Nvidia is now worth $102M per employee
sherwood.newsBelieve it or not, elite universities are reaching similar heights. MIT is at roughly $20M endowment per faculty member.
All caveats apply. 20≠100, employees might include staff and grad students, etc.
Still, I would never advise anyone to donate to any institution like this.
Looks like Princeton tops the Ivy League (and perhaps all universities?) at $27M per faculty member. I often like to call it "a hedge fund with hobbies"
> Believe it or not, elite universities are reaching similar heights. MIT is at roughly $20M endowment per faculty member.
If you ran MIT, what do you think a good endowment amount per faculty member would be?
I’m not sure $20m is far off from optimal.
I might go down to $15m, with an annual draw down $750k so (5%) per year to cover burdened labor rate of average professor and certain percentage (half?) to go to “support” (i.e., everything else, including libraries, staff, maintenance, maybe grad student funding, etc.). Even that seems sort of on the tight side.
$20M might be optimal if tuition was $0 and they didn't receive any research grants.
> if tuition was $0
https://sfs.mit.edu/undergraduate-students/the-cost-of-atten...
I couldn’t find the number quickly, but tuition is not a big funding source for the university. They are quite generous with tuition grants. This might not be true for the MBA program, but probably is for many/most other programs.
> … and they didn't receive any research grants.
Yeah. This is where I think one needs to look at exactly where expenses go. Grants don’t go as far as some folks think they do.
Anyway, what’s your number with things being as they are?
I’m still ok with $15-20m, and I am not particularly fond of university admins and the way they run things.
> tuition is not a big funding source for the university
Princeton reported a paltry ~$154.5M in net tuition and fees for FY2023, which seems low ($18K/student.) They claimed $20K "net cost after aid" for students on financial aid.
https://finance.princeton.edu/report-treasurer
https://www.princeton.edu/news/2023/03/31/princeton-trustees...
Indirect cost, how I loathe thee…
I'd go down to about $2M, which is where MIT was at a quarter century ago, when it was it's prime.
Being answerable to funding sources, tuition, etc. is a good thing. Having money to soon off from is not.
An endowment is entirely different from a market capitalization.
How can you write "employees might include staff" and keep a straight face? Yes, they do, how is this even in question? MIT has 1089 faculty members but 17,180 employees total. That's a 15x difference.
Yes. Do you realize how insane this is? A colleague of mine used to have a chart of staff per faculty member at MIT versus time. That figure has gone up around 20-fold on that chart.
It's almost all bloat and overhead made possible by being over capitalized.
Nvidia is clearly overvalued, but betting against it is too risky. Nobody knows when its price will return to a reasonable value. It could be months or years.
The market can remain irrational longer than you can remain solvent.
Every employee at NVDA is basically worth more to the financial markets than any Series A startup lol
TSMC apparently doesn’t charge nearly enough for their services.
TSMC can charge as much as the market will bear. Nvidia already moved their production once to Samsung for the 3000 series GPUs just to call TSMCs bluff.
TSMC are the best in town, but they're also not the only game in town. Pretty sure Intel and Samsung are constantly tempting Nvidia with good offers, which Nvidia can use as leverage in negotiations to keep TSMC prices in check.
It does seem strange, because ok there’s competition, but there’s very limited competition – we’re talking maybe five or six chip manufacturers? And it seems they would all know the metrics of price and can ratchet upwards what they would charge Nvidia, like a cartel without officially forming a cartel.
If TSMC doubled or 10x what they charge Nvidia on Monday, Nvidia can’t switch suppliers overnight, and it would crash Nvidia’s valuation, leaving TSMC to say “ok we have the capacity for anyone who wants to take on Nvidia”.
Another side of that, while the whole “China invading Taiwan thing” isn’t exactly on the cards, there’s now $1+Trillion at stake shorting Nvidia stock alone if China felt like putting a few boots on the ground for a week.
