Yes, designing chips is hard, it takes a lot of knowledge. This is why medical doctors need to go through all that schooling... designing a tiny chip with more transistors running software that does amazing things is very difficult.
My Ph.D. is in computer engineering, specifically VLSI and chip design. This was from a few years ago. I _probably_ should have gone into industry, I mean, after all, it is what I went to school for and wanted to do. However, the starting salary for a chip designer (Intel / AMD / HP / IBM) was literally less than I was making at a side job (I worked my way through my Ph.D) as an IT sysadmin. Not only that, people that I knew well that graduated before me would call me up and tell me it was worse than hell itself. 80 hour weeks? Completely normal, outside of the 2 hours of commute time. Barely make rent because you live in California? Check. Pages / Calls all hours of the day outside of work? Check. 80 hours? You mean 100 hours a week leading up to a release, right? Check.
Looking back on it, it seems this was "the challenging" and if you made it past this (something like 5 years on) things calmed down for a chip designer and you moved into a more "modest" 60-80 hours a week role with less pressure and somewhat of a pay increase.
Yes, how do you attract talent under those conditions? It is not flashy work, takes a lot of schooling and the rewards are low. At least medical doctors can kind of look forward to "well, I can make _real_ money doing this", and have the satisfaction of "I helped a lot of people".
I have never done any chip design work, but I have seen some hobbyists going on with HDL without a degree. It's definitely not professional level but I suspect they are hireable materials at least. This leads me to ask the two questions:
1) If chip design (or X, anything) is so vital, so important to national security, why do universities insist that a degree of X include a lot of unrelated courses? You can argue that universities are not just for employment (yeah, as if most people go to university just for fun), but by the name of God, I really hate it when my university forced me to go through all those BS selective courses to reach 120. If you ask me, it's just money grabbing.
2) Why can't students go straight to a fab or whatever after bachelor and do their masters THERE? Isn't the industry a much better place to do that? Actually, why don't the industry simply hire high school students and go from there? Companies used to do that in the 50s/60s. I don't know if they still do that but I think it's rare.
Are you asking why college isn’t a vocational school or technical school? Or are you wondering why the USA doesn’t have apprenticeships in the way Germany and others have?
A college isn’t really meant to be a narrow tracked vocational school - but it’s fair to ask why aren’t their more vocational schools for high tech fieldss
For number 1, there really does need to be some wider reforms, but I fear that won't happen until the whole student loan paradigm crashes and burns. I had a couple of fully paid semesters that were 100% electives I had no interest in. I would have preferred to graduate a year earlier (and thus, a year richer) or take courses that were actually relevant. Problem is that universities talk out both sides of their mouth -- raking in huge amounts of cash like a corporation, feasting off guaranteed loans from 18 year olds, while espousing some nobler concept of education/enlightenment on the other hand.
These are horrendously difficult questions, though a partial answer to 2) is that labor (with "Baumol" training costs factored in) was so cheap up till the early 60s that high schoolers were easily competitive with college grads..
Ah, thanks for the link. I do hope that labor gets more share in the revenue. I kinda love the idea of "cooperative corporations" where shares are more or less evenly shared by the employees (there are ofc differences between different levels, and I'd imagine a large amount of stocks are "frozen"). I think it's going to be a lot harmonic for all employees (even the managers).
>It is
largely for this reason that performing arts organizations in financial
difficulty have often managed to shift part of their financial burden
back to the performers--and to the managements, who also are generally very poorly paid by commercial standards.
Imho the paper continues to function after you patch it with your favorite calling, design-driven industry or organizational structure :)
I know of a bunch of people, some who only have high school degrees, who are entirely self-taught and are doing tinytapeout (https://tinytapeout.com/) chip designs. Yeah, that's not nearly at the skill or scale of designing CPUs for Intel/AMD/NVIDIA/Apple/Marvell/etc but it's still chip design!
Your concerns about horribly long hours and lower than IT/software pay are the most concerning part to me. But, if there's really a shortage of engineers who know how to do chip design, hopefully the market will take care of that via supply/demand at least once things get really out of whack.
It's always crazy to me that you can build crappy websites in the most flexible environment imaginable and make way more than those doing the actually deep and challenging work required for those websites to run in the first place.
These truths mostly also apply when answering the eternal* question:
Why is everybody outside music, movies, crypto & pizza struggling to attract talent?
* Snow Crash (1992) might turn out not to be so precisely prescient due to upcoming dedollarization, AI democratization/bubble burst (the exact option depends on your personality type), & the solid state battery boom:
>When it gets down to it — talking trade balances here — once we've brain-drained all our technology into other countries, once things have evened out, they're making cars in Bolivia and microwave ovens in Tadzhikistan and selling them here — once our edge in natural resources has been made irrelevant by giant Hong Kong ships and dirigibles that can ship North Dakota all the way to New Zealand for a nickel — once the Invisible Hand has taken away all those historical inequities and smeared them out into a broad global layer of what a Pakistani brickmaker would consider to be prosperity — y'know what? There's only four things we do better than anyone else:
Looks like some kind of a gatekeeping ritual to rationalize why upper management salaries are in millions. I think we can also see in other industries too.
