SoftBank boss says the AI revolution will cost $5 trillion a year, also, there's no bubble

[bigreddyde]

No people weren't spending money on IBM mainframes, which are just highly specialized servers, because they were spending money on memory, storage and GPU's, basically AI hardware, instead.

What is a "long run" job? "People once claimed electric machines would eliminate most jobs" Did these people claim, "most jobs" or did they claim, "their jobs"? Hundreds of thousands of textile workers lost their jobs to machines in a short few years period and those machines created an insignificant number of jobs. Not only did they lose most of their jobs, but because so many people were looking for work wages dropped in many industries by as much as 80%.

People manually switched phone calls for 100 years and for 300 years people worked as computers. Both of those jobs were replaced by computers. What about factory jobs? It is estimated hundreds of millions of factory jobs have been eliminated by automation and machines globally since the dawn of the Industrial Revolution.

Do you think a company wouldn't cut payroll as much as they could get away with? What jobs are being created by AI and do you think it's significantly more than it's eliminating? An oversupply of former IT workers would definitely lower wages in any IT adjacent industry.

It's hard to say what jobs will be created, but I can think of a few obvious ones.

Someone will be needed to tag training data. This might seem insignificant, but remember how much data will need to be tagged. Also remember that the quality of that work directly correlates with the quality of the training. I'm sure you have noticed the decline in quality of information published in recent years. This even applies to well known publishers. I see articles in major news publications now that would have received a solid 'F' from any highschool English teacher 20 years ago.

Which brings us to another possible job. Filter.
People who specialize in separating good data from bad. Rocks on pizza, anyone???

 

[Loadedaxe]

It will definitely be worth it, in the future. But first, there will be pain....

I think that is probably the most realistic outcome.

AI may absolutely be worth it in the long run, but that does not mean every company, product or data center being funded today will survive. There will likely be overbuilding, failed businesses, layoffs and investor losses before the genuinely useful applications and sustainable companies separate themselves from the hype.

This pattern has repeated itself since the Industrial Revolution. New technology attracts massive investment, too many companies rush in, speculation outruns reality, and then a painful correction clears out the weaker players.

The technology can succeed while the current investment cycle still ends badly.

 

[Puiu]

Yet inflation as a whole is far lower than a few years ago, and far lower than the historical average for the US:

"...The annual core inflation rate in the U.S. is 2.6% for the 12 months ending in June, cooling down from 2.9% in May..."

Except AI-based models are *increasing* food production ... and doing so with less water, fertlizer, and pesticide use.


IBM is not considered an AI stock.

He was obviously talking about computer parts not the general inflation numbers.

"Except AI-based models are *increasing* food production ... and doing so with less water, fertlizer, and pesticide use." - So much BS. There are examples where short term gains were made, but those were on the back of the soil's health. In reality the AI cannot optimise like they want to (like in the lab) because of how much soil quality differs even on a small patch of land. On the other hand, large agriculture corporations have been stealing the data of all of the farmers who have been using AI (exact planting dates, yield maps, soil boundaries, and even farm financials) and they can manipulate the market and prices with that data.

Data centers are actively competing with the food industry for land, water and power. This is driving up the cost of farming. And this affects small farmers massively and benefits large corporations.

AI tools are being marketed as "a way to optimize crop yields out in the field", but in reality the physical infrastructure required to run AI is simultaneously raising the baseline cost of the land, water, and power needed to grow food (the land especially gets really expensive). Then you can add the increased temperature in the region which affects crops that are near the data centers and the constant noise that affects livestock.

 

[Puiu]

It's hard to say what jobs will be created, but I can think of a few obvious ones.

Someone will be needed to tag training data. This might seem insignificant, but remember how much data will need to be tagged. Also remember that the quality of that work directly correlates with the quality of the training. I'm sure you have noticed the decline in quality of information published in recent years. This even applies to well known publishers. I see articles in major news publications now that would have received a solid 'F' from any highschool English teacher 20 years ago.

Which brings us to another possible job. Filter.
People who specialize in separating good data from bad. Rocks on pizza, anyone???

Tagging has already been taken over by AI so that's out. The same with "filters".

 

[Endymio]

"Except AI-based models are *increasing* food production ... and doing so with less water, fertlizer, and pesticide use." - So much BS.

Oops!

