The other day I was watching Dave Jones, a video blogger that I find entertaining and informative. His blog, the EEVblog, is catnip for nerds who like to solder stuff and use oscilloscopes.
Recently he did a short segment where he answered a question from a student who was upset that his teacher told him that EE was perhaps not a great field for job security, and he sort of went on a colorful rant about how wrong the professor is.
The professor is right.
Electrical engineering employment is indeed in decline, at least in the USA, and I suspect, other development countries. It’s not that EE skills are not helpful, or that understanding electronics, systems, signals, etc, are not useful. They are all useful and will continue to be. But I think more and more of the work, in particular, the high paying work, will migrate to software people who understand the hardware “well enough.” Which is fine. The fact is that EEs make good firmware engineers.
I think someone smart, with a solid EE background and a willingness to adapt throughout your entire career, should always find employment, but over time I suspect it will be less and less directly related to EE.
I mostly know Silicon Valley. Semiconductor employment is way down here. Mostly, it is through attrition, as people retire and move on, but nobody is hiring loads of young engineers to design chips anymore. It makes sense. Though chip volumes continue to grow, margins continue to shrink, and new chip design starts are way down, because “big” SOCs (systems on chip) with lots of peripherals can fill many niches that used to require custom or semi-custom parts.
I suspect that the need for EEs in circuit board design is also in decline. Not because there are fewer circuit boards, but because designing them is getting easier. One driver is the proliferation of very capable semiconductor parts with lots of cool peripherals is also obviating a lot of would-have-been design work. It’s gotten really easy to plop down a uC and hook up a few things over serial links and a few standard interfaces. In essence, a lot of board design work has been slurped into the chips, where one team designs it once rather than every board designer doing it again. There might be more boards being designed than ever, but the effort per board seems to be going down fast, and that’s actually not great for employment. Like you, I take apart a lot of stuff, and I’m blown away lately not by how complex many modern high volume boards are, but how dead simple they are.
The growth of the “maker” movement bears this out. Amateurs, many with little or no electronics knowledge, are designing circuit boards that do useful things, and they work. Are they making mistakes? Sure, they are. The boards are often not pretty, and violate rules and guidelines that any EE would know, but somehow they crank out working stuff anyway.
I do hold out some hope that as Moore’s law sunsets — and it really is sunseting this time — there will be renewed interest in creative EE design, as natural evolution in performance and capacity won’t solve problems “automatically.” That will perhaps mean more novel architectures, use of FPGAs, close HW/SW codesign, etc.
Some statistics bear all this out. The US Bureau of Labor Statistics has this to say about the 2014-2024 job outlook for EEs:
http://www.bls.gov/ooh/architecture-and-engineering/electrical-and-electronics-engineers.htm#tab-6
Note that over a 10 year period they are predicting essentially no growth for EE’s at all. None. Compare this to employment overall, in which they predict 7% growth.
One final note. People who love EE tend to think of EEs as the “model EE” — someone clever, curious, and energetic, and who remains so way for 40+ years. But let’s remind ourselves that 1/2 of EEs are below median. If you know the student in question, you can make an informed assessment about that person’s prospects, but when you are answering a generic question about prospects for generic EEs, I think the right picture to have in mind is that of the middling engineer, not a particularly good one.
I’m not saying at all that EE is a bad career, and for all I know the number of people getting EE degrees is going down faster than employment, so that the prospects for an EE graduate are actually quite good, but it is important for students to know the state of affairs.
I think you are generalizing. EE is a broad field. I agree some sections might be declining in the USA, but some are actually flourishing. I myself, as an EE, design HW, write FW and most of the time I give advice to my SW colleagues on how to write their codes to better interact with the HW. Getting an EE degree means you do many fun thins from HW to low level SW design. there are other fields such as Power Systems that are growing thanks to new battery powered devices and cars. You mentioned that HW has become so easy that anyone can put it together and it works. I can say that in silicon valley, completing a three months SW boot camp will make a SW engineer. Also getting to a reputable EE program is much harder than getting in to a computer science program in any state. And last, who designs those chips with many peripherals? an EE graduate.
I am not generalizing, so much as aggregating. It is a statistical fact that EE is in decline, and enrollment in EE programs is also in decline. The chips needs designers, and the designs of the chips are getting more complex, but the number of unique chips is dropping from what it was 20 or 30 years ago. Dropping a LOT.
A lot of EE work is being done by people who are EE-adjacent like SW folks with tinkery inclinations. There will always be tricky stuff that requires people who know what they are doing, but for a given design, the fraction of the design that requires a “real” EE is going down, and many designs don’t require the real EE at all, unless they get in trouble. (signal integrity, etc).
Not sure where you get your statistical data. Check out the latest US Occupational Outlook from Bureau of Labor Statistics (see the link below). As I mentioned previously, EE is a diverse field. One could be designing 100KV power lines and the other might be writing low level system Firmware using C language. Moreover, chips getting complex does not mean less of an expertise needed. It’s still an analog design packed in a smaller space. Finally, not true that SW folks could take the task of designing the hardware, while I have experienced that the opposite is true. Many EEs are capable of taking the task of designating system software and write code, it’s natural to most EEs. To be honest, I have worked with a few SW folks on an FPGA design. They were good writing Verilog code, but had a hard time explaining to them that what they are developing is a piece of hardware, events can execute concurrently in parallel, opposite to C code that executes sequentially line after line. I agree that there are more jobs posted for software developers, but there are more of them around as well. That said both fields are valuable. In a system, software and hardware have to work together in harmony.
https://www.bls.gov/ooh/architecture-and-engineering/home.htm
The source of my stats was the BLS, as linked in the original piece. At the time, their latest numbers were for the 2014-24 period and showed negative growth. Their newer stats, for 2018-28 predict 2% growth, which is below average for most job categories and also below the rate of growth of the economy overall, suggesting that EE is becoming a smaller part of the overall economy over time.
I don’t think you understood my other points at all. Nobody said EE wasn’t important or varied or interesting or wonderful. I said it was in decline, as a career. It is.
Yes, ICs are getting more complex, but there are many, many fewer of them. This means less work for EEs. Also, system-on-chips can cover large market swathes because they have so many peripherals on them. As a result, those peripherals are now integrated by the IC designer, making board-level designs with them *much simpler*. Even the interfaces themselves tend to be self-clocked serial connections — easier to design for than wide synchronous buses. As a result, you need a lower caliber of engineer to design products than you once did. Or look at analog design. Again, there’s just less if it needed because so much can be done in the digital domain where is becomes software. These are just facts. Less skilled folks can do board-level work that used to require real EEs.
As for software people doing EE work, it is happening all the time. They may do it badly, but they do it adequately. How do I know? Because silicon valley is full of folks doing just this.
Yes, EE is still important and varied, but as a career choice, it’s less appealing than it once was.
I really appreciate this post. I have been struggling to figure out if i should double major in Mathematics and Electrical engineering or Mathematics and Computer Science. From this information and a quick jobs search I have decided the Computer Science will go better with my mathematics degree. As much as I like EE i dont want to gamble with my future knowing that im a parent so thank you again for the post.