the convergence.
5. What next for Conexant? The food is good in Newport, as is the sailing, so they should be able to hire a top notch CTO and CEO and CFO team that may take advantage of the lower debt load now, and maybe they can hire back a few of the best and the brightest and maybe change the brand name?
I'm not so sure the voice processor will be in high demand given today's processors are so powerful. Considering power efficiency, 1 less chip will definitely reduce overall power consumption.
The future of a company is typically relying on the vision of the management team. The vision may turn into gold. Unfortunately, if the team fails, the workers suffer. It is unfortunate that Conexant has to file Chapter 11. I do hope the team successfully revives the company.
Conexant had the best engineers, advanced products, and strong financial lineup.
"Forgive but never fortget their name." John F. Kennedy.
Dwight Decker,Dan Artusi, Scott Mercer, Christian Scherp, Sailesh Chittipeddi and its cronies effectively wiped out everything with their financial expertise.
Chapter 11 is the best solution for Conexant and put it out of misery. CNXT shareholders and its employees (ex and current) are the only victims, gone with their saving, trust, and retirement. RIP.
I always find it odd when companies involved in what should be a thriving business area, end up struggling or failing. But then again, back in the '90s and '00s, wasn't it obvious to everyone that this ever-important "growth" was primarily achieved through acquisitions? How can any sensible person get excited about that kind of growth?
Conexant spun off from my company, and our group was sold off to another company. Many top notch engineers at Conexant. I have to believe the analysis in the article is correct. Too much of this "growth by acquisition," as opposed to growth by creating new market demand, and an underestimation of the cash suction those acquisitions would create.
Lessons one wishes the politicians and government bureaucrats would also take to heart. Staggering and growing debt will ultimately be the undoing of any organization, including a government.
There was huge mismanagement at the CEO level at this company that has left a lot of people lose their jobs. Not a single CEO had the right technical or marketing background to steer the company and so the downward spiral precipitated by financial mismanagement. The current CEO was a "get lucky person" with an operations background who has neither a marketing or technical background to retarget it to new markets. It still has very good engineers who stuck with the company. It needs a good CEO. It has the DNA to execute well on the engineering front.
Here's an example of what sort of problems Conexant has, and why it's so difficult to work with them. Great FAE - absolutely fantastic FAE. Great support, wonderful products.
Then comes the reality: Order a dev kit. No inventory. Order samples. No inventory. Order parts. No call back from the Distributor, nuHorizons. Period.
Our FY2013 for the CX93610 is 25,000 units.
Our FY2014 for the CX93610 is ~240,000 units.
And possibly higher.
So after designing their part into our product, we can't get the parts, and we need to go in another direction. Such a bitter pill to swallow.
That is Conexant.
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.