You hire hundreds of interns and entry level workers to let them fight in the bloodbath for 100h a week. Pay peanuts. Let them do all the work.
The ones who survive get a bit bigger salaries. Those who still persist in upper level bloodbaths are upgraded into millionaires. And paying them millions looks acceptable as it is so hard to reach the top.
While you clearly could share all those millions between entry level and paid internships, don't have 100h weeks and have a healthy industry.
The job is too niche. This means there are very few employers worldwide that can make use of your experience. That actually puts downwards pressure on your salary (the employer is a quasi monopsony), not to mention that you are tied to a few locations.
The reason software pays so well is that there’s a lot of demand by a lot of companies for the best in a global market AND the enough of the best can counter those offers by going to VCs to fund their own company. EE market is not as global and much more difficult to fund. The only ones paying the biggest salaries are the big companies
I this strikes me as less than half the reason, though still a very big factor.
If I had to guess it would be that every business needs (or thinks they need) a custom website, and often some custom software processes, and definitely their own custom network and IT, and likely some slightly customized ERP. But rarely, very very rarely, do any businesses want custom chips or computers such that a CE is required. The demand for software is so high.
What is interesting in these kinds of discussions is that intuitively, nobody really beliefs in the law of supply & demand. On an emotional level, people always think that effort and difficulty should be rewarded on their own - which is actually correct from a social and ethical POV, and yet:
markets are markets operating under the iron law of supply and demand.
It is more intuitive if you spend more time in academia / higher education. Even around me, with a bunch of mid-lower tier schools, PhDs and masters are being pumped out semester after semester in very niche areas. In all my time searching through job boards, I've probably seen a tiny handful of chip design jobs compared to the absolute boat load of things like more generalized engineers, business analysts, financial/accounting, and so on.
Yep, just because something is difficult doesn't mean it's valuable. E.g. a lot of art is extremely difficult to create but not desired by anyone other than the artist. The value is not proportional to the difficulty.
Likewise just because something is potentially worthwhile doesn't mean it's valuable in a specific instance. E.g. drinking water is required for survival and thus extremely worthwhile, but so cheap that restaurants give it away for free when you order anything else.
If you can do ASIC design, you can probably do FPGA design. FPGA design is a pretty easy thing to do remotely, it's not ALL that different from software work. I know a bunch of consultants who make very good money living where they desire, working from home, doing FPGA design for industrial customers. Granted, these people aren't fresh college graduates since they're experienced engineers, but there are a good number of them out there.
It's quite hard to get motivated to study anything for a career when everything is posed to be replaced by AI. Not saying that is what will happen, but it seems daunting to spend 5-6 years getting into chip design only to be replaced...
The chip industry obviously isn't struggling to attract the next generation, as can be seen by the fact their salaries in the article are about in line with everyone else's. I've worked in industries that struggled to attract people, it looked like 2x, 3x, 4x and 5x salaries for what a given skill-set would reasonably be valued at and they still sometimes have to hire unqualified people. And train them, shock & horror.
These articles crop up from time to time and I really don't think the frame is hitting a high standard of thoughtfulness. Or it is a surreptitious attempt to influence government policy in which case fair enough. The problem with the chip industry is that Asia has a comparative advantage at it, probably because that is where all the advanced industrial capital investment is happening. It has nothing to do with people - it is usual for these matters to turn out to be a regulatory problem when the superficial issues are peeled back.
>The problem with the chip industry is that Asia has a comparative advantage at it, probably because that is where all the advanced industrial capital investment is happening
Asia has a comparative advantage because labor is much cheaper there. PHDs working at TSMC in Taiwan make less than fresh grads working in software in the US.
> Asia has a comparative advantage because labor is much cheaper there.
You started the sentence with comparative advantage then transitioned smoothly to an absolute advantage - if you want to see low wages, look to Africa. Relative wage levels doesn't determine directly cause comparative advantages. I also note you're also implicitly disagreeing with the article and suggesting that compensation gap is not a myth.
At least in the UK it just doesn't pay enough vs. the effort required, it also seemed like there were also a lot less jobs going compared to software.
The article shows some pay figures but those are American where everything pays insanely well compared to here so I'm not sure how relevant those are.
It's odd that it doesn't discuss the size of the industry though - I always thought of it as a small, relatively niche industry compared to software dev. and while there is probably less competition for that smaller amount of jobs, there's still a smaller amount of jobs.
I think it's a mistake to draw a hard line between "systems" software engineers and hardware engineers - e.g. I would see myself as a software engineer, but an pretty comfortable with HDL, or even modifying it, even if my primary job was on the driver side.
