The Software Behind Silicon (with Synopsys Founder Aart de Geus and CEO Sassine Ghazi) 01:16:04

hello， acquiredlistenerstoday is a very special treat our guests are the founding ceo of synopsis are digious and the ceo today， sasine gas synopsis is the eightbillion dollar company that makes the software that chip designers rely on to do their jobs it is one of the two big players along with cadence designs systems the field is called electronic design， automation or eda its a crude analogy but you can think about it as the productivity software for chip designers like the Microsoft Excel or figma for that profession and so much of the complexity of chip design these days has been baked into the eda software that it makes entirely new types of chips possible that you couldnt do without them they are the essential infrastructure behind the AI era and all the semiconductor innovation that we are experiencing today no ai applications would be possible without eda and the incredible optimizations that the software does for chip designers and in fact in a full circle moment synopsis even uses ai now to design the software to design chips so with that onto the interview with art and suscine art ansa cene welcome to acq to thank you for having us we wanted to do a deep dive for listenerers its been a while since we were in the land of semiconductors weve covered Nvidia and tsm， c and apple and arm so many of your customers and companies that you work with， but we have never hit the world of eda directly and so listeners artdigous is one of the storied pioneers of the semiconductors industry 37 years ago founded the company and really evolved to become an essential part with synopsis of the the semiconductor value chain and whole ecosystem today and recently susane you transitioned and took the helm going from COO to ceo so we have the unbelievable privilege of having both of you with us here today pleasure to be here yes， thank you all right so this is acquired and we love history so i think to groundthe current state of the semiconductoricosystem why not we wind back to the beginning of synopsis？

so what did the lay of the land look like then and how crazy was the idf for what would become eda when you were first getting started well？

so what we talking about here is mid eiders right and so just put i guess a little stake in the ground i was a gentle electric designing at about four micron i know you dont remember that that existed but yeah they were this big that also， says that general electric was actually in simiconductors at some point time they invested in sort of the factory of the future this was the future， sort of the same as ai is now everybody needs to have it well then it was semikin doctors things when pretty well until they didnt go so well and then go so well meant in hindsight that in nineteen eighty five was the worst downturn in the history of simiconductors， in the eids and nines， and i think it hits general electric card because that was very large company they had a very steady state sort of divident driven investor a group and so these ups and down in the center doctor industry turned out to not be really there thing long story short we gonna be laid off and so it was completely accidental， but it was also accidental that in the five years after that i work there and especially last three we have developed a number of design tools that actually were very innovative one of those was synthesis we were somewhat known because of that and while on one hand i literally actually interviewed for a job we gonna be laid off after all simultaneously we have this rebellious idea of you what if we took the technology and i looked if we could start to a startup and we did it with i think great care of thinking because we decided very quickly were gonna do that in full light of general meaningtell them about it and not take anything it right was a great company they have taken very good care of us and actually it was just writing to do and there was an opportunity to advocate this spin out with the technology， which always gonna be lost they were gonna get out of of this field with a small group of total people of seven we essentially got out of g with their supports both some financial support and the transactional the technology for whats the equivalent of a million dollars of value and fast forward in by time when public 23 im still proud of that because they really deserve it it was the right outcome。

but it was all pretty accidental how that came about its so rare the incorporate spin out into a little venturestyle venture goes well thats a busy ache wo were you all designing microcessors or was this more specialized？

no， this this was so called gagara is and these were essentially chips for other customers its hard to remember a gateway， but a gateway was essentially a long series of transistors that have been prefabricated they are sitting in rows you use the first layer of metal to essentially take those transistors and make certain gates out of it the nangate and org ate an inverter that was pretty much the choice set and then you use the the rest of the layers to connect those to actually create the actual circuit g would do that for their customers and then manufacture these jobs and gateway thats the ga in fpga right yes， i yeah， i think so actually yeah because the same concept yes！

it is exactly the same concept yes！

i never hurt that question is yeah well！

its funny you said gatara is oh， oh， i havent thought about that in a long time， but fpga is like thats the current hotness you absolutely right well！

yeah， so were hot already then i guess yeah！

so this idea that you have that you could turn it into its own company was there a blueprint that rd existed for chip designers need great software to do their jobs？

well， or was that sort of a novel idea that that could be an independent company well？

you really have two questions at the same time here why did we get into synthesis in the first place， theres a technical question and then oh yeah， how to get to a company its sort of funkey how we got there？

because while it was there， there was a guy a g who had explained in some seminar that if you used multifluxers， you could actually create circuit that would be denser than just and or invert or nand or in inverts and so i talk to one of my designer friends say can you put together the footprint that you need to do a multiplexer and he did and put that in the library and then will design with those and get smaller circuits the pro is none of the designers knew how to use them and so after some reflection yeah， what why dont reduce automatically design that and somehow we manage to write a program called socrtes that actually did that and got quite good results although it turned out in a long term that multiplexes were not a good idea because multibloxes are not restoring logic meaning you put three in a row， your signal degrades whereas with all the others you have the signal stays a very square wave， which is what you wanted？

but in the process， we became quickly known to some paper publish as being on the Frontier of this thing called synthesis and of course， by that time， we discovered that ibm had that worked on it for long time and so did Fujitsu and Oshiba and a whole bunch of large companies but at the same time， we knew we had something because the results were stoundingly good compare to uh what was the manual design done before？

then within g they used it on the gate raise with great results and then the the whole notion of well now suddenly were gonna be laid off and all of that is gone gradually morphed into talking to some vcs， talking to some designers and say well what about creating a company yet to understand at that point time i was a very young person and i had a bunch of way younger people because six out of the seven had all been summerstudents that sort of the cast of characters and the notion of writing a business plan was interesting concept and i still have a couple of the books that i bought in the local bonds and noble of how to write a business plan you like Jenson did the same thing when he had the right of business plan for Nvidia well， although he was already closer to the business side when he worked up to lsi logic same concept fundamentally the one thing i just couldnfigure out was what is the difference between orders， revenues， and sales， and to this day i dont quite understand the developers between sales on revenue in orders but but you know for that we have people now as they say hey you！

