There’s a saying I’ve heard that this reminded me of, “if you owe the bank two dollars, that’s your problem, but if you owe the bank two billion dollars, that’s the bank’s problem.” I think it’s relevant because it shows how responsibility shifts as things scale. If you owe the bank a little money then you are just a regular customer and you deal with the bank’s policies. But if you are a big player then it’s up to the bank to negotiate with you and come to mutually agreeable terms.
These companies that have many big data centers are now big players. If they are stressing the grid they will need to be part of the solution in expanding the capacity and infrastructure, either by paying more to the electric companies or by including power infrastructure in their vertical integration. Microsoft has the right idea by investing in Three Mile Island, I think other big players will do similar things.
Unlike banks, grids have an obvious solution to this problem: stop supplying these players with energy when the grid is at significant risk of a black out.
Banks have an obvious solution as do grids. Build their reserves and infrastructure and scale to their biggest customers. A stressed scale out system just requires more scaling. It feels weird to say any other solution, especially when the customer is a paying customer.
I would say the same is true for crypto miners and their increasing need for increasing kilowatt power to solve their algo's.
Are you suggesting it should be illegal for these players to buy energy on the open market?
yes.
there is not and should not be an "open market" for power. the sale of electricity from the power grid is highly regulated, and for good reason. delivery to residential customers and essential services should always be prioritized over industrial bulk purchasers.
bulk purchasers already have the opportunity to use power sources that are not connected to the grid, if they want to use grid power they should be prioritized according to their benefit to society, not their cash on hand.
> they should be prioritized according to their benefit to society
The fact that they have the capability to pay for and consume this power is the strongest signal that they are indeed providing value to society. The money comes from somewhere. If that turns out not to be the case, they have the strongest incentive to stop.
But this sort of rationing at the hands of some regulatory body has historically proven to be a far more deadly cure than whatever the presumptive disease was.
That argument would require regular socialization and redistribution of all wealth. The market does not prioritize what the society needs, because individual needs are weighted by wealth and income.
In a Western society, everyone's needs are supposed to be equally important. We often don't even make majority decisions, because we assume that the constitutional rights of a minority are more important than what the majority finds reasonable.
> The market does not prioritize what the society needs
What do you think prices do? What do you think income is reflective of?
Companies that make lots of money provide things that lots of people are willing to pay for. This is the reflection of what people need. Exxon makes money because people want to stay warm. Cargill makes money because people want to eat. Apple makes money because people want to be entertained. Eli Lilly makes money because people don't want to die.
What do you know that millions of people all acting in their own self interest don't know?
right, hospitals and farmers should just massively raise their prices so they can better compete with crypto miners, that will make the world better...
money is not a good indicator of value provided to society. there is a ton of ways where it falls apart, and there are a lot of people devoted to finding the ways they can collect the most money while providing the least benefit to society.
> money is not a good indicator of value provided to society
Ok, it's the least worst system we have...unless you have an alternative. I'm all ears.
Usually these things end up with “well obviously my personal priorities are objectively true…”
Charge them more until it's worth it.
And sure, tell them they get disconnected first when there's a shortage. But a hard block is worse than taking their money and using it for improvements.
the grid gets to choose how it prioritizes power. if they want too much power at a time the grid doesn't have enough to go around, they're likely fairly early on the cutting block
No, but we should have a discussion about if it is worth it to expense so much energy for something with very little hard, real use case. Al be it bitcoin or LLM's.
there is a lot of money behind AI, but i have yet to see any of them actually get us a use case which work reliably.
Not to mention, AI is fun and all, but that energy could be far better spend making the economy less reliant on fossil fuel.
> there is a lot of money behind AI, but i have yet to see any of them actually get us a use case which work reliably.
Classic HN. You personally can't see the value of X therefore the entire market for X is wrong.
The suggestion is that they can buy it at certain times, but not when the grid is nearing capacity. And yes, that's a perfectly reasonable thing for a society to decide.
It's shared infrastructure, mainly paid for by public funds. If a government decides the priority should be the millions over the few, that's what happens in a functioning government.
Or they can pay for their own power stations and infrastructure to be built.
Markets are neither open nor free, they are closed and regulated. So yes, block their purchasing, tell them to sort their own. Here's the thing, do you cut off one customer when supply is tight or do you risk political suicide by kicking the bucket to millions of small customers? Power utility companies will kick one customer, if they don't they will lose their social license to operate and get legislated out of business.
For emergencies, there are already methods to shut off power to low priority businesses. And say, keep hospitals open.