And also as a Brit who has since got a green card in the US a couple of years ago - I'm not sure if the UK is a great comparison point - brexit decimated that level of engineering in the UK - the sort of companies that need that expertise are at the multinational level or have fabs, none of which is really true for the UK isolated from Europe. Every company I worked for/with in the UK has either moved offshore or completely closed down since. And low demand causes low wages.
And then goes to show how a SW engineer makes 30% more. Where’s the myth? Especially given that EE requires a lot more work so that 30% gap is actually worse.
My degree is in EE. I do software now. In the groups that I've worked in at large and small companies, the talented software people have always been paid better than the talented hardware people. But, on the consulting side for experienced engineers, the hardware people seem to get the same or better pay than the software people.
I’m a SW engineer at Arm, and as far as I know (and what I see on levels.fyi) the hardware guys in the same level (e.g. Staff vs Staff) make about the same as I do. The apparent difference is probably more caused by the type of companies that employ lots of hardware engineers. Chip designer at google probably makes about the same as a software engineer at google, but both make way more than the chip designer at Intel. But Google employs mostly SW people and Intel employs mostly HW people.
"It is hard to make any sweeping conclusions that software pays more than hardware, or vice versa. The myth of software always being more lucrative may be unfounded, but the sentiment does exist among young professionals looking to choose career paths."
I.e. there surely exist hardware positions that pay better than some software position.
One has to be careful with statistics, you can't start with the aggregate and predict the individual case.
You are too tied to the whims of the capital to risk going down such a specialized field where you need multibillion dollar worth of machinery to get started, I would say.
Simpler explanation: it’s grueling work that doesn’t pay market rates.
Intel has been the semiconductor industry standard for comp for years, and they’ve lagged in this department for a decade. This has depressed the whole industry’s pay as so many chip companies view Big Blue as the market linchpin to set comp off of.
The exception here is for Apple and nVidia - where the work is still pretty grueling, but the pay is excellent.
You gotta be the cream of the crop to get in either of those places. The hiring bar for both is high, and getting higher.
Source: me. I run a job board for FPGA and RTL engineers.
The is THE limiting factor of many fields. Bioscience etc. even car mechanics for that matter.
It all starts to click when you have a practical and physical application.
A garage is readily available and not expensive.
A lab on the other hand… imo there should be more low cost options to produce chips. There are a few projects trying to do that.
Some universities have some cooperation with fabs for dead/spare space.
I don't understand why they don't make certain studies cheaper or free for students that qualify, or even give scholarships. It's very important to nudge more students towards politically or economically vital studies.
If it's economically vital, why doesn't it pay as well as industries competing for similar graduates? (not just programming, but also finance sucks up a lot of mathematically inclined people)
I'm reminded of COVID where the most "essential" workers inevitably meant the most expendable.
People make value arguments but these are incredibly naive. Look at the companies with the top market caps. They'd tumble if TSMC or ACML went under.
So there's something else to how price gets defined. Be careful to think the status quo is "rational". Sure, there's an explanation but an explanation is not a good explanation. Usually it's a simple explanation to an incredibly complex topic which no one really can explain
Why do you assume they must survive? Here's an analogy: if the food stops growing, everybody dies. That analogy is a bit extreme but at this time Intel (who just laid of 20% of its staff) and Samsung are the only meaningful competitors and TSMC produces the lion's share of chips. If TSMC goes under Apple, Nvidia, Microsoft, Google, and Meta are all also going to tank too. Those companies don't really have competitors. It's highly centralized and yes I know that AMD exists and they'll tank too
Hmmm what is a simple "hello world" project in chip design?
In computer science courses, that's as simple as a println().
In machine learning courses, that's training on mnist dataset to do character recognition.
In electrical engineering, that's buying a raspberry pi to blink led.
In chip design ... Chatgpt says to design a 1-bit full adder using verilog?
...
I understand why the article thinks the market is looking for graduate education. To design a simple chip requires an *initial investment* (as with all hardware startups really). This is different from software where one can simply launch a web app with a container hosted on your preferred cloud provider...
... That said, with the rise of LLMs lowering the barrier of entry of software even lower (e.g. vibe coding), may we see more rise of hardware startups/innovations?
FPGA dev boards are cheap nowadays, and you can start coding in a hardware definition language with a simulator. The ChatGPT answer of doing a 1-bit full adder as "hello world" makes sense.
You are obviously not going to etch silicon at home, but the design part is rather accessible as far as hardware goes.
You can absolutely etch silicon at home. Processes like wet etching (KOH, HF), reactive ion etching (RIE), laser ablation, and even electron beam lithography using repurposed CRTs are all viable at the DIY scale.
They're not used in high-volume manufacturing (you’re not replacing ASML), but they’re solid for prototyping, research, and niche builds.