>If TSMC doubled or 10x what they charge Nvidia on Monday
Chip manufacturing isn't like selling lemonade. You're not paying spot prices. You negociate fixed prices upfront for the duration of the manufacturing contract with various exit clauses and penalties for both sides.
>Nvidia can’t switch suppliers overnight
They can't switch overnight, but you also don't make good business by screwing over your customers for a one time hitjob just because you've locked them in for a few years. That trick will only work once. After that they'll avoid you with a 3,048 meter pole.
You stay afloat by building long term trust with your customers.
I realise I’m oversimplifying yeah, but it still seems like TSMC has a fairly huge amount of leverage that they’re not wielding.
Fascinating, for whatever reasons, I thought they were years ahead, and Nvidia’s (and apple) only option.
Thanks!
>Fascinating, for whatever reasons, I thought they were years ahead, and Nvidia’s (and apple) only option.
It's easy to get that impression based on all the recent talk about TSMC and Taiwan's importance to the global economy. My understanding is that while TSMC is currently ahead, the lead over Intel and Samsung is very slight. Intel was historically the global leading-edge producer until relatively recently, and there is no reason to think that it (or Samsung, or even someone else) cannot regain the lead. Intel is, as I understand it, the first purchaser of ASML's newest lithographic machines, in hopes of leapfrogging TSMC.
Yeah I thought this as well, guess I got the wrong impression.
I've wondered this for awhile.
Theres a very recent comment from the CEO stating something similar... But historically it's seemed they've allowed their customers to take most of the margin.
They should have a lot of leverage being on the most advanced node, but doesn't seem they use it.
Likely a cultural thing, but if somebody has more insight would be curious
We consider repeatably providing good value to customers as some kind of weakness?
Nvidia is getting ~90% margins on their cards, which are manufactured by TSMC. Nobody else can manufacture to the same levels of density/performance/efficiency. TSMC is clearly not capitalizing well on their process advantage.
Apple and Nvidia can switch to other fabs, but performance and competitiveness will be impacted. TSMC seems to be running more like a charity than a business with the amount of margin their customers get on their products (both Apple and Nvidia)
Perhaps the better question is why doesn’t ASML charge far more?
Brilliant. Thanks for posting. Actually quite a good metric for thinking about valuation, ie it’s 5x what it should be right now.
I think about it the other way. Businesses with high $/revenue per employee, high margin, etc. tend to be fundamentally better managed and more robust.
I'm not claiming that applies to Nvidia either way specifically, but it's a good sign in the generic:
1. They're efficient.
2. They've avoided the curse of overhiring (hi, Google!)
3. They can stay afloat even if business drops a lot.
But this article is about market cap not revenue. But even revenue is almost meaningless for the things you mention -- it's profit per employee which would be interesting. And for that, tech companies are nowhere near the top.
Check this https://airleasecorp.com/press/air-lease-corporation-announc... they made 753.6M in 2023 with 163 employees or 4.6M per employee. No one comes even close. No one.
I believe Fannie Mae is the second who made 17400 with 8100 people or 2.17M per employee.
To compare, nVidia made 4368 with 29600.
If you take their latest quarters results ($6B profit) you get $220k/employee per quarter or $880k/year
why bother
Probably the Sicilian mob too. So what? Criminal enterprises do not count.
Interesting. Yeah would be cool to look at profit
Or here's another perspective:
Remember the point of "AI" is getting human-like productivity out of machines.
So should it be surprising that the company at the forefront of this is has a low human-to-everything ratio?
I wonder if Nvidia would consider acquiring either Qualcomm or Mediatek. Both have very little competing against Nvidia. Gets Nvidia into the Smartphone and Wireless game. Which is crucial if they intend to bet big in ARM PC. Broadcom is way too big, Marvell is also another target but may be a little overlap in terms of product range and doesn't offer any consumer profile products.
They can't pay everyone $102M, that's the market cap (the total value of all stock shares added together). In terms of profit, it is only worth $2M.
Any nvidia employees here? How much are you worth now? ;-)
Depends on how much the company pays me :)