you invented great product and those things will figure themselves out deferred revenue bookingsbilings one thing i want to understand quickly before we get to the company before synthesis and software how was chip designed done was for planning done like architext， liquid drafting boards was it pending paper？

oh， yeah， youre right with so many of those pieces but the first thing to understand is this fundamentally two layers this the functional layer and then is the physical layer when you talk about the layout， you already have an understanding of what the function is and what the buildings blocks are now you actually have to physically design them and physically connects them right， we were working at the functional level and and there the notion is you have some complicated mass function a digital mass function that you want to implement and you need to choose the right gates and theres a number of methods simplify that， but ultimately you build a settle building blocks that you， then connect you typically did it on paper or then gradually on on a schematic sentry type thing and then comes the question all how good is it well fewer gates is better your area was not really used the substitute of that point time was just the number of gates， because if you knew that the rest was sort of the determent and then the other thing that was important then that will turn out to be absolutely crucial on how we differentiated is we understood that the speed was key and the speed is determined by whatever is the longest paths through your design and so we we could judge if the circuit was getting better， not only was getting smaller， but also was it getting faster and that combination turned out to be the key differentator fascinating。

so suscine we have not yet gotten to your role in the story and so i want to start from sort of your beginning with synopsis you join the company in nineteen 98， but im sure that in your jobs at at Intel at elsewhere you sort of came across synopsis before so do you remember your first experience？

yeah， i mean as artist describing to you the synthesis， the gates， the function than the place and loud so my first experience with synopsis i was doing my masters an electrical engineering actually， i was more on the control system side so i did not touch synopsis at all after i finish my masters degree i realize thats not the field i wanna be n because most of the job opportunities at the time were controlling massive mechanical stuff be at oil diggers or a giant the satellite or what have you then i started my PhD and vlsi design and this is where synopsis my first introduction to synthesis and as you describing it out how do you build a library the building block you synthesize back to your question David the largest design at the time that a single engineer could do was very limited by the number of gates because the actual software from a capacity point of you could not manage just the clock time to run and synthesize will limit how much can you design in terms of size of design wait so the physical onship limitations were actually not the bottleneck it was the ability for the design software to handle the complexity its both right because first you need the the how much the software can handle the complexity and still meet your performance target and artist right at that time performance target was the key a power area was so secondary， which ten years later， it became performance power than kind of area now you optimize to the end all at the same time in order to make your requirement when i started my career at Intel believe it or not a lot of the stuff that the synthesis creates they were manually verified so you lay out the transistors you make sure you have the right with right length and how you connect them together to create the actual cells so my experience with synopsis was grad school than of course。

at Intel i used many of the synopsis products and thats when the opportunity to join synopsis came along i was super familiar with the company and the support the product rnd and the rest of history and so you join because we were the only company have no buds right this is that right he can you confirm that exactly doing this synthesis class a problem this is really really hard to deal right and what made it particularly hard was there were of course techniques to optimize just a functionality and many of those were algorithmic we added to that we then called an expert system， which was lookout certain situation the circuit say this doesnt look good but i know heres a better version and so you would add so called rules to make it better and you add a rule against better you ads five rules against better you add one more rule gets worse because now you need a rule to manage the rules i always like to highlight that it wasan expert system because that makes us say now kings of ai you know thirty years later but the fact is it was limited in its capability but it was dramatically better than humans and so by the time and it this was not even the first product it was a prototype of the first product that we had as a minute who became a company we talk to customers and they would give us one of their sockets that had maybe max a couple gates they had worked on it for many， many， many， many weeks and then a matter of few hours we could literally give it back to them thirty percent smaller and smaller meant thirty percent fewer gates and thirty percent fast are meaning shorter critical path they would look at it and then its a its impossible theres no way you did that and they would go away literally for two weeks and then they would come back and say well i have checked and i checked its actually doing it and then the expectations of course， with immediately way higher than what we could do because they had just encountered magic right out of that interaction something very profound happened is for the majority they became our friends customers because they could use say yeah but you know what you did here thats not that great and by being able to look at our circuit and say thats not that great it made them great， but they gave us a gift of feedback that two weeks later we had fixed based on their input and therefore， they become you know a parents of the tool to right everybody had added something and that whole first generation of two thousand 3 dozen companies over time， they all had the same behavior， which is they routed for us because they？

they could see it happening on their own circuits it doesnmatter what kind of circuit youmaking whether its a microprocessor or an analog system or a gater a like you need this technology， you need this optimization and so intelppy that you do you are getting better！

even though thats also serving ti or yeah yeah multiple boxes here for starters we were strictly a digital company well， today we do of right these things on alloc circuits automatically， but that was far away plus you know， this was a corner stone to really the digital age and before synopsis and in all fanas， i should say before synthesis and before placing about the two go hand in hand the field was called computer ade design you did stuff on the computer， but computed essentially help you do stuff that you did yourself what was so great about synthesis？

it actually created something and so we were part of the transformation of computerated design to electronic design automation i remember you know we had a moniker for that is you know with the only ones that have license to kill you know what one of her double sevel because license to kill means we can actually change a circuit and that was completely taboo before if a tool did that means put somebugs in it， right， right？

you wouldnt want software intervening in your own workflow the creativity was reserved for the human for the designer you know， please only aid me do not automate for me well！

it was more than creativity it it was the trust that it actually would work people really could not trust that tools would change it for the better yeah and you know its amazing even in 2018 when we introduce ai for this synthesis and place， them loud believe it or not the resistance from our users was i wanna know what the ai change will like， but thats the idea you cannot is just how many parameters do you want to understand and there was a lack of trust for about the first two years even though the outcome the results were always better using the ai system they meaning the users could not trusted or use it because they want to answer the question you know， engineers， but i need to understand what did it do of course， like now was a different story， ai is so well excepted that question is gone what i love about was a scene？