Just don’t underestimate the safety aspect—some of these chemicals (like HF) are genuinely nasty, and DIY high voltage setups can bite hard.
You're not hitting nanometer nodes, but for MEMS, sensors, and basic ICs, it’s totally within reach if you know what you’re doing.
> rise of LLMs lowering the barrier of entry of software even lower
Getting to your first wafer costs something like $250k and upwards of fab costs, depending on what process you're using. Hence much of chip design effort is already spent on verification, it's probably over 50% by now. This is the exact opposite of vibes because mistakes are expensive.
Businesswise it's quite tough B2B sales because you're selling into other people's product development pipelines. They need to trust you because you can sink their project, way over and above the cost of the actual parts.
Edit: I cannot emphasise enough how much more conservative the culture is in chip design and EE more broadly. It belongs to a world not just before "vibe coding" but before "web 2.0". It's full of weird closed source very expensive tooling, and is built on a graveyard of expensive mistakes. You've got to get the product 100% right on the first go.
Well, maybe the second go, production silicon is usually "B" rev. But that's it. Economics dictate you then need to be able to sell that run for a few years before replacing it with an upgraded product line.
The rule of thumb I use for chip design is that verification takes at least 2/3s of development. Sometimes more. 50% would be nice but I think is optimistic
Verification is indeed the majority of the time spent. Unlike programming, Verilog and VHDL and higher level things like Chisel aren’t executed serially by the hardware they describe like a von Neumann machine. Hello World for a chip isn’t designing the circuit, or simulating the circuit, or synthesizing the circuit to some set of physical primitives. No, it’s proving that the circuit will behave correctly under a bunch of different conditions. The less commoditized the product, the more important it is to know the real PDK, the real standard cell performance, what to really trust from the foundry, etc. Most of the algorithms to assist in this process are proprietary and locked behind NDAs. The open source tools are decades behind the commercial ones in both speed and correctness, despite heavy investment from companies like Google.
And so my point: the place where people best know how to make chips competitively in a cutthroat industry is NOT in schools, but in private companies that have signed all the NDAs. The information is literally locked away, unable to diffuse into the open where universities efficiently operate. Professors cannot teach what they don’t know or cannot legally share.
Chip design is a journeyman industry. Building fault-tolerant, fast, power-efficient, correct, debuggable, and manufacturable designs is table stakes. Because if not, there are already a ton of chip varieties available. Don’t reinvent the wheel because the intersection of logic, supply chain logistics, circuit design, large scale multi objective optimization, chemistry, physics, materials science, and mathematical verification is unforgiving.
I think you would just buy a cheap FPGA board and use that wouldn't you? No need to do a full chip until you know what you are doing. That would be like building a server farm just to do your software hello world
In the software field, people with no education or education from a country with a low standard are normally a pain in the ass to work with.
And in hardware mistakes are more costly, while in software most of the developers work on completely useless projects that are doomed to disappear soon.
Because we all eager for blog posts how someone discovered how to use transitors to create a memory module, with a 555 to keep the refresh clock going.
In this article, we will discuss 6 reasons why the chip industry is struggling to attract new talent.
Theory-first education: In an effort to build from fundamentals, there is too much emphasis on theory rather than a focus on applications.
Compensation myth: There is a feeling that software pays more than hardware. Reality is not so cut and dry.
Graduate degrees: A lot more employers ask for graduate level degrees to enter chip design creating bottlenecks in talent supply.
Early specialization: Highly niche skillsets are less marketable and career limiting.
Documentation shortages: Hardware design is entirely tribal knowledge and hard to self-learn.
Chip design culture: Hardware companies have a retro feel to them, deadlines are tight, and mistakes are deadly.
EEs also have an awful "holier-than-thou" attitude in Engineering.
And then you make it only worse by requiring "Masters only or above". Well guess why, because your graduation was spent going around stuff that goes from nowhere to nowhere else.
"Graduate degrees" listed as a reason.
Yes, designing chips is hard, it takes a lot of knowledge. This is why medical doctors need to go through all that schooling... designing a tiny chip with more transistors running software that does amazing things is very difficult.
My Ph.D. is in computer engineering, specifically VLSI and chip design. This was from a few years ago. I _probably_ should have gone into industry, I mean, after all, it is what I went to school for and wanted to do. However, the starting salary for a chip designer (Intel / AMD / HP / IBM) was literally less than I was making at a side job (I worked my way through my Ph.D) as an IT sysadmin. Not only that, people that I knew well that graduated before me would call me up and tell me it was worse than hell itself. 80 hour weeks? Completely normal, outside of the 2 hours of commute time. Barely make rent because you live in California? Check. Pages / Calls all hours of the day outside of work? Check. 80 hours? You mean 100 hours a week leading up to a release, right? Check.
Looking back on it, it seems this was "the challenging" and if you made it past this (something like 5 years on) things calmed down for a chip designer and you moved into a more "modest" 60-80 hours a week role with less pressure and somewhat of a pay increase.