"...AI is revolutionizing agricultural water conservation by integrating real-time climate, soil, and crop data to drive precision irrigation. This intelligent automation significantly reduces water waste—frequently cutting irrigation usage by 25% to 40% while preserving or boosting crop yields..."

https://www.sciencedirect.com/science/article/pii/S2772375525002151

Smart Farms: How AI Can Solve Water Wastage | Earth.Org

With water increasingly valuable against the backdrop of a rising human population and climate change, can AI prevent water wastage?

earth.org

"...Artificial intelligence is revolutionizing agriculture by shifting the industry away from broad-spectrum chemical sprays. By combining computer vision, IoT sensors, and robotics, AI pinpoints, targets, and neutralizes weeds and pests, reducing total pesticide and herbicide use by 50% to 90% while keeping crops healthy...."

PerPlant's Cameras Help Farmers Target Pesticide Use

The startup utilizes AI cameras to get a better picture of crop needs.

www.automate.org

"AI is revolutionizing agriculture by shifting farming from a blanket approach to pinpoint precision. By analyzing real-time data from satellites, drones, and soil sensors, AI optimizes resource use, detects diseases early, and accelerates the development of climate-resilient crops to significantly boost overall food yield)...

https://www.weforum.org/stories/2025/06/ai-food-systems-agricultural-revolution/

https://www.sciencedirect.com/science/article/pii/S2666154325001334

 

[Endymio]

Data centers are actively competing with the food industry for land, water and power.

You're setting new records for hilarity. Globally, agriculture uses over two quadrillion gallons of water, compared to which all datacenters combine use only 0.1% as much -- mere rounding error. Worse, datacenters are either sited in areas where water is cheap and abundant, or else they use closed-loops systems which consume almost no water at all.

Your land argument is even more comic, as acreage used for farmland in the US totals 900 million acres and has been steadily declining for well over a century, with the end result that many millions of acres (far more than all datacenters combined) isn't even being actively farmed; it's merely sitting unused.

 

[n3o127]

Either AI does work the way it's being sold to us as and it crashes the economy because noone has a job or it doesn't work and it crashes the economy because the businesses can't pay back investors creating something similar to the subprime loan crisis.

Investors and shareholders loge AI because it sounds like an infinite money glitch. The problem with having infinite money is that it becomes worthless and it's no different from having zero money

Hope you're wrong, though I agree with you 100%.
This gets compared by some to PC boom and dot com boom that followed, both which intent by both was to become a tool for people vs a replacement where possible. The next 10 or 15 years is going to be an interesting ride with AI, politics, mega corps, mega pharma, mix in the divide in opinions and such....

I do feel though that Robots could be a more positive impactful tool much like the move to computers, and dot com -- perhaps also the best fit for autonomic or AI.

 

[Loadedaxe]

I agree that robotics could be one of the more positive and practical uses of AI, especially for dangerous, repetitive or physically demanding work.

Robots are not new to manufacturing, though. The auto industry has used them for welding, painting, lifting and assembly for decades. The difference is that most traditional robots perform one narrowly programmed task, while AI could eventually make them more adaptable and useful outside highly controlled factories.

I would also say computers and the internet were not purely tools rather than replacements. They improved productivity, but they also eliminated and transformed many jobs. Technology has almost always done both.

The real question is whether AI and robotics are used to help people and improve productivity, or simply to remove workers as quickly as possible while all the benefits flow upward.

 

[Puiu]

Oops!

"...AI is revolutionizing agricultural water conservation by integrating real-time climate, soil, and crop data to drive precision irrigation. This intelligent automation significantly reduces water waste—frequently cutting irrigation usage by 25% to 40% while preserving or boosting crop yields..."

https://www.sciencedirect.com/science/article/pii/S2772375525002151

Smart Farms: How AI Can Solve Water Wastage | Earth.Org

With water increasingly valuable against the backdrop of a rising human population and climate change, can AI prevent water wastage?

earth.org

"...Artificial intelligence is revolutionizing agriculture by shifting the industry away from broad-spectrum chemical sprays. By combining computer vision, IoT sensors, and robotics, AI pinpoints, targets, and neutralizes weeds and pests, reducing total pesticide and herbicide use by 50% to 90% while keeping crops healthy...."