just said is we are essentially a company that has repeated its own history over and over again and i was almost tempted to say that you know would learn it all from ronal rig and trust but verify yeah， so herethis ai stuff and then you still need to simulate a lot to make them sure that you didnt have an error in it but the value of trust is extremely high， but the necessity for verification is also less like right because the cost of going to manufacturing of something that has a bug is wow！

you made a big decision there often i dont trust ai tools because when i do look at the output。

im like its not clear to me that this is better then me doing it and there are many situations where it is a many where isnt do you feel like eda is uniquely well suited toward a designer just letting go and saying i trust the machine and i dont have to understand every little input maybe what usually ferling to Ben is generative ai when is generating something through a natural language and you say um this is 九十 percent accurate not a hundred percent accurate im assuming thats what youreferring to in the eda world were all about optimization massive optimization problems i mean were talking about many billions of transistors that your JAMA in a small silicon area and your plank to optimize where do you place it？

how do you routed to how？

do you get to the performance， the power and you know this is no way for the human to do it so so our indusy has been very much in the space of using technology to optimize so what we do for the ai we were doing generative ai， but put that aside for the moment is using machine learning and ai algorithm to optimize for an outcome but you always have as art mentioned many steps before you committed to manufacturing and say its gonna work meaning you dont just say oh thats what since this is all ai generated is gonna work ill go to who memory your foundry and you spend many， many， many millions and the chip does not work so you have many checks of vertification what we have pioneered with ai for eda starting in 2017， right now is used by a dozens of customers in production meaning they are trusting that the trusting the outcome but there are all the other checks you have to go through to verify that its gonna work once you many factuid and to be clear theres basically specific goals around size。

around power efficiency， and around you know overall performance。

and is basically optimizing within that set of constraint so you you can sort approve what its done that its its better than what happened before the ai to came in those are the outcome metrics right then you just met should see a four and thats what we optimize for but you know uh what that forgets its not only that there is the other ten trillion constrains that you have to meet that tolerate zero error and you the very big difference between many of the ai optimization things that we see in the world and some hallucenate more than others there many very good ones is that we have a constraint that is much much harsher， which is absolute correctness in functionality and by the way the scene jumps quickly 25 years and in those 25 years， this been twenty five years of revolutionary techniques and enhancements not only to what we do but to the circuits that we do with our customers and you know sitting on an expanential of rate of change that is a logical of revolutionary changes and every single one every single one has been delivered in a evolutionary way with other words you forget one lesson learn in yeah， i dont know nineteen 97 crosswalle capacitance you forget that nothing works today nothing while he made it sound so simple yeah， you are over now ai the heck out of it we do but that ai itself works on unbelievable number of parameters and the the rules specifically for the layouts have have no tolerance for error you vilate one of those the yield will go zip down the drink but see thats why when the question often comes up why are there eda startups？

why is the market consolidated to just two its that exact point that are just made the learning its not just hey can i train a model and then create an output and is im there the accumulative knowledge to get to the current state before you look at the future state is massive it strikemean that theres actually！

a lot of parallls to the foundry business just on the software side you know you cant just go recreate tsmc obviously as we are seeing its all those cumulative years of learning about how to do this and this is the same thing in the software yeah！

you know， its interesting because DSMC was found it three months after synopsis thats in hindside super interesting because that was simultaneously a change in the industry where the focus was gonna go towards， fabless design and then founders that would do the manufacturing and remember before that time will many companies were idms they have their own foundry and the if real men have fabs， i think was the quota today we would there say it like that rightfully so， but it was very much added to still to this day people that spend a lot of capital are are really proud of spending the capital theyelse have no choice right there is a slide little fun antidot， which is a morescan founded the company， but the first ceo was a guy by name Jim Dix who happens to be the general manager i work under a g and when they closed e e when there so you know its one big family enterprise here and the other end of dot i like to bring up we over we talk about those customers that use our stuff and we learned from i dont know if you recognize that the names Chris Malacosky and courage a screen of course， well， they were at sun microsystems and they were among our first customers and actually very good guys to work with you know them extremely well， and then a few years later， yeah， jenson， the showed up because he was the career taker from lsa logic， because that is where a sum manufactured its cheps and of course。

then they teamed up but i forget twitch year with ninety three or something like that and so you know of those three companies we are proud to say that we are the ones that have survived the longest so great so seen you gotta tell us the story so i was watching and they will come back to everything else that were talking about here but i was watching the keynote and it appeared that maybe Jenson arrived like literally seconds before he was about to run on stage because it was the maybe concurrent with the week of gtc yes！

yes， its funny my team thought i was joking like i was setting it up this way have that no i mean it when i went on stage he was not there yet he was not in the building but but he was texting he is like im in the parking lot im like OK great im about the hop on stage so um you know initially the the idea was about four five minutes then i bring him on stage and im looking at the clock thinking have that let me burn more time than and all right should i continue with my presentation or wait yeah it was live its wild ive sure this existing other industries but it has to be a very special thing in your industry that sometimes theres a company that becomes an unbelievably important company in the world and you sort of get to be a huge partner to them in that success mean its absolutely fair to say that theres no chance Nvidia could do what they do without synopsis software？

how do you think about the role and sort of the importance in the world that you really become as i mentioned in the keynote？

i want to say even in jensons words because i was not putting words in his mouth that the synopsis is mission critical to Nvidia success we dont think that likely when we know we are a mission critical to many billions of dollars of our customers revenue its a huge responsibility its a responsibility to continue on innovating because theyre inspiring to build the bigger， product， the next big thing， and if i were software， and i were support， our ecosystem engagement is not able to stay ahead i dont wanna say with them ahead， so when they are ready， we have it it wont happen and right now where its happening is with our chip customers as their architecting the future product and with foundly and those two are becoming so important， so theres the triangle always us customer foundly that we are working on architecture， on physics and manufacturing and software to bring it all together and i want to say that has accelerated in terms of being interconnected in the finet world you know， when the transistor start moving to more complicated manufacturing， and of course， the last five six seven years， when we say its impossible to design and manufacturing， those chips without our contribution as an indusy， its not then overstatement theyre sort of also。