Yes, how do you attract talent under those conditions? It is not flashy work, takes a lot of schooling and the rewards are low. At least medical doctors can kind of look forward to "well, I can make _real_ money doing this", and have the satisfaction of "I helped a lot of people".
I have never done any chip design work, but I have seen some hobbyists going on with HDL without a degree. It's definitely not professional level but I suspect they are hireable materials at least. This leads me to ask the two questions:
1) If chip design (or X, anything) is so vital, so important to national security, why do universities insist that a degree of X include a lot of unrelated courses? You can argue that universities are not just for employment (yeah, as if most people go to university just for fun), but by the name of God, I really hate it when my university forced me to go through all those BS selective courses to reach 120. If you ask me, it's just money grabbing.
2) Why can't students go straight to a fab or whatever after bachelor and do their masters THERE? Isn't the industry a much better place to do that? Actually, why don't the industry simply hire high school students and go from there? Companies used to do that in the 50s/60s. I don't know if they still do that but I think it's rare.
Are you asking why college isn’t a vocational school or technical school? Or are you wondering why the USA doesn’t have apprenticeships in the way Germany and others have?
A college isn’t really meant to be a narrow tracked vocational school - but it’s fair to ask why aren’t their more vocational schools for high tech fieldss
For number 1, there really does need to be some wider reforms, but I fear that won't happen until the whole student loan paradigm crashes and burns. I had a couple of fully paid semesters that were 100% electives I had no interest in. I would have preferred to graduate a year earlier (and thus, a year richer) or take courses that were actually relevant. Problem is that universities talk out both sides of their mouth -- raking in huge amounts of cash like a corporation, feasting off guaranteed loans from 18 year olds, while espousing some nobler concept of education/enlightenment on the other hand.
These are horrendously difficult questions, though a partial answer to 2) is that labor (with "Baumol" training costs factored in) was so cheap up till the early 60s that high schoolers were easily competitive with college grads..
https://en.wikipedia.org/wiki/Baumol_effect#Education
Easier question to answer:
Even easier question to answer:Ah, thanks for the link. I do hope that labor gets more share in the revenue. I kinda love the idea of "cooperative corporations" where shares are more or less evenly shared by the employees (there are ofc differences between different levels, and I'd imagine a large amount of stocks are "frozen"). I think it's going to be a lot harmonic for all employees (even the managers).
Note the original paper by Baumol and Bowen:
"On the Performing Arts: The Anatomy of Their Economic Problems" (1965)
https://sci-hub.ru/https://www.jstor.org/stable/1816292
>It is largely for this reason that performing arts organizations in financial difficulty have often managed to shift part of their financial burden back to the performers--and to the managements, who also are generally very poorly paid by commercial standards.
Imho the paper continues to function after you patch it with your favorite calling, design-driven industry or organizational structure :)
I know of a bunch of people, some who only have high school degrees, who are entirely self-taught and are doing tinytapeout (https://tinytapeout.com/) chip designs. Yeah, that's not nearly at the skill or scale of designing CPUs for Intel/AMD/NVIDIA/Apple/Marvell/etc but it's still chip design!
Your concerns about horribly long hours and lower than IT/software pay are the most concerning part to me. But, if there's really a shortage of engineers who know how to do chip design, hopefully the market will take care of that via supply/demand at least once things get really out of whack.
It's always crazy to me that you can build crappy websites in the most flexible environment imaginable and make way more than those doing the actually deep and challenging work required for those websites to run in the first place.
Fundamentally, in programming the programmers have the means to compete with their employers if they want to. Shifts power a lot.
Every time I read a post like this I'm thankful the market all but forced me into software.
These truths mostly also apply when answering the eternal* question:
* Snow Crash (1992) might turn out not to be so precisely prescient due to upcoming dedollarization, AI democratization/bubble burst (the exact option depends on your personality type), & the solid state battery boom:>When it gets down to it — talking trade balances here — once we've brain-drained all our technology into other countries, once things have evened out, they're making cars in Bolivia and microwave ovens in Tadzhikistan and selling them here — once our edge in natural resources has been made irrelevant by giant Hong Kong ships and dirigibles that can ship North Dakota all the way to New Zealand for a nickel — once the Invisible Hand has taken away all those historical inequities and smeared them out into a broad global layer of what a Pakistani brickmaker would consider to be prosperity — y'know what? There's only four things we do better than anyone else:
music
movies
microcode (software)
high-speed pizza delivery
Looks like some kind of a gatekeeping ritual to rationalize why upper management salaries are in millions. I think we can also see in other industries too.
You hire hundreds of interns and entry level workers to let them fight in the bloodbath for 100h a week. Pay peanuts. Let them do all the work.
The ones who survive get a bit bigger salaries. Those who still persist in upper level bloodbaths are upgraded into millionaires. And paying them millions looks acceptable as it is so hard to reach the top.