PerPlant's Cameras Help Farmers Target Pesticide Use

The startup utilizes AI cameras to get a better picture of crop needs.

www.automate.org

"AI is revolutionizing agriculture by shifting farming from a blanket approach to pinpoint precision. By analyzing real-time data from satellites, drones, and soil sensors, AI optimizes resource use, detects diseases early, and accelerates the development of climate-resilient crops to significantly boost overall food yield)...

https://www.weforum.org/stories/2025/06/ai-food-systems-agricultural-revolution/

https://www.sciencedirect.com/science/article/pii/S2666154325001334

This is so funny. Most if what's written there is just prefaced by the word "could". Especially the water thing. You don't need "AI", systems already exist that help with giving general information to farmers for this.

"Their AI system can learn to identify damaged areas in a pipe without the need for manual inspection." - Huh? What BS is this? You don't need AI for this, just some sensors.

As for the 30% numbers in that article, those were some tests done in 2016, before any of the "AI" you know in the modern sense. It basically did some weather predictions on draught and gave that info directly to the farmers in the pilot program which delayed the date of the planting of the seeds. We can do that without any of the data centers that we were talking about today, with hardware from over a decade ago. And in fact we are already doing it. It's all BS. The real star was weather forecasting.

The AI irrigation systems mentioned in the study are just theoretical and even in that study they admit that real world feasibility is very poor because of major issues with the sensors and other problems.

In fact the vast majority of the things that "AI" is being used for are just not needed. The vast majority of the gains (water savings) are from knowing the soil general moisture with a lot of IOT sensors and predicting if it will rain; TL;DR knowing if turning on the irrigation system is ok or not which can be done with regular easy to use and cost efficient algorithms if you have the data.

These studies always pit "traditional" farming to uber digitized systems with a lot of sensors, imaging systems etc. It's the sensors that are doing the heavy lifting in everything. Not the "AI".

For example: If soil moisture drops below 35%, open the valve; if it hits 60%, close it - that is not AI. Then if you add the weather forecast, air moisture and temps as variables for that day you can effectively do everything these "magical AI systems" are doing with simple algorithms.

In the end it's all about investing massively in the sensors and the infrastructure, not the AI. The AI here is useless and adds unneeded cost and complexity.

Then you literally have what I wrote about previously:
"Furthermore, scale-specific dynamics were noted: smallholder farms typically face delayed ROI due to higher per-unit capital costs and limited technical infrastructure, whereas commercial farms demonstrate more favorable economic metrics. Qian et al. [82] underscored this gap by quantifying crowding-out effects and investment risk disparities. These findings underscore the need for context-sensitive economic frameworks and funding models to support widespread adoption."

As for the pesticide... the ideas are even dumber. It literally requires per plant spraying systems and it assumes that farmers only blanket spray.

"Smart smart-nozzles on automated machinery instantly turn on and off, spraying only the specific, infected plants rather than the whole field, cutting pesticide volume by up to 80%."

Like seriously, what level of investment are we talking here? A single drone costs thousands and the IOT sensors are simply unfeasible for most but the largest farms in the world. This is just stupid. And the proposed solution for money? Ask farmers to take a picture on the field and send it to the cloud if they see a pest... yes, it's just that dumb.

People really love to think that AI is magical... these people have simply turned off their brain.

 

[Puiu]

You're setting new records for hilarity. Globally, agriculture uses over two quadrillion gallons of water, compared to which all datacenters combine use only 0.1% as much -- mere rounding error. Worse, datacenters are either sited in areas where water is cheap and abundant, or else they use closed-loops systems which consume almost no water at all.

Your land argument is even more comic, as acreage used for farmland in the US totals 900 million acres and has been steadily declining for well over a century, with the end result that many millions of acres (far more than all datacenters combined) isn't even being actively farmed; it's merely sitting unused.

Some people simply cannot understand statistics and numbers. You are one of them.

WTF do global numbers have to do with this? It is and has always been a local issue where the data centers are being built. Large data centers can use as much water as a small city of 50k people.

Here's a real statistic that should scare you: 40% of US data centers are located in areas facing high or extreme water stress. Google's data centers in the Dalles accounts for about 40% of all of the water usage in that region (which grew from about 15% before the AI craze.) And they still have data centers being constructed there. The water bills for people living there are expected to rise by almost 100% in the next several years because they have to expand the water system to give google more water.

