55% net profit doesn't include NRE right? The thing about selling fewer, bigger-ticket items is that the non-recurring engineering costs are amortized over fewer sales. Not to say they aren't printing money, but the unit cost to produce the second GPU pales in comparison to the effort to produce the first one.
How many racks are they selling? Is that 50% of their revenue? 10%? How sustainable will that be? I understand AI will probably continue to grow, but can they continue cornering such a market with 55% margins?
Fortunately, I opted to pivot towards ratio of total equity, the per-unit activity was a very rough attempt at moving away from abstractions, and that is obviously one of the many flaws in such an exercise.
I already noted the profit margins are incredibly unstable, so I don’t trust the reported figures where they quadrupled inside of a decade. I’m not suggesting it isn’t real, only that it isn’t possible to pin that 55% down as sustainable for any significant period of time, certainly not the 30-50 years is it would take to realize $4T of value at their current pace.
It's close to 90% of their revenue. They will sell about $115B this year, $180B in 2026 and $230B in 2027, with margins staying fairly constant. Their only real competitor is Broadcom, who has slightly worse margin on AI chips.
So my broader argument is that their current performance isn’t a reliable indicator for the future, as so much of their current position is circumstantial. Any investments that rely on their sustaining this sort of performance are inherently flawed as a result.
For the record, most recent graphics revenues were $14.3B, with $116B in compute/networking. [1]
So quite lopsided. That curtails my expectations even further, as this represents a lot of initial infrastructure investments whose long-term expenditures (and viability) remain to be seen.
Healthy relationship with food does not involve restriction or conscious attempts to loose weight. You eat when hungry and stop when not hungry.
The thing that makes anorexia possible (among other things) is you being able to ignore hunger. Healthy organism will instinctively eat when hungry or missing something. The instincts takes over, body produces hormones to override behavior and diet ends.
Yeah but people in this conversation and other conversations about calorie restriction, are not arguing from the standpoint of someone being healthy, and then indulging in unhealthy relationships with food. They are talking about someone who has an unhealthy relationship with food and their body, demonstrable by their excess weight, and talking about ways to correct the poor health by having a healthy relationship with their own will. You need a healthy will in order to manage weight loss due to caloric restriction.
I think a lot of people talking past each other on this topic are really just disagreeing about what healthy will power actually is. To be specific, comments along the lines of "it's not my/their fault, it's the fault of our environment, and the availability of unhealthy food".
I think this is just having an unhealthy will. I think this is also the whole divide on things like ozempic - some people view it as enabling people to have unhealthy will power. Other people view it as the only way someone can have healthy weight. I don't think either party is wrong, I think they are just talking past eachother.
School districts spend most of their money on humans.
There is a LOT of infrastructure. The schools have to be built, they have to be staffed(there is a fair bit of support staff, to keep the teachers in classrooms teaching).
From cafeteria workers and janitors to IT people and payroll, back office and grant people. High Schools might have a few thousand students that congregate there every single day, that's a small town. Our local high school's have a dedicated full-time HVAC person for instance, just to keep the insane temperatures outside from getting into the classrooms.
It doesn't say how much is for normal instruction of normal kids, how much is zoo keeping of unruly kids, how much is remedial teaching of the slow kids (or sometimes the ones who had other reasons for being behind), how much is AP teaching (or better) for the smart kids.
It just says:
Pct of Per
Pct of Public Student
Total Exp. School Amount
i. Teachers $10,124,379 32.1% 37.7% $9,914
1) buildings, air conditioning, buses and bus drivers, lunches, “resource officers”, the second(!) assistant superintendent’s second(!) secretary (true story, not a big district either) et c.
2) Some kids with certain needs require one entire staff member with them at all times. There may be a few of these in a given school.
3) Somewhat more require something like 1/5-1/10 of one or (more often) two staff members’ time (think “special ed” rooms)
4) These figures may include serious outliers. Selective public schools often have very high per-pupil spending, and schools attached to correctional institutions sometimes register six figures of spending per student, to pick a couple examples.