While you clearly could share all those millions between entry level and paid internships, don't have 100h weeks and have a healthy industry.
The job is too niche. This means there are very few employers worldwide that can make use of your experience. That actually puts downwards pressure on your salary (the employer is a quasi monopsony), not to mention that you are tied to a few locations.
I'm surprised if a chip expert wouldn't be very employable by various downstream businesses?
The reason software pays so well is that there’s a lot of demand by a lot of companies for the best in a global market AND the enough of the best can counter those offers by going to VCs to fund their own company. EE market is not as global and much more difficult to fund. The only ones paying the biggest salaries are the big companies
I this strikes me as less than half the reason, though still a very big factor.
If I had to guess it would be that every business needs (or thinks they need) a custom website, and often some custom software processes, and definitely their own custom network and IT, and likely some slightly customized ERP. But rarely, very very rarely, do any businesses want custom chips or computers such that a CE is required. The demand for software is so high.
> The demand for software is so high.
What is interesting in these kinds of discussions is that intuitively, nobody really beliefs in the law of supply & demand. On an emotional level, people always think that effort and difficulty should be rewarded on their own - which is actually correct from a social and ethical POV, and yet:
markets are markets operating under the iron law of supply and demand.
It is more intuitive if you spend more time in academia / higher education. Even around me, with a bunch of mid-lower tier schools, PhDs and masters are being pumped out semester after semester in very niche areas. In all my time searching through job boards, I've probably seen a tiny handful of chip design jobs compared to the absolute boat load of things like more generalized engineers, business analysts, financial/accounting, and so on.
> It is more intuitive if you spend more time in academia / higher education.
And this is precisely why credentialism and large parts of academia must go. The incentives are perverse.
Yep, just because something is difficult doesn't mean it's valuable. E.g. a lot of art is extremely difficult to create but not desired by anyone other than the artist. The value is not proportional to the difficulty.
Likewise just because something is potentially worthwhile doesn't mean it's valuable in a specific instance. E.g. drinking water is required for survival and thus extremely worthwhile, but so cheap that restaurants give it away for free when you order anything else.
Employable by multiple downstream businesses able and willing to pay high prices in the location you want to live is the full parameter.
If you can do ASIC design, you can probably do FPGA design. FPGA design is a pretty easy thing to do remotely, it's not ALL that different from software work. I know a bunch of consultants who make very good money living where they desire, working from home, doing FPGA design for industrial customers. Granted, these people aren't fresh college graduates since they're experienced engineers, but there are a good number of them out there.
It's quite hard to get motivated to study anything for a career when everything is posed to be replaced by AI. Not saying that is what will happen, but it seems daunting to spend 5-6 years getting into chip design only to be replaced...
The chip industry obviously isn't struggling to attract the next generation, as can be seen by the fact their salaries in the article are about in line with everyone else's. I've worked in industries that struggled to attract people, it looked like 2x, 3x, 4x and 5x salaries for what a given skill-set would reasonably be valued at and they still sometimes have to hire unqualified people. And train them, shock & horror.
These articles crop up from time to time and I really don't think the frame is hitting a high standard of thoughtfulness. Or it is a surreptitious attempt to influence government policy in which case fair enough. The problem with the chip industry is that Asia has a comparative advantage at it, probably because that is where all the advanced industrial capital investment is happening. It has nothing to do with people - it is usual for these matters to turn out to be a regulatory problem when the superficial issues are peeled back.
>The problem with the chip industry is that Asia has a comparative advantage at it, probably because that is where all the advanced industrial capital investment is happening
Asia has a comparative advantage because labor is much cheaper there. PHDs working at TSMC in Taiwan make less than fresh grads working in software in the US.
> Asia has a comparative advantage because labor is much cheaper there.
You started the sentence with comparative advantage then transitioned smoothly to an absolute advantage - if you want to see low wages, look to Africa. Relative wage levels doesn't determine directly cause comparative advantages. I also note you're also implicitly disagreeing with the article and suggesting that compensation gap is not a myth.
Everything hardware on all levels pays significantly less than software, be it US, UK or Asia.
At least in the UK it just doesn't pay enough vs. the effort required, it also seemed like there were also a lot less jobs going compared to software.
The article shows some pay figures but those are American where everything pays insanely well compared to here so I'm not sure how relevant those are.
It's odd that it doesn't discuss the size of the industry though - I always thought of it as a small, relatively niche industry compared to software dev. and while there is probably less competition for that smaller amount of jobs, there's still a smaller amount of jobs.
I think it's a mistake to draw a hard line between "systems" software engineers and hardware engineers - e.g. I would see myself as a software engineer, but an pretty comfortable with HDL, or even modifying it, even if my primary job was on the driver side.