a historical perspective that yeah， we can only be thankful for that we were part of this， while now it feels old more slow essentially was exhibit of what an expanential is and an exponential is easily the toughest mother of mathematical function because staying on that socker there is going fast right， we were lucky that we had similar technology at a moment where similar technology was needed again to to move forward and and continue to move at an unbelievable speed not necessarily exactly the exponential that Gordon have predicted， but still the exponential that change mankind right！

i think this is an important point that i dont want to just over here people take more slaw as if its some derivefrom the natural universe property， the same way of like equals i may or something its not it literally realize on companies like synopsis getting clever again like every time moreslowhappens itve because somebody got clever again and oh my god we just barely made it youre the causeof it not the result of it well its interesting because more so of course。

started just as an observation right he had seen this curve as moving up rapidly， and then he made some prediction that it probably would continue for a while that prediction， then became sort of day well， you have to do it， because otherwise you not use you， you not with the team here right there the races on and then that race itself started to selftime itself against that， and by the way that includes also， the different switches to to the different sizes of wafers yeah， the ability to manufacture and you know it was not always the perfect exponential， but the guest out of it was absolutely and this is not something new a in human kind history the printing press had exactly the same characteristic how in a essentially fifty years from virtually zero books at when to twenty million and change the world then of course， in many ways， the industrial aids have the same cattristic again what is so exceptional about this one is you know if we look at what we have done sanops is weve contributed about ten million x in productivity ten million x and you say well you know are you gonna do another point five now hell no， we need to do another ten to hundred ax and of course， thats not gonna be possible who is just doing that on a one chip and we ill sure well get to the whole motion of this more and how that is changing things but what is important that two things one is that in order to stay on that exponential you need to race like crazy and the way you do that is you race with people that are crazier than you with other words you go to those customers that are even more paranoid of not being successful and that are thankful but never happy that thats a polite version thats a good version and they drive you are crazy and weve had the good fortune to i want to say for i dont know seventy five percent of our products always be the state of the art for all this time right， so weve been there and i like to compare sometimes to you the two of the France yeah， you see all these guys by king like crazy and then only the three guys that move away from the poletone and by the time those three guys are a couple， hundred yards away， the others will never catch up and theres a reason for that the others can not team up welven off whereas， those three guys by necessity and by scale you know every uh， you know fifteen seconds or whatever it is they change the guys upfront but they chase each other until the last hundred yard and then it see how everybodys on their own， but in all case。

the race never finishes how to the franciis now thirty seven years and you need to keep going out at but that combination is has been unique in in this industry well listeners this is the perfect opportunity to introduce a new sponsor here on ack to quarter their new product quarter pro launched about a year ago and is already adopted by several fortune five hundred companies and some of the worlds largest hedgefunds and equity research departments yeah！

this research platform is transforming the way qualitative public market research is conducted heresalequarterpro works you can get every piece of first party information from public companies all in one single place thats live earnings calls with real time transcripts company filings， slidedex and more quarterpro has built a world class user interface for this yep lets say youre an invester or a podcaster and youve got the use case where you need to look up a company such as novonordescaremes or visa you can open their platform and search guidance or marketoutlook quarterpro then immediately identifies all instances where a company has historically mentioned and discuss these topics in all of their ir related communications or heres another pretty crazy thing theyve done thats difficult to get anywhere else you can actually search through literally every individual slide in quarters database covering 9000 public companies and millions of slides for any keyword mentionbased on quarters aicapabilities this trulymakes it easier than ever to conduct qualitative analysis of entire industry values。

chains and specific companies so whether youre an equity research channellist， an asset manager or an investor relations professional this platform will help you increase your productivity through their livecall transcripcomponentsai， powered summaries and a feature allowing you to visualize the entire timeline and changes of specificslides three woutquarters quarters also offers their database as an API solution this enables other companies， such as trading and research platforms as was ai and lm companies to build a custom solutions and integrate this database into their offerings or adfunctionality on top of the data yep to find out why leading companies globally are choosing quarterpro in their data day work and to experience the platform first hand request a personal demo by visiting quarterdot com slash acquiredthatsqa rtr no 一 qu artr dot com slash acquired work click the link in the shownotes to get the personal demo from the quarterteam are thanks to quarter so obviously for these advancements for moresla to stay true it requires incredible collaboration between say tsmc to figure out how to you know reduce the number of Adams in between you know the two transistors or whatever it is that is measured as three nanometers and it requires you know asml to make an even better elaser and requires the eda software to become even better and requires the cleverness of the chip designer and at everything has to work together in the same generation at the same time betting on each othersdependencies does it feel to you that ekingoutthat next generation performance gain is harder than ever or has it always felt this hard well。

i would say it always felt this hard i think its different than it was but in the description that you gave you have a relationship between the foundry and the equipment vendors and you mentioned asml applied materials those would be the ones that essentially focus on so how many Adams are i specifically and an asmr would be focusing on how many photoms do you need to get at， which frequency in order to get really small lines by the way we are in that domain two because the officer is the leader in t cats technology computer ated design， which is yet another simulation or modeling if you like of of the truly menu school， right and then at the same time， you alluded to the fact that you have to align how you design with the building blocks that you have available and you can say well， lets optimize the building blocks what you designing， or lets optimize what you designing for the building blocks that you have this is often called DTCO design technology co optimization where synopsis is a leader in and what you notice in this story is they do the designs they do the manufacturing， but we make it all happen right somebody builds the legal block somebody dos the castle we make sure that all these things hang together while i agree art its always been difficult。