TL;DR if we could suss out what’s actually spent on a “normal” student at a “normal” school, it would be way lower.
> the second(!) assistant superintendent’s second(!) secretary (true story, not a big district either)
I bet if you go ask what those secretaries actually do during their day, you will find they are very busy doing lots of stuff that has nothing to do with the traditional "secretary" role. It was probably the only way to get it stuffed into the budget, so the actual work they do can get done.
Certainly some Administrators have actual secretaries that do traditional secretary duties, but they are few and far between in my experience. Most of the ones I've met do lots and lots of other things totally un-related.
As for multiple Asst Superintendents, that's totally normal, usually one is in charge of "back office" payroll, HR, the business end of things and another is in charge of education type things. At larger districts you might have 1/2 a dozen of them, all specializing in a particular area.
The mention of air travel was strange. I wasn't aware of anyone who thought long range flight would ever be electrified. At least not without some fundamental breakthrough.
S-curves are hard to predict. Basically every time someone attempts to do it, they are way off. This [0] is a neat paper that addresses the question. We've blown past every single prediction.
Why do you mention long range flight? I don't see anything in the article saying batteries will take 100% of the airplane market.
It does say batteries will start to take market share in 2030. That's almost certainly true. It's a high priority for the Norwegian company to electrify the short distance airplane network in the next coming years. There are already battery electric planes coming out. And battery chemistries suitable for short range planes are starting early production.
I suspect battery electric plane will get a surprisingly good range once we start to get highly optimized battery chemistries and optimized airplane designs for that market. The hardest part is to get the first few products to mass market.
They might creep into the medium range market by 2050.
But long range? It might never happen. Unless we get something like aluminum-air batteries that can exploit oxygen in the air somehow. But it doesn't matter. Long range flights are not the majority of flights. It's a small enough market that e-fuels could cover it.
Since flying battery electric will be so much cheaper it's also possible people will have to switch planes multiple times on a journey. Maybe there will be some innovations/optimizations that make that faster and easier.
Long haul might not be the majority of flights by number, but they account for ~40% of the emissions from commercial aviation (well more accurately, wide-bodies do.) Regional flights that are prime targets to go fully electric only account for around 6% of emissions.
But you're right that starting somewhere is better than not doing that.
There are some very early stage tests, there is some kind of island hopper electric airplane that flys regular service, and it's only like 5 or 10 miles across water.
Batteries will get more energy dense, the range will increase a bit. But yeah, it's hard to see it getting to a few 100 miles.
Testing and development by an actual operational airline, but running into regulation and certification issues. Could be a while even for this relatively narrow use case of seaplane flights of under an hour duration. Interesting update. https://harbourair.com/earth-day-eplane-update/
In terms of battery density, the fact that they have an operational, flyable aircraft, just stuffing batteries and an electric motor into a 60 year old air frame... pretty good and only going to get better!
There are people researching it, I believe Airbus is about to test flying with hydrogen. It's the usual thing though for "green hydrogen", there's not much green hydrogen, there are some testbeds but just like for cars it seems to be mostly extracted from natural gas. You can extract it with any energy source like solar power. There's still the challenge that hydrogen fuel is not very compact, so it's hard to carry enough energy (in a car or plane) for much use, you end up with very very high pressure tanks. I think hydrogen will make sense eventually for trucks, tractors maybe. The question is will the massive investments in improving batteries make hydrogen vehicles obsolete or not.
If advances in solar continue, yes. Currently, the IRA provides very lucrative investment and tax credits for green hydrogen projects (solar and wind powered electrolyzers). Power producers like AES are already building multi-billion dollar projects, and there are a lot more in the pipeline. One day the tax credits may not be needed for this to be economically feasible.
Companies like terraform industries are doing something similar, but creating natural gas. With enough cheap solar, all hydrocarbons are pretty much on the table as well.
It'll be a decade or more until this is scaled up and not dependent on subsidies.
The comparison to Oil is interesting. Because people have also been saying we would hit a production wall there, and have been saying that for about 90 years now.
