ESDA CEO Confab: It was a Dark and Stormy Night
April 13th, 2017 by Peggy Aycinena
Something eerie and inexplicable happened on Thursday evening, April 6th. Out of nowhere, an intense storm swept through the Bay Area, unannounced and without warning. The skies darkened, the winds howled, severe rain pelted the crowded, suddenly dangerous freeways, and hundreds of thousands lost power.
Meanwhile, exactly in the midst of the most violent part of this mysterious storm, the CEOs of the four most important companies within the ESD Alliance sat on stools in front of an audience assembled at Synopsys and chatted about this, that, and the other. Seemingly oblivious to the profound violence unleashing itself just outside the windows, they acted as if nothing was amiss.
Everything in the industry – and the world – was in order: Wonderful, with the data pointing continuously up and to the right, and everywhere ample evidence for a bullish, optimistic, and excited outlook on the future of EDA and IP.
No matter that Nature was having its way out there in the darkness, that the U.S. had bombed Syria the hour before their discussion began, that the drumbeat for answers about entanglements with Russia was quickening, or difficult conversations with the President of the PRC were underway that very day in Florida – the CEOs of Synopsys, Cadence, Siemens/Mentor Graphics and SoftBank/ARM sat relaxed and easy, basking in the evident vitality of the EDA and IP industries, and allowing themselves to be shepherded through a congenial confab of confident chit-chat by Ed Sperling of Semiconductor Engineering fame.
That fact that the vagaries of Nature never came into the conversation was not surprising; the fact the Mr. Sperling refused all opportunities to bring what he termed as “politics” into the conversation was quite the opposite. Surprising, that is.
Would EDA companies continue to operate on a business-as-usual manner in Russia? Would ARM be changed by Brexit, Article 50 having been invoked by Prime Minister May that very week? Would SoftBank investors, Saudi Arabia, Japan, China or Abu Dhabi, change ARM in any way? Would Mentor be influenced by its new German owners? Would Synopsys continue to operate as if free trade and globalization were still the order of the day? Would Cadence be influenced by the rising power of the PRC?
None of this was on the table at that April 6th CEO Confab of Collegiality and Good Will. And why was that?
Is it possible that the CEOs of publicly traded companies get paid to be positive-outlook kinds of guys? Is it possible that the CEOs of all companies, publicly traded or not, need to maintain a veneer of steady optimism and aura of control over the arc of history of their organizations?
Is it possible that the world will continue to turn, and commerce will continue to operate, no matter the vagaries of political firestorms or howling nor’easters that should be kicking the shit out of Gnarly New England, not Sacred Silicon Valley?
Is it possible that “if we ignore it, it will just go away” is the best way to conduct business in industries as completely/totally global as EDA and IP?
Who knows? These questions were never asked on April 6th.
What was discussed generated two categories of responses: the Good and the Difficult. But other than a stirring endorsement of China as a market and mentor, nobody discussed the geopolitical situation or the weather outside.
Until the very end, that is, when suddenly the panelists turned brutally candid – either in preparation for forging out into the raging storm to reach their cars in the dark, wind-whipped parking lot, or because representatives of the U.S. Military seated at the back of the room began to ask questions.
The Good …
In the category of happy talk – trains, planes, cars, IoT, complexity, global markets, and investment opportunities – all of the responses supported the thesis that everything’s up and to the right for the EDA and IP industries.
Synopsys’ Aart de Geus: The industry is going through a phase shift and rejuvenation. Also, the only thing better than something smart, is something even smarter, so let’s praise the advent of the IoT.
SoftBank/ARM’s Simon Segars: There’s been an uptick in optimism about the next 10 years in the semiconductor industry. And just as the automotive industry eventually benefited from regulation after 60 years, the Internet will, after only 25 years, quickly be regulated, making it more safe and secure.
Siemens/Mentor’s Walden C. Rhines: This is a really good time for the industry. People who never did design before – Google, Facebook and Amazon – are becoming big customers for EDA, along with all sorts of companies who find they need to design to differentiate. And although we’ve pretty much automated the whole process of IC design, the design of trains, planes and cars is still largely un-automated. This market will be enormous compared to the semiconductor market.
Cadence’s Lip-Bu Tan: There are waves coming, and you should ride the wave even if you don’t create the wave. There is a lot of new opportunity, and new technology that needs innovation – things that are more data-centric, with more storage, less latency. And there’s the IoT. I’m putting my bets on intelligence at the edge for manufacturing. Overall, EDA is tremendously undervalued, which is good news from an investor’s point of view.
Simon Segars: For over 30 years, we’ve aggregated the industry – EDA and IP – so you don’t have to reinvent the wheel with every design. Although in some ways it’s business as usual today, complexity is going up and up. This creates opportunities.
Aart de Geus: Complexity is our middle name in this industry.