but ill say the last six plus years its been much more difficult or and the reason im saying that if you look at synopsis relationship with foundly say six seven years ago， we used to get the input from foundry called enablement and you enable whatever they create in our product and you provide the product to the customer so you become the bridge between foundly customer some i conduct the company designing on that foundly through enablement so you take whatever they created and then you put it in your product and you give it to the customer the last five six years its impossible any longer to just do enablement we sit with hundreds of engineers， at TSMC， at Samsung， at Intel， at GF sitting during the process development the technology development its no longer enabelement because enabelement is impossible you have to invent stuff with them to see will your physics the way， your pushing it will at handle the design your your espiring to design on it and thats a big change that happened that is different than before actually！

i buy what you say i was coming from the its the same because we have always worked as hard as we could yeah thats that that does not change and the pressure for drivings but but i think what you introducing is actually， the notion that beyond scale complexity with now engine systemi complexity for every level and the systemic complexity at the manufacturing size is that its not sufficient to just understand sub pieces you need to understand how the pieces work together and i think that we benefit a lot from the understanding from the foundies but they benefit a lot from the optimizations we do that now help them get faster and better its even to continue scaling i mean if you look at it now。

its really unbelievable to think about that the indusy stocking about eighteen ankstrum fourteen ankstrum no longer nanometer those are not only from a physics limitation you are hitting the limit once you want to put them into a production on a chip like you know the latest announcement from Nvidia was black， while two hundred and eight billion transistors imagine the heat those transistors are generating， so just from a thermal so in physics when youre designing that transistor， you say oh， i to it will work but once you gem them together and you run the software at the full workload the heat that is generating， so even though the physics from a manufacturing it works there are many aspects that you hitting the wall that you need to plan design architect for is it reasonable to say that before you just had manufacturing limitations working against the edge of whats possible。

and now were actually bumping up against the edges of physics governing whats possible completely。

when you think of the new wave of designing chips and this is what art was talking about you start designing through advance packaging multiple dies sitting in an a package， electronically you can design it to function correctly， and then you can manufacture it and package it to function correctly once you start running get in the field with the software workload， then you run into all kind of physics issues thermal is the biggest one， but thermal when it creates heat， it may create warpages， it may create cracking， so things will start cracking mechanically so you need to take into account all these physics effect during the design stage and during the process technology development so its absolutely multi dimensional design factor you have to take into account if i can just entertain a thought exercise whatwe need to do to get 4X？

8X， 16X more performance from here what are the innovations that need to happen for that to be possible the two things？

the first one is is well understand from many years ago， which is develop the hardware for the specific workload， somewhat overly simplified the first years of moreslaw is here are more transisters better circuitry right yourself tware and people said wow， i can now do so many more and then is oh， you need more memory is more memory right yourself where and make the world happen right and then then can gradually sort of this condrum of all yeah， but can you not make it fast like a lot fast？

especially all that visual stuff on the screen its so slow and then out of nowhere somebody says yeah， what why dont we not use a general purpose processor to do pixels and then the thing becomes called the gpu and what does the gpu do it loves pixles it doesnt only pixels it can do them forward and backwards and sideways and song and out of that is essentially a specialized accelerator and then of course they discovered that well would be better to have two or four or sixteen actually， multicore even smaller processes and essentially what you have is now a workload that has determined the hardware that you need now advanced thats to a of twentyyears later and say the workload is driving a car without accidents you can imagine that by saying oh， yeah to take your old the three eighty six and see what you can do with that you gonna go nowhere right you need actually， a whole bunch of specialized machines from the anything that takes the many sensors they die and compresses it to our transport at ansall to ultimate vdai algorithms that can run preferably a real time to drive the car and so one of the the statements on that is called software define the architectures and i i show it sort of as this v from top down because you starting with high level function how he drive it correctly and get there at the same time and its literally at the same time you come to the conclusion that you know chips that are more than one and a half inch square and i know that some people do whole wafers but yeah， it quickly gets to an end and adding another zero in the number of transistor is gonna be a really long， long home and so you say well， what if if we split functionality into multiple chips， what if we brought them really close together and there in lies the essence the word close together because the notion of heavy multiple die maybe on an interposer， which is itself a check right is not new， but a it was difficult， it was expensive， and b it was slow and if you look at the evolution of the last twenty years， the single thing in my opinion that is empowering multidie is connectivity， meaning we have improved dramatically dramatically not only reducing the distance but the band was meaning how many pins you can do how small these pins are and how little energy they need to flip a bit or to pass a bit from one chip to another still way more than keep in on the same chap if could just keep on the same chap， that would be cool。

but thats not gonna be possible and this is of course the black well thats an example is the new Nvidia black well chip is that the silicon interposer between two dies that enables super fast information to flow between the two dies and into an amd have very similar constructs and they all increasingly now look like there are to twenty or so chips and by the way。

the these chips dont all dont have all be processors actually need memories and the cool thing was memory she gradually can stack them and they stack potentially better because they doncreate the heat that suzine was creating in his processors thermal is absolutely one of the big killers and all of those and a few others that the enabler is connectivity and so if you now look at a picture of sort of bottom up from physics， you got to this whole new architecture， thats really connectivity driven you come down from software as in software driven the word architecture has a functional perspective and has a physical perspective and so that opens an entire new age we call it this more so systemic complexity with a moreslow exponential ambition and i like to use the word exponential because im a strong believer that what we see happening is another twenty years of additional complexity and speed may have to be a redefined as well we do a whole bunch of things in parallel but thats a different form of speed right but any speeds you can improve is still valuable so with system or what youre basically saying is we are going to abstract up one level what the notion of the system is were not measuring wars loss specifically on this one chip anymore were measuring it for your whole system that were the goal might be drive this car safely are we able to optimize more components of it to work together harmoniously to continue to achieve more slow like outcomes yeah， i accept what i would add is it is not abstracting one level weve been abstracting more levels already for many years and i think that includes the software， the embedded software， the software that connects to other other pieces then ultimately the various forms of ai optimizations， and then still the domain specific knowledge of that a a great example of this is if you were to ask us hey， you know if you really wanted to cut another twenty percent of the power！

