An $800 Billion Revenue Shortfall Threatens AI Future, Bain Says (bloomberg.com) 43
AI companies like OpenAI have been quick to unveil plans for spending hundreds of billions of dollars on data centers, but they have been slower to show how they will pull in revenue to cover all those expenses. Now, the consulting firm Bain & Co. is estimating the shortfall could be far larger than previously understood. Bloomberg: By 2030, AI companies will need $2 trillion in combined annual revenue to fund the computing power needed to meet projected demand, Bain said in its annual Global Technology Report released Tuesday. Yet their revenue is likely to fall $800 billion short of that mark as efforts to monetize services like ChatGPT trail the spending requirements for data centers and related infrastructure, Bain predicted.
The report is set to raise further questions about the AI industry's valuations and business model. The increasing popularity of services such as OpenAI's ChatGPT and Google's Gemini, as well as AI efforts by companies across the planet, means demand for computing capacity and energy is rising at a rapid clip. But the savings provided by AI and companies' ability to generate additional revenue from AI is lagging behind that pace.
The report is set to raise further questions about the AI industry's valuations and business model. The increasing popularity of services such as OpenAI's ChatGPT and Google's Gemini, as well as AI efforts by companies across the planet, means demand for computing capacity and energy is rising at a rapid clip. But the savings provided by AI and companies' ability to generate additional revenue from AI is lagging behind that pace.
"Is Enron Overpriced?" (Score:5, Insightful)
Re:"Is Enron Overpriced?" (Score:5, Insightful)
I see that we are now about to exit the Peak of inflated expectations and are now going into the Trough of disillusionment defined in the Gartner hype cycle [wikipedia.org].
Re:"Is Enron Overpriced?" (Score:5, Funny)
Re: (Score:2)
Mail Gartner about that one!
Re: (Score:1)
It's smelled bubbly for a while already. When will this ZitGPT pop already?
Intel may actually be a good hedge stock since they don't (yet) make notable AI chips. Investors will see them as more reliable because they sell chips for actual and time-tested products. They'll look smart for not hopping on the AI bandwagon so quick, after the bursting.
Re: "Is Enron Overpriced?" (Score:1)
Re: (Score:1)
There are only 3 viable x86 makers: Zhaoxin, Intel, and AMD. AMD also has a lot tied up in AI, so they'd take a big hit when the bubble poppage is "official". Zhaoxin is too small to matter much. Thus, Intel's minimal exposure to AI will help it survive the bubble better.
x86 chips are still a staple of almost every business. Investors love staples during bubble slumps.
Hype (Score:5, Insightful)
Re: (Score:2)
I think it's important to understand the distribution of revenue and profit for the competing companies.
Some are like OpenAI with some revenue but huge losses, mostly from capital expenditures for computer equipment rather than operating expenses. These companies project dramatically increasing revenue but still significant losses for several years. These companies will complete for at least several more years.
Some are like the smaller companies that are struggling. They spend lots of money with minimal rev
hyperscalers... (Score:2)
Except those hundreds of $10,000,000,000 datacenter projects are going to be canceled mid-construction since there's no need to build for future demand that will never materialize. The REITs that were building the datacenters don't really care, as they still own the land and can find something else to do with it or sell the land itself for a profit.
Re: hyperscalers... (Score:3)
Those data center projects will crash head first into the decades long power plant building process. I bet most will not be able to operate.
Re: (Score:2)
Except those hundreds of $10,000,000,000 datacenter projects are going to be canceled mid-construction since there's no need to build for future demand that will never materialize. The REITs that were building the datacenters don't really care, as they still own the land and can find something else to do with it or sell the land itself for a profit.
The whole present business model is screwy. The whole idea of re-commisioning nuclear rfeactors in order to have them supply power to AI is kinda telling. So you need 819 Megawatts (TMI unit 1 output) in order to power AI. Just curious - are you going to need to boil the Susquehanna River to cool these data centers? That's going to be a lot of pretty concentrated heat that needs dissipated.