 

[BadThad]

I've use AI every single day since it came out. It's as significant of a game changer as the invention of PC. It has improved my professional work by leaps and bounds. So far, it is NOT a panacea, it requires human review, intelligence and clever and repeated querying to be useful. IT MAKES MISTAKES!

This is certainly a bubble of some sort. Just like the .com boom, there will be losers and there will be winners. It's going to take 10 years just to start to sort it out. I think it overshot a lot and now it's (already) pulling back some. I believe that will consolidate and stabilize soon as the AI capital dries up and funds are reallocated.

 

[BadThad]

You're setting new records for hilarity. Globally, agriculture uses over two quadrillion gallons of water, compared to which all datacenters combine use only 0.1% as much -- mere rounding error. Worse, datacenters are either sited in areas where water is cheap and abundant, or else they use closed-loops systems which consume almost no water at all.

Your land argument is even more comic, as acreage used for farmland in the US totals 900 million acres and has been steadily declining for well over a century, with the end result that many millions of acres (far more than all datacenters combined) isn't even being actively farmed; it's merely sitting unused.

America has an near ENDLESS supply of land and most areas have endless supplies of water. Just one flight across our great country and the ignorant will be shown the light. Most of America is trees, rivers, fresh water lakes and open space with a few clustered cities. We have the GREATEST country in the world.

 

[BadThad]

This is so funny. Most if what's written there is just prefaced by the word "could". Especially the water thing. You don't need "AI", systems already exist that help with giving general information to farmers for this.

"Their AI system can learn to identify damaged areas in a pipe without the need for manual inspection." - Huh? What BS is this? You don't need AI for this, just some sensors.

As for the 30% numbers in that article, those were some tests done in 2016, before any of the "AI" you know in the modern sense. It basically did some weather predictions on draught and gave that info directly to the farmers in the pilot program which delayed the date of the planting of the seeds. We can do that without any of the data centers that we were talking about today, with hardware from over a decade ago. And in fact we are already doing it. It's all BS. The real star was weather forecasting.

The AI irrigation systems mentioned in the study are just theoretical and even in that study they admit that real world feasibility is very poor because of major issues with the sensors and other problems.

In fact the vast majority of the things that "AI" is being used for are just not needed. The vast majority of the gains (water savings) are from knowing the soil general moisture with a lot of IOT sensors and predicting if it will rain; TL;DR knowing if turning on the irrigation system is ok or not which can be done with regular easy to use and cost efficient algorithms if you have the data.

These studies always pit "traditional" farming to uber digitized systems with a lot of sensors, imaging systems etc. It's the sensors that are doing the heavy lifting in everything. Not the "AI".

For example: If soil moisture drops below 35%, open the valve; if it hits 60%, close it - that is not AI. Then if you add the weather forecast, air moisture and temps as variables for that day you can effectively do everything these "magical AI systems" are doing with simple algorithms.

In the end it's all about investing massively in the sensors and the infrastructure, not the AI. The AI here is useless and adds unneeded cost and complexity.

Then you literally have what I wrote about previously:
"Furthermore, scale-specific dynamics were noted: smallholder farms typically face delayed ROI due to higher per-unit capital costs and limited technical infrastructure, whereas commercial farms demonstrate more favorable economic metrics. Qian et al. [82] underscored this gap by quantifying crowding-out effects and investment risk disparities. These findings underscore the need for context-sensitive economic frameworks and funding models to support widespread adoption."

As for the pesticide... the ideas are even dumber. It literally requires per plant spraying systems and it assumes that farmers only blanket spray.

"Smart smart-nozzles on automated machinery instantly turn on and off, spraying only the specific, infected plants rather than the whole field, cutting pesticide volume by up to 80%."

Like seriously, what level of investment are we talking here? A single drone costs thousands and the IOT sensors are simply unfeasible for most but the largest farms in the world. This is just stupid. And the proposed solution for money? Ask farmers to take a picture on the field and send it to the cloud if they see a pest... yes, it's just that dumb.

People really love to think that AI is magical... these people have simply turned off their brain.

Rarely do we agree but most of this I do agree with. Overuse of AI already exists and it will continue to be overused. If someone figures out how to AI a bubble gum machine, does that REALLY matter for