which layer would you start with i can tell you would not be the transistor it will be the software somewhere its kind of like whenever im tempted to buy a lighter carbon bicycle i realize that instead of spending three thousand dollars to shave announce i could probably lose a pound and it would be nothing but advantageous yeah！

but i example， in case you are the software and youre allowed to softer。

oh， of course， that whats whats the old phrase about bicycles and plus， one is the right number of bikes to have yeah！

exactly！

i have a t shirt that says just one more guitar and its to say one more guitar and you gonna be a great musication the difference between you and greatnesses right there uh thats awesome OK？

so if i could perhaps paraphrase the two things that you said its this idea that hey， what if we admit that density is going to be really really hard from here to get even more density on a chip， so either a we can stop making so many trade offs in the hardware to accommodate general purpose computing and just make specialized hardware and be we can horizonally scale we can just connect more dized together so we basically have have more compute and yeah， its gonna take twice as much space for twice as much compute。

but at least we get twice as much compute the only thing i i would slightly tune in what you said is all these things multiply and so you should have dont care at which layer of abstraction you can have an improvement if you can improve the transices by five percent thats still five percent that applies to a lot of things right and not all things benefit equally and so its been interesting to listen to some of the people that that manage big compute centers they would say you know i dont care how much bar use on the processor because if you make the processor faster， i can leverage that on all the other chips that are expensive to buy at or or to run so systemi complexity is fundamentally defined in the simple mass of multipllications where are scale complexity and mostly additions and so yeah， we like to have more transistors。

but its the multiplication of effect that changes what you can do in my mind there are other factors to since you listen to my keynote i i call that on two vectors one is the March to anks from that is always the opportunity to advance on most law than there is the March to trillion the trillion transistors， which will only happen through multidii architecture decisions you need to make technically theyre both doable the decisions are what customers are making out financial decisions does it make sense lets say for your next phone to have a chip that may cost fifteen thousand dollar the answers know but hey you can run ai on the edge is gonna be very cool is gonna be super fast yeah sure but you can not afford it some of those chips were talking about theyre selling for 2530 K a pop uh for the certain applications because the yield is horrible because youre pushing the limit of everything so the architecture is not only technical decision its a economical decision you have to decide and then how much do you go down the ankterm？

how much do you go up the architecture for that blade off put another way this is an end dimensional space where the dimensions are actually different for every customer and use case and yet you are still trying to optimize them and produce sort of the best product suite that you can exactly thats why right now when we are talking to our customers？

were not talking about hey， we have this product design whatever you want with it， we have and market specific we talk to automotive customer is a very different conversation than the mobile than the data center for all those reasons that changed again six seven years ago， we did not have those and market focused discussion because as the same product you can develop on the same rhythm of more zla and life is great now there are all these tradeoffs that you need to take into account the foundation is the same tech anomics right every technical decision the simultaneously an economic of decision be it for the build or for the use side of things if you go back to the very point。

we started， which was years a synthesizer that creates functionality， which is the value， and it does that uh with a performance site and a number of gates and the number of gates is essentially the anomics that that determines how expensive is gonna be to to manufacture that， and that is now taken so many dimensions and if you look at lets say the manufacturing side the these expension boosts have been almost order of magnitude over time because OK， are you gonna do a three on a millimeter fat now？

we did two on a millimeter well， the entire industry has to retool for that and you bet it becomes very more， much more expensive and right now theres no visibility to do format millimeter partially because its too hard to coordinate an entire industry to get to that point and so the economics at some point in time taperoff and this is where innovation comes in？

of course。

can you do a differently right and so multidize only a is an answer to that fascinating this question always seems to be devisive for people in this industry is ev lithography a technology thats going to get us through over the next decator to or do we need to find a new better way yes！

the bout yes！

and yes， i mean theres still much to do where the new generations of these machines coming at the same time theres also a lot of development in the manufacturing from a material side with as much as possible selflining devices so where where you dont need a mask for every layer that you depose and also for places where you can actually let me call it a road material under other things on the side way fashion and so the reason i am on purpose a little bit open ended on this is because we have learned many many times that saying no always turns out to be wrong and yeah being as an advance symic conductor exhibition， or i should take conference as a undergrad student in nineteen， seventy， eight and the leaders in the fields were all unanimous electronic gonna be big and you know one micron of course， is the physical limit many years later， i had your opportunity give a metal to one of the guys had said that and of course， couldnt resist bring it up what he had said， but but at the same time so great to give the metal to the very person who had predicted impossible and then was an engineer and made it happen to get around it and this is happening in the core extreme race as as a scene mentioned it is happening right there and remember its what is it fifteen years ago finfet it will never happen these vertical woebly things that you can barely it will never happen and for sure they will never be in cars here we are engineering is is very different than science we work around science yeah！

you visa have a lot of mileage that can serve i dont know im not too familiar with what will come next after it but but its still an early adoption from a process technology point of view and when was the the transition happened， i mean number of found reasresisted for cost reasons found behind then you like you know what forget the cost from the state of elevent， i need to be on it whats next i dont know well！

i mean the the asmr folks， the technology leader say they see another decade of the delivery but yeah， we all think so right and so yeah it has my legyeah！

yeah！

and if they can we send office will help work around it yeah！

engineering！

engineering！

yes！

yeah on the Fred of kind of curious snops is a wonderful company great revenues incredible market cat at all these things， but a something not log the way like you cant became something more to you are one of a few lintpends in the system did you see that in the beginning of like oh you look at any expination function and you it goes as long as moresliw effectively has like its gonna undergure the world eventually， you when did this become a parent to you well。