Point is, this business as they are looking at it isn't the proper path. Maybe it is early on in AI, but if it req
Re: (Score:2)
During the early years the laser was called "a solution in search of a problem". Don't try to estimate what current AIs can do by the applications that they are currently shoe-horned into.
OTOH, every speculation as to how AI will develop further is *speculative*. That explicitly includes the speculation that it will not get any better or more efficient. (And I'd call the speculation that "we've reached top AI" at least as silly as "AGI will show up tomorrow and solve all our problems".
Re: (Score:2)
I don't think that it's a problem of there not being demand. There probably just won't be enough demand to support all of the infrastructure being built. These big AI-players have a choice. They can either bet big and build big now and hope to be one of the last standing when the market shakes out, or they can not play at all and lose for sure. When it's all said and done (5 years?) there will only be a few big AI players still standing, and they are probably names we already know. That is the nature o
Re: (Score:2)
Motivations and evaluation criteria.
Most employees are evaluated negatively if they spend a bunch of corporate money or invest a bunch of time and human resources in an activity or asset that proves to be a failure.
The top level decision makers though are evaluate on their 'vision' and being able to take advantage of opportunities in the market place.
Look at it this way, how many people are still talking about how much MS wasted building Zune's nobody wanted, vs their early Internet, and mobile phone misses
Re:Hype (Score:5, Informative)
Meanwhile, all us peons down here are saying "how could this possibly add up" while great business leaders get sucked in by the hype.
No, the C-suite type didn't "get sucked in" by the hype, they jumped in to *ride* on the hype.
Regardless of outcome, those C-suites get to spend millions to billions of other people's money, for years, all the while paying themselves tens/hundreds of millions every year, to just do something related to the hype to keep the bubble going.
Yes, some years down the road, the hype crashed and burned along with the companies and all the workers there, but the C-suite just jump off with their golden parachute to the next hype, with millions of dollars more than before.
Who did you say got sucked again?
Re: (Score:2)
I guess these "great minds" are too stupid and too ignorant to actually understand anything and are just following the hype.
Re: (Score:3)
I mean, they're being sold a potential solution to that pesky "employee" problem so their companies can be converted into pure money-printing machines once they figure out how to get rid of the other liability... customers. If there's just a 0.001% chance it will work, that's too good to pass up.
Re: (Score:2)
Well, yes. Only problem, if printing money gets that easy, money loses all of its worth.
Re: (Score:1)
Re: (Score:2)
"great business leaders" have fallen for 1 hype or another for a very long time.
the ones who've been taken in by the likes of SBF should never be allowed near civilization ever again
So they don't need revenue (Score:3)
The combination of replacing workers with plummeting wages because of those workers competing for jobs is worth trillions.
One of the major problems people have is they think like an employee and not like an employer.
Your boss is thinking about the entire labor market and how it affects how much he spends on labor.
It's like how when you go buy eggs in the back of your mind you're thinking about the supply of eggs and how that will affect the price.
In this analogy you are the egg and you are going to be cracked to do an omelet and eating.
Re: (Score:2)
Re: (Score:2)
For example a mechanical engineer making $150 K / year it's not uncommon to pay maybe $7500 for CATIA.
For a stock trader it's about $30K / year for a Bloomberg terminal.
Or maybe it will be like operating an MRI scanner which is $5 million to buy, and the MRI technologist who operates it makes $88K / year.
Or a mining dumptruck that costs $7M and the driver is paid $80K.
Nobody knows where this ratio will
Zuckerberg spent something like 50 billion (Score:2)
The thing you don't get is that the ruling class can divert unlimited resources without the need for profit now. They are completely insulated from and untouched by capitalism and meritocracy.
So they don't need a profit because they are above all that. If something is useful to them personally it gets done regardless of whether or not it's a useful product for literally any of the other 8 billion people on this planet.