And also as a Brit who has since got a green card in the US a couple of years ago - I'm not sure if the UK is a great comparison point - brexit decimated that level of engineering in the UK - the sort of companies that need that expertise are at the multinational level or have fabs, none of which is really true for the UK isolated from Europe. Every company I worked for/with in the UK has either moved offshore or completely closed down since. And low demand causes low wages.
> The Hardware vs. Software Compensation Myth
And then goes to show how a SW engineer makes 30% more. Where’s the myth? Especially given that EE requires a lot more work so that 30% gap is actually worse.
My degree is in EE. I do software now. In the groups that I've worked in at large and small companies, the talented software people have always been paid better than the talented hardware people. But, on the consulting side for experienced engineers, the hardware people seem to get the same or better pay than the software people.
I’m a SW engineer at Arm, and as far as I know (and what I see on levels.fyi) the hardware guys in the same level (e.g. Staff vs Staff) make about the same as I do. The apparent difference is probably more caused by the type of companies that employ lots of hardware engineers. Chip designer at google probably makes about the same as a software engineer at google, but both make way more than the chip designer at Intel. But Google employs mostly SW people and Intel employs mostly HW people.
No, they don't.
"It is hard to make any sweeping conclusions that software pays more than hardware, or vice versa. The myth of software always being more lucrative may be unfounded, but the sentiment does exist among young professionals looking to choose career paths."
I.e. there surely exist hardware positions that pay better than some software position.
One has to be careful with statistics, you can't start with the aggregate and predict the individual case.
You are too tied to the whims of the capital to risk going down such a specialized field where you need multibillion dollar worth of machinery to get started, I would say.
Simpler explanation: it’s grueling work that doesn’t pay market rates.
Intel has been the semiconductor industry standard for comp for years, and they’ve lagged in this department for a decade. This has depressed the whole industry’s pay as so many chip companies view Big Blue as the market linchpin to set comp off of.
The exception here is for Apple and nVidia - where the work is still pretty grueling, but the pay is excellent.
You gotta be the cream of the crop to get in either of those places. The hiring bar for both is high, and getting higher.
Source: me. I run a job board for FPGA and RTL engineers.
Theory first
The is THE limiting factor of many fields. Bioscience etc. even car mechanics for that matter.
It all starts to click when you have a practical and physical application. A garage is readily available and not expensive.
A lab on the other hand… imo there should be more low cost options to produce chips. There are a few projects trying to do that. Some universities have some cooperation with fabs for dead/spare space.
I don't understand why they don't make certain studies cheaper or free for students that qualify, or even give scholarships. It's very important to nudge more students towards politically or economically vital studies.
> politically or economically vital studies.
If it's economically vital, why doesn't it pay as well as industries competing for similar graduates? (not just programming, but also finance sucks up a lot of mathematically inclined people)
I'm reminded of COVID where the most "essential" workers inevitably meant the most expendable.
What other industries?
People make value arguments but these are incredibly naive. Look at the companies with the top market caps. They'd tumble if TSMC or ACML went under.
So there's something else to how price gets defined. Be careful to think the status quo is "rational". Sure, there's an explanation but an explanation is not a good explanation. Usually it's a simple explanation to an incredibly complex topic which no one really can explain
>Look at the companies with the top market caps. They'd tumble if TSMC or ACML went under.
No they wouldn't because their competitors would be in the same situation as them. If nobody has advanced chips, nobody has a competitive advantage.
Why do you assume they must survive? Here's an analogy: if the food stops growing, everybody dies. That analogy is a bit extreme but at this time Intel (who just laid of 20% of its staff) and Samsung are the only meaningful competitors and TSMC produces the lion's share of chips. If TSMC goes under Apple, Nvidia, Microsoft, Google, and Meta are all also going to tank too. Those companies don't really have competitors. It's highly centralized and yes I know that AMD exists and they'll tank too
Never mind "competitive advantage" — if nobody can make cutting edge chips, most (or all?) of them no longer have a product to sell.
It's the diamond/water paradox. Diamonds are expensive but water is cheap, even though water is much more important than diamonds.
https://en.wikipedia.org/wiki/Paradox_of_value
Because capitalism doesn't work like that?
Hmmm what is a simple "hello world" project in chip design?
In computer science courses, that's as simple as a println().
In machine learning courses, that's training on mnist dataset to do character recognition.
In electrical engineering, that's buying a raspberry pi to blink led.
In chip design ... Chatgpt says to design a 1-bit full adder using verilog?
...
I understand why the article thinks the market is looking for graduate education. To design a simple chip requires an *initial investment* (as with all hardware startups really). This is different from software where one can simply launch a web app with a container hosted on your preferred cloud provider...
... That said, with the rise of LLMs lowering the barrier of entry of software even lower (e.g. vibe coding), may we see more rise of hardware startups/innovations?
FPGA dev boards are cheap nowadays, and you can start coding in a hardware definition language with a simulator. The ChatGPT answer of doing a 1-bit full adder as "hello world" makes sense.