i you know i think sucene alluted to to one of the aspect on technology side， which is when the relationship with the top foundaries started to change， because only they have touch some boundaries that they couldnt get beyond and we need it to get their information in order to be able to model what the circus actually would be able to do and so i will put everyone of those under the notional systemic complexity and by the way systemic complexity is not a last thirty years， not and future years whatever you do want you reach some boundaries susteme complexity becomes the thing that you have to handle around that thing and the systemicomplexity of a single transistor today is unbelievable right， but we would have thought of it us super simple legal blocks， and so so this is happened the second thing is when the architectures started it to somehow have a wish list on physical behavior， which was far away from where they were and so certainly， they have certain desires of how fast to access uh the memories in order to be certain computations in the vice versa to domains that have been nicely separate for good reasons and once they they get closely closer sonlythey are one and that is a moment of systemic complexity there was a movement coming down from that perspective and then in a whole different game was the notion of globalization had been successful and suddenly you delled with parties literally all over the world that could only be successful by having a a chain of participation and collaborations， and so yeah， if theres a singlar skill that matters more in systemicomplexing than anything else， its a combination of trust in collaboration， and i think synopsis emerged as hopefully trustfully good enough， but also needing an intending on collaboration and that was fantastic and of course， you know the fact that theres deep loalization in the world in the last what seven eight years complicate things for many people but at the same time you know the its the skill said that still relevant for the future is gonna be way more relevant not less yeah！

maybe another way to answer your question David if you go back， maybe fifteen years ago i would indusy was not that exciting it was so hard to reclude it was so difficult to bring gain young fresh a blood out of school into not only eda eda and some i conduct i remember when i was a gm of the rnd product development， one of the initiatives was how do we excite the next generation to study electrical engineering to study computed engineering because it was like now， its not exciting i wanna study be more on the on the software side maybe thats what you were in your background when i started in venture in in 20。

we had a startup eda type company that was in the portfolio and it was like the black sheep we yes like it yes， we want to find like Facebook click thats right！

thats right！

exactly who now i believe is a customer of synopsis and designs their own chips thats right exactly so and now is very different！

very different because theres a recognition that in order to drive that ambition of software of applications etc you can for sure by a general purpose chip that youre not going to be competitive so how do you customize from the silicon all the way up to the system to the application that you designing thats why many companies who can afford that theyre trying to develop their own silicon or architect their silicon because they know the importance of the silicon in the context of the software and the apps the building if you ask me fifteen years ago， do i envision were gonna be at this point？

i didnt see it we could see that were gonna March down mos law but that now with ai as a huge opportunity to the super every market then every market needs to go through its own transformation at the software level system level the way they have designing that and product and whats powering it is the silicon but not by itself an isolation silicon in the context off for each and market application so seems just trigger to a thought by beinging in the vertical market and making this vertical movement what is changed is we started in a technology with it was a technology push and then there was an economic success of the people applying it too softer whatever it was what is happened now is the technology push continues but there is an end market and markets blural pull and having a technology push an a end market pull accelerate things substantially right and of course， meanwhile， everybodys enondated by big data what the help do you do with that？

well， you need to process it somehow and by the way its gonna change your business well， those are very big statements right， and then they come to the semiconductor world and say yeah i need something much fast or much bigger， and we are like racing forward but we have direct impact in their pnl， on the private part， on their differentation whereas in the past， they partially hope that us well， yeah， you expensive tools and now its like we open the door with them and for them its a heck of a tail wind as the world has this pull that youtalking about。

but also， as specialization of computing becomes more and more important to eke out that next frontier you know you used to just sell to a handful of companies and now theres a strong incentive for many more companies to design their own silicon specifically i think its true that eight of the top ten market cap companies in the world design their own chips so the only companies that dont tominknowledge may be its some secret project are berkture half away and soudy your Ramco and so your your customer base has exploded berksr is half of apple thats true half of berkser is apple so they they do you know anybodythere we could call to help them right？

right yeah， right？

right your customer base has exploded and the sort of level of importance of silicon in their business has also exploded so you have this like dual access tail when the toping you yes！

yes， actually， the the number that i typically share that people get big eyes when they hear it fifteen years ago， pretty much a hundred percent of synopsis revenue was some iconductor companies today， 45 percent of our revenue and of course， we went from the a billion and a half to six billion and revenue so the base got much bigger for five percent of our revenue are system companies system companies meaning those are end market oems that they develop an and they dont sell chips theyre selling a product so that gives you a sense of exactly the point youre making fifteen years ago i cant imagine like the co of Toyota would like come see you guys but today they are right or four i think literally is an example of four designs their own chips i mean maybe every car company does now， but that was always a thing that they bought through intermediaries exactly the key point though even if youre not designing your own chip， right now， youre talking to synopsis， so say you are an automotive oem that you have no intention to design your chip however， you need to architect your electronics given the context of electronics is gonna get higher， bigger and bigger and bigger given electrification， autonomy， etc etc so you are hiring you the automotive or em you are hiring chip architects without an intention to design a chip so you can architect your electronics in the car and those are the customers of synopsis because we have software that enables them to virtualize the entire electronic system thats the exciting opportunity of the future fascinating OK！

i cant believe weve gotten this deep in the episode of that as in the question in January the news。

broke that you are making quite a large acquisition of a company called answers what is the logic there and how does it all work together you know we touched on number of the y and how the world is changing i want to describe it in two reasons why were doing it a reason number one deep in our core business as i uh mentioned earlier the challenge of going down the north law is no longer an electronics only challenge is electronics and deep physics when it comes to putting these devices in a chip， thermal， structual， etc and ansas is the leader in simulation and analysis in those spaces， so thats an hour deep quota business the other vector is what we just touched on as well， which is many system companies less continue picking on automotive as a system， or em theyre trying to figure out how do i design my whole card that has bunch of electronics that is gonna trigger a mechanical action thats gonna trigger a number of other physics action they call to multiphysics， meaning different type of physics analysis that you need to do how do i design the card with a way to simulate everything up from i e， a digital twin of a card， including the electronics， the mechanical， etc and again answers is the leader in the simulation and analysis of that multiphysics so we see the opportunity at the silicon level and at the system level， and thats why we have describing our company as the design solution from silicon to system and were looking forward when you bring to great companies to really deliver the engineering platform of the future its an awesome opportunity will excited about i may be over simplifying here。

but you know theres some set of things on the eda side that synapsis does really well， theres something that cadence does really well but in simulation。

theres basically just ansas everybodyneds does that feel like its a reasonable characterization that remember the discussion we had in the beginning theres the cumulative learning that you have in order to be the trusted similar an ansas a number of similation when i say there， the indusy leader meaning they are the trusted similar because theyve had a history of forty plus years of cumulative evolution of their simulation to be clear though in every space same same as you describe in eda that is synopsis and there are number of other companies that to we compete with in their space is the same is the same but the history of that cumulative learning they are the leader in having that history and that whats called the sign of trust meaning once you do the simulation can you sign off that i can trust that outcome and thats a key in what they offer and what synopsis offer the fascinating the simulation is such an interesting area because on the one hand。

it can help your customers do better like literally use your own existing software packages better， and then in addition to that now that all this new hardware complexity and ai exists were going to get better as a species of simulating way more things in the world， and so it also creates more demand for everything else that you make to the extent that the market for simulation broadly is going to grow exactly so one of the thesis that we explain to avoid investors after announcing！

the acquisition is picture the world five plus years from now that physical testing physical testing in the context of whatever that end device that your physically testing is going to become more connected and smarter right because its gonna have some sort of chip in it， because its interconnected and smarter it gets too expensive and too long， and just not practical too physically test stuff so simulation plays a huge role that in that same context when you think of simulation， you think of digital twining stuff， you think of virtualizing stuff， and this is where we see our core competency of what weve done at the silicon level because thats you do when you design a chip， you virtualize you model you simulate you can do with now for much bigger systems than the chip simulation has always been a good idea it just wasnt technically possible for that many use cases before， and it seems like were now getting more and more fidelity on the physical world of simulating more complex projects and you have accelerated compute where before it may take you weeks to simulate a very small function now with accelerated compute one of the slides was presented that GTC and in my keynote was ten fifteen twenty x speed up due to accelerated compute thats a massive speed up then you lay up on top of it ai for the further acceleration how do you get smarter in what do you similate more effectively， efficiently， etc using ai Technix so its opening up the door exactly what you said Ben for more applications to simulate you know youdescribing the company mostly through technical terms right？

but the reality is it is a group of people first and foremost one of the things that has helped set ops is precisely in this notion of all this learning that that succeed was talking about over the years accumulating that has been enhanced greatly by having many people here work here for a long time and you know both the seen and i are sort of examples of that and at the same time continually rejuvenates both with new people or different people， but also in our own learning of how you do things and i think that is a a fairly unique cattristic， and of course that some companies that we admire greatly because they showcase there she mentioned and video， but i would only put to tsmc also， in that category of yeah never quite being satisfied with yesterday yet tomorrow is impossible， but you know tomorrow was only twelve years away and so you better get going right there is a passionfor making the future happen that is probably half a yeah grounded in other paranoia， as an only different partner as survive had and partially also， in a in a belief that things are possible that we still have to invent that is a very unique recipe for companies and i think that that is certainly one of things up characterizes on options well！

thank you both for your time my closing question that ive been enjoying recently is lets flash back all the way to are we started the episode both of your first experience with synopsis what is the same today as it was then i whats something that couldnbe more different why dont you start to see the passion towards innovation its always been there from day one as are just made the last comment he made its always been what we are working on to model is almost impossible！

very difficult to do thats an indusy thats a privilege in our indusy because thats the key to innovate you talk to avoid engineers they love the fact that working on the most complex things known for human kind thats not only not stopping the opportunities right now to monetize the opportunity to be at the center of what youre working on is so relevant to many inflexion points that that happening in front of us thats thrillling in my mind i probably may be not by accident land on sort of the same as the scene has。

which is web touch the exponential and its in our dna and that socker is not going away it takes different forms right， but tenx is still ten x but the the the next tenx is of the old tenx right that rate of change is just exceptional and have even been in some way， somewhat central and a big piece of that roadmap is is a privilege that is amazing at the same time if you look at the rate of change for us as a company in terms of size and then of course complexity but also of the world you know we started this where the far east was not very important yet and today its one of the dominant parts of the high tech ecosystem， and its also part of one of the big stress fields in the world that that adds an enormous amount of complexity and so being now， a company that is in the middle of these types of things， thats only needs to have opinions， but also careful actions of how you behave in a political mindfields how do you behave in a situation where you see our industry is going to go to about tenx， more energy utilization and without all the the ramifications of touching what is happening in terms of climate change an industry that simultaneously we have had multiple countries in various states of doing well or deep war and how do you deal with that brings a set of questions to us that as leaders we have to learn just as much there as anywhere else and its interesting to what degree companies are now counted on as both sort of i dont want to say too strongly moral centers of gravity， but certainly value centers of gravity as people are finding difficulty， finding it in the political environments or in some cases， finding is really well， some cases not were finding it at all in relision and so now the question is you know what societall groupings matter and we for a long time said that they who have the the brains to understand should have the heart to help with other words we call responsible for the communities that that were part of and by now the community is humanity right and so we started to modify the old bit into a they we have the brains to understand should have the courage to act and that is different than before i can imagine i was part of the business plan that youre going to bars a noble to uh figure well， you know we we participated in a March for in the support of people having aids in the nines and they were very strong opinions that said well， thats not cool because， because all things that today we think as you know this was middle ages thinking and theres still a lot of middle age thinking now and so i dont know where that leads and i you know i think we we have the the great fortune to have not only a new the leadership that can give the next decay of moving its forward in a company that does well， but is at the same time the question so what position do we take in the worlds that is more than send options as a tech maybe super tech company， but as a tech company， as a human company。

that those are interesting questions i cant imagine a better place to leave it were going to have to do another episode to explore all of that thank you both so much for your time thank you for your uh great engagement yeah that was fun thank you both， thank you listeners well， see you next time。