Re: So they don't need revenue (Score:2)
They will untimely kill their own market and go out of business. A few rich households donâ(TM)t consume much and when the majority of the country is on payment plans for individual pop tarts deflation will be rampant. Capital will dry up, stocks will crash because no one exists to buy in, infrastructure will crumble, and all the big powerful companies will be shown as paper.
Re: (Score:2)
What are the analogous problems in China?
Did the king need customers? (Score:2)
The problem here is you're still thinking and acting as if the system is put in place to protect your interests still exist. They were systematically dismantled over the last 65 some odd years.
Bad In-Roads = Borrowed Infrastructure (Score:1)
There's plenty of demand for free services... (Score:3)
Are today's "AI" companies important to future? (Score:5, Insightful)
The Generative AI companies did their thing. It was overall very impressive, even if they massively overstated its usefulness. ChatGPT is a great early demo of this infantile, currently-almost-useless-but-very-promising tech! Now someone simply (heh) needs to get the compute requirements down two to four orders of magnitude.
If companies like OpenAI can (and want to) work on that, great! Or others can build on the work that's been done up to now. I don't think anyone will miss the current companies, though they might currently be employing people who likely have a leg up (thanks to their familiarity with the subject) on addressing the compute resources problem.
But whenever (if ever) it gets done, people are going to run it on their own machines, not your servers and jail. Lock-in has always been, and will always be, an adversarial force to be eliminated by progress. If that means OpenAI's long-term plans won't work out, well, too bad.
Re: (Score:2)
Are today's "AI" companies important to future?
In the future, I think that today's AI companies will serve important examples of the perils of over-hyping technologies that are still going through heavy development. If not businessmen then (intelligent) investors will absolutely learn about the dangers of over-hyped and underdeveloped technology.
That said, it's absolute gold mine for people looking to churn out huge amount of puppets and astroturfing on social media.
I suspect it will or even already has been used to simplify complex manipulations to mak
Re: (Score:1)
"The poor downtrodden rich need our help now! It's why Thurston Howell had to rent a 3 hour cruise from 2 bumbling morons in a dodgy boat."
Let them die a natural death. (Score:2)
Millions of people ask themselves every day "how do we stop this AI shit-pocalypse?!". Perhaps we don't need to do anything after all. Perhaps all we need to do is to ride the storm, let the economy of it catch up with everyone, and then we'll get our popcorn and watch it burn with a very smug "I told you so" on our faces.
Re: (Score:2)
Shades of 2008. That recession occured because banks had over-invested in mortgage backed securities. Which collapsed in value, scrubbing billions off the banks capital balances and tying up their ability to lend.
Move forward 17 years and now it will be loans for data centers, power generation and grid improvements. All backed by the banks. And all worthless now. Because the value of an asset is based on its ability to generate revenue and service debt.
At least the data centers can be used to house the n
It is called "over capacity" (Score:2)
When you have the capacity to create more, much more, stuff than you can sell, no, it is not a "revenue shortfall", it is called "over capacity".
And the right response, in a free market, to over capacity is to let the least efficient producers go bankrupt, and hence reduce the waste of producing stuff no one buys.
It is also called market consolidation, when smaller producers merge to force bigger, more efficient companies in order to survive the culling.
But, when you allow a monopoly to essentially own the
Odd AI generated slop isn't worth much! (Score:2)
I do use many automated tools I have written over the years, but they take defined blocks of(lets say record variable names) and spit out a nice blocks of add/display/modify/delete base report application code for me. That I usually don't need to spend much time debugging.
The apocalypse has been delayed. (Score:1)
That threat will not be evenly distributed (Score:1)
Yeah, lots of "AI" companies will fail. Their future is threatened, to be sure. But AI future *in general* is not threatened. Just as with the internet bubble of the 1990s didn't threaten the internet itself, this bubble isn't going to threaten AI, just those who don't figure out a way to harness it profitably.
too slow (Score:2)
How can we accelerate this financing situation? I'm going to get AI right on that.