And, if you want to pay, you can pay a premium to have more 'secure' power. This is an option today, already. No need for any discussion about markets and 'government is holding back the free market'. There is already a market to pay for premium power. And the utilities can charge AI customers more if they want.
this is actually a major problem in the netherlands at the moment.
a part of the energy grid cannot handle the load being generated by all the solar being deployed, and upgrading the grid takes years. In the meantime, bussinesses and even households simply are not getting a new power connection unless they are deemed of enough importance.
If upgrading the grid takes years then invest until it doesn't take years.
There have been many years of build up to solar power. If the people running the grid haven't done their job then replace them with people who can.
Sorry hospitals, these AI jerks have more money! /s
AI jerks will just start putting big red crosses on their buildings, open one 10'x10' space in their facility with some barely passed the exams to get a license type of doctor and claim it to be a medical facility of some sort. Just like other hospitals, they'll have a shortage of beds.
I'm only half joking as from what I've seen of those involved, I would not put this time of consideration past them.
I think you might misunderstand your analogy. If you owe the bank two billion dollars it's their problem because a missing 2 billion dollars is large enough to be painful for the bank.
And the bank stands to lose far more money than the company they have lent to who will just fold.
In renewable energy it is actually common for the funding institutions to be paid a cut of the profits and to have no recourse to repossess assets. They take all the risk because they provide all the funding. The job of the actual developer is to provide a project worth funding.
Are these AI companies already profitable or is it speculation that's causing them to get a ton of funding? AI is such a polarizing subject and I find it difficult to sift through all the hype and hate surrounding it.
They generated a whooping three billion dollars last year.
OpenAI is expecting $3.7 billion in revenue this year, and is going to lose $5 billion, according to this article: https://www.nytimes.com/2024/09/27/technology/openai-chatgpt...
So they are currently printing 1 dollar bills at a cost of $2.35 each.
Point to one single transformative invention that was accretive from day one. The ChatGPT moment isn't even two years old. I don't know for certain whether any of these investments are going to work out, but all of the current cash burn is forward looking. It may all come crashing down, but this is not the right analysis.
This certainly the model that basically all VC funded tech companies use. Though OpenAI is notable, because I don’t think we’ve seen another company burn this much money this fast.
Not a value statement on whether is this is right or wrong.
This predates modern VC by centuries. Early trade routes which connected humanity across continents were highly speculative, requiring huge amounts of capital, and were only profitable years later.
That's not right. They're paying $2.35 to get a $1 per year revenue stream.
Losing money on every transaction, but making it up with scale!
Cosmo Kramer: It's a write-off for them.
Jerry: How is it a write-off?
Cosmo Kramer: They just write it off.
Jerry: Write it off what?
Cosmo Kramer: Jerry, all these big companies, they write off everything.
Jerry: You don't even know what a write-off is.
Cosmo Kramer: Do you?
Jerry: No, I don't.
Cosmo Kramer: But they do. And they're the ones writing it off.
But what was the cost of producing that revenue? At what point do they expect to break even on their investment?
I can very quickly generate a lot of revenue if my investors will let me buy eggs at $0.50 each and resell them for $0.05, but I don't think I'd get any bites unless I could explain how that enterprise will eventually turn a profit.
it does seem very likely that $ per token of inference (at fixed quality) will continue to go down rapidly as hardware and software continues to improve.
But this is what the food delivery companies were (are?) doing. This is what Uber is doing. This is pretty much what every VC funded company does. They burn through the investors' money (because who cares its not their personal money) until they get a foothold.
That's what they were doing when money was free. Money is no longer free, unless you're an AI company.
> They burn through the investors' money (because who cares its not their personal money) until they get a foothold.
Otherwise known as investing. Sometimes it works. Sometimes it doesn't. And the last 20 years has been chockfull of excesses absolutely. But Uber is profitable these days, and has genuinely transformed how people get around. It looked like a negative unit margin dog until it didn't.
In 2021, SoftBank is very interested in your egg selling company.
After all, you can't make an omelet without breaking the bank on some eggs.
And it will go from 9% to 90% over the century. Big tech companies might as well avoid the middleman and evolve into the power generation business, although it's not hard with an open design of a solar farm in the desert.
Giant buildings in the desert with solar cells on top. No need for distribution lines. How many people are actually necessary to staff one once it is up and running? Bury the building with the desert soil for insulation.
Your idea could perhaps work only in space, a bit closer to the sun, operating at peak power generation 24x7. On the ground it takes a lot more area in panels than is used by a datacenter.
I think you massively overestimate the generation capacity of a PV cell. If you cover all the surface area of a datacenter with PV cells, it will generate a small fraction of the power required.
Unless we come up with newer more efficient algorithms and specialized hardware for implementing AI.
glances at Bitcoin mining ASICs
Any day, now.