You are obviously not going to etch silicon at home, but the design part is rather accessible as far as hardware goes.
You can absolutely etch silicon at home. Processes like wet etching (KOH, HF), reactive ion etching (RIE), laser ablation, and even electron beam lithography using repurposed CRTs are all viable at the DIY scale.
They're not used in high-volume manufacturing (you’re not replacing ASML), but they’re solid for prototyping, research, and niche builds.
Just don’t underestimate the safety aspect—some of these chemicals (like HF) are genuinely nasty, and DIY high voltage setups can bite hard.
You're not hitting nanometer nodes, but for MEMS, sensors, and basic ICs, it’s totally within reach if you know what you’re doing.
> You are obviously not going to etch silicon at home
No? https://www.youtube.com/watch?v=IS5ycm7VfXg
> rise of LLMs lowering the barrier of entry of software even lower
Getting to your first wafer costs something like $250k and upwards of fab costs, depending on what process you're using. Hence much of chip design effort is already spent on verification, it's probably over 50% by now. This is the exact opposite of vibes because mistakes are expensive.
Businesswise it's quite tough B2B sales because you're selling into other people's product development pipelines. They need to trust you because you can sink their project, way over and above the cost of the actual parts.
Edit: I cannot emphasise enough how much more conservative the culture is in chip design and EE more broadly. It belongs to a world not just before "vibe coding" but before "web 2.0". It's full of weird closed source very expensive tooling, and is built on a graveyard of expensive mistakes. You've got to get the product 100% right on the first go.
Well, maybe the second go, production silicon is usually "B" rev. But that's it. Economics dictate you then need to be able to sell that run for a few years before replacing it with an upgraded product line.
The rule of thumb I use for chip design is that verification takes at least 2/3s of development. Sometimes more. 50% would be nice but I think is optimistic
Verification is indeed the majority of the time spent. Unlike programming, Verilog and VHDL and higher level things like Chisel aren’t executed serially by the hardware they describe like a von Neumann machine. Hello World for a chip isn’t designing the circuit, or simulating the circuit, or synthesizing the circuit to some set of physical primitives. No, it’s proving that the circuit will behave correctly under a bunch of different conditions. The less commoditized the product, the more important it is to know the real PDK, the real standard cell performance, what to really trust from the foundry, etc. Most of the algorithms to assist in this process are proprietary and locked behind NDAs. The open source tools are decades behind the commercial ones in both speed and correctness, despite heavy investment from companies like Google.
And so my point: the place where people best know how to make chips competitively in a cutthroat industry is NOT in schools, but in private companies that have signed all the NDAs. The information is literally locked away, unable to diffuse into the open where universities efficiently operate. Professors cannot teach what they don’t know or cannot legally share.
Chip design is a journeyman industry. Building fault-tolerant, fast, power-efficient, correct, debuggable, and manufacturable designs is table stakes. Because if not, there are already a ton of chip varieties available. Don’t reinvent the wheel because the intersection of logic, supply chain logistics, circuit design, large scale multi objective optimization, chemistry, physics, materials science, and mathematical verification is unforgiving.
I think you would just buy a cheap FPGA board and use that wouldn't you? No need to do a full chip until you know what you are doing. That would be like building a server farm just to do your software hello world
I guess open something like simulator.io and start designing?
The one from Sebastian Lague is great too: https://sebastian.itch.io/digital-logic-sim
In chip design it is using languages like verilog that don't mean what they actually mean and just confuses everybody
If they paid new grads $200,000/year I can guarantee you that they would not have any issues attracting new talent.
Summary of six key reasons including an insistence on requiring graduate education.
In the software field, people with no education or education from a country with a low standard are normally a pain in the ass to work with.
And in hardware mistakes are more costly, while in software most of the developers work on completely useless projects that are doomed to disappear soon.
Because we all eager for blog posts how someone discovered how to use transitors to create a memory module, with a 555 to keep the refresh clock going.
In this article, we will discuss 6 reasons why the chip industry is struggling to attract new talent.
Theory-first education: In an effort to build from fundamentals, there is too much emphasis on theory rather than a focus on applications. Compensation myth: There is a feeling that software pays more than hardware. Reality is not so cut and dry. Graduate degrees: A lot more employers ask for graduate level degrees to enter chip design creating bottlenecks in talent supply. Early specialization: Highly niche skillsets are less marketable and career limiting. Documentation shortages: Hardware design is entirely tribal knowledge and hard to self-learn. Chip design culture: Hardware companies have a retro feel to them, deadlines are tight, and mistakes are deadly.
Agree a lot with the article
Especially 1 and 3
EE education has absolute dog-crap didacticism.
EEs also have an awful "holier-than-thou" attitude in Engineering.
And then you make it only worse by requiring "Masters only or above". Well guess why, because your graduation was spent going around stuff that goes from nowhere to nowhere else.