The Difficult …
Although unwilling to introduce politics into the conversation, Sperling did ask his panelists to address the difficult topics:
* Security as a hardware/software issue.
* Stubborn designers resistant to change, even in the face of growing design complexity.
* The steady migration of EDA customers from West to East.
Simon Segars: We are not making progress fast enough. People are building things which are unsafe and dangerous. And, there will be more incidences of large-scale hacks while the supply chain catches up with what needs to be done in the hardware so the software is safe. It doesn’t requires rocket science to see that people are designing things today and pushing them out to the market without thinking about security. The state of the art is pretty grim.
Wally Rhines: There are all sorts of technologies available today for designing chips that are more secure. The problem is that people who design those chips really don’t want to pay a lot for that additional security. But eventually, there will be a trojan in a chip that causes harm. So now there will be an additional line added to the purchase agreement: ‘If we’re buying this chip from you, you’re guaranteeing there are no trojans.’ Only when the customers begin to say that, will security be a critical part of how people design chips.
Aart de Geus: It’s a very complex problem. You could argue that only the smartest people design chips, but the smartest people are working to hack those chips – people so smart that they would be Fellows in conventional companies. Also, there are a deep-rooted set of issues around security that are not well understood – and different issues between hardware and security. Meanwhile, the intersection between these two is very vulnerable. We need to systematically build security into a design that meets regulations, and we need a design to be secure by construction. Gradually we will have to regulate security.
Lip-Bu Tan: Cyber-security is important, and companies must be aware of the problem and know how to prevent it. Look at our key customers – the verticals – they each have different requirements for security, and how much they are willing to pay for solutions. We’re all working on it, but not until some really big companies make it happen … even now, people know they have to be aware of it, but may not yet know how to prevent it.
Wally Rhines: The truth is, we have lots of new capabilities in our tools and IP, but the designers will continue to use older tools and methods as long as they can get the job down – tools and methods that are slow and obsolete, but reliable. Only if the designers absolutely cannot solve a problem, will they move to new tools.
Lip-Bu Tan: The growth opportunities are exciting, but on the EDA side some of the tools are very old and need a complete rewrite. Now there are chips with 100-to-200 CPUs, and every vertical market has different requirements for the tools, so there’s a steep learning curve and we’re all having to move up that curve.
Aart de Geus: As system complexity is changing dramatically, we continue to be on just as fast a wave of evolution forward as at any other time in the 40 years of our industry. But with the speed of change in system complexity, partitioning the project has become just that much more important.
Wally Rhines: China is investing at least $20 billion per year to stimulate their industry – initially investing in fabs, but now investing in the fabless infrastructure, reporting 1300 fabless IC companies in the country. That’s why China and the Pacific Rim are creating new growth that overshadows the old areas.
Lip-Bu Tan: Clearly IP is very important, and we’re seeing IP is starting to have a trend of more outsourcing, but trying to catch up with the protocols involved in using IP is creating a nightmare to support it all. Nonetheless, I’m excited about growth – even though Moore’s Law is clearly slowing down.
The Downright Bad …
Only when pressed for answers from representatives of the U.S. military at the back of the room, did Sperling allow the conversation to get prickly. When asked what the Government could do to help keep the EDA and IP industries alive and vital, the responses were sharp and immediate.
Aart de Geus: Work on the H1-B visa issue! Since 9/11, the U.S. has shut its doors indiscriminately. Although we rely on having the best people from around the world to grow our industry, it’s becoming impossible for the best people in the world to work in our industry because of these visa restrictions. This is the singular biggest detriment to our making progress. This is state-of-the-art technology that has nothing to do with borders. The whole situation around the H1-B visas today is a tragedy.
Wally Rhines: Be very careful about implementing export controls. If the global markets for our products are restricted, we simply won’t develop the new technologies. We’re not like the big companies who can monitor and control things. When you put those kinds of export restrictions on our products, our industry is made less viable.
Simon Segars: You need to help get people more interested in science and technology. Now there are more lawyers than engineers – the supply/demand here has gotten out of kilter.
Lip-Bu Tan: The Chinese Government is putting their money behind real innovation. If you and the U.S. Government don’t change at the top, you will no longer be competitive! I know a lot of universities where the professors are telling the students to go into software or social media. But hardware needs the best talent out of the universities. Do more to encourage that to happen.
And the Wind howled on …
Will the wind ever remember
The names it has blown in the past
And with this, crush its old age and its wisdom
It whispers, no this will be the last
And the wind cries Mary.
Tags: Aart de Geus, ARM, Brexit, Cadence, Ed Sperling, ESD Alliance, H1-B visas, Lip-Bu Tan, Mentor Graphics, Semiconductor Engineering, Siemens, Simon Segars, SoftBank, Synopsys, Walden C. Rhines