Bitcoin is definitionally inefficient to mine. That’s the whole point.
The point is that difficulty is adjusted in accordance with network saturation to prevent a 51% attack. The network is not inherently inefficient unless you're attempting to monopolize it - you are still rewarded for using more efficient mining methods than your competitors.
Which leads us to the GPUs. For a while everyone said ASICs (and the courageous ones, FPGAs) were the future of mining because they were inherently more efficient than GPU hardware could ever be. But it turns out that it's the other way around - mass-manufactured GPUs have access to more efficient silicon than any FPGA on the market could hope for. Similarly, AI chips that want to beat Nvidia on efficiency terms first have to climb a mountain of power usage that is inherent to their exponentially less-dense hardware.
It is the first meaningful demand growth this century for the electricity industry
How is this bad? Are the utilities just lazy?
But that's what the article is about.
The utilities are being asked to make massive speculative investments in transmission infrastructure, but are dubious that the demand will actually be there when those projects finish in half a decade, and they'll be left holding the bag.
Also, infrastructure investment has a very long return on investment. (power plants for instance need to run decades to make back their profits, especially nuclear power plants.
Other forms of energy generation come and massive ecological costs. Are we really prepared to spend this much energy on a speculative technology? especially when it is destroying the planet while doing so?
we should already be spending a ton on renewables and transmission capacity.
this is a good way to ensure emissions become sustainable
This is why long term offtakes exist. The customer can prove their conviction and share in the risk. Extremely commonplace.
how is this stressing a grid that has to shed huge amounts of load to maintain availability off peak? instead of shedding the load, just price it effectively and the demand will even out and be a more efficient use of the fuel or generating station. the problem with our energy infra is a lack of consistent demand, and both AI and proof of work create the demand and a consumer of last resort that would make generation way more efficient.
afaik, the main compute load is done in training, not in the queries. this article sounds like astroturfing a pretext for regulation under the auspices of "climate."
also, the diminishing value of cash is accelerating so taking todays dollars and putting them into tomorrow's energy is probably the most fundamentally sound economic use of capital today.
power generators were loaded with debt and used as public sector slush funds (in canada where they were nominally publicly owned) and now the companies are not very productive because they have to pass on the debt service cost to consumers, and prevent anyone from entering the market to offer competitive service that would interfere with that debt service. if you build new energy infra, you need to protect it from managers piling it with debt to loot it, but if you solve that real problem, demand is not a "problem," it's the most valuable thing in the universe.
> American Electric Power, which owns the line, has proposed a new, higher electricity rate for data centers and cryptocurrency miners that would lock them in as customers for a decade. The companies are balking.
Look at that! Net Neutrality comes to the electrical wire. They could have charged different rates based on industrial/residential or volume but that wouldn't be optimal rent extraction.
This is actually a really great article about the weird contract dynamics of large power customers and the chicken-egg problem of needing commitment for a certain amount of guaranteed spend to justify the cost of servicing them. It has nothing at all to do about AI.
Comparing this to net neutrality just poisons the well when discussing NN. Internet service providers being able to arbitrarily adjust pricing based on data source stifles competition. Is there competition in the power supply industry where a client can pick their source? Well, aside from having their own generation. In this case transfer is the service and if one day huge demand cluster disappears the capacity for it is left to rust.
100 years ago was there this much bad-faith handwringing about all the new roads they needed?
What do mean by "they"? EDIT: who do you mean by "they"? (Rephrasing since some readers don't get that intent of the original question.) The handful of corporations interested in transport? Roads were a broad-based need. Roads have been built, used, maintained for thousands of years. The interstate highway system came from a government initiative, if there was a "they" it was broad-based enough to be "us". Railroads served needs that had existed for a long time and it was very clear what their use was from the time they were conceived, and there were more than a handful of participants in the movement of goods and people, since producers wanted greater reach and consumers wanted better access. By contrast to the very small number of "they" asking for more elecricity for data centers. How many people involved in the discussion are asking for the AI companies to be getting this juice?
"They" is a third-person pronoun in English. It refers, in many cases, to a plurality of subjects, often people, not including the speaker. If the speaker is included then it would have been "we" instead, but in this case I was not alive 100 years ago so I use "they" instead of "we".
Hope that helped. Best of luck learning English!
Can you believe these “mammals”, burning energy around the clock to keep their blood warm all the time?
You could go on newspapers.com (free for 7 days) and look at those old papers. I know that the Transcontinental Railroad was intensely controversial.
What's "bad-faith" about it?
OP is saying the opposite: the TCR required huge investment, and was massively speculative, with a huge bubble ensuing. But rail travel did revolutionize the world, and few people sat around asking whether or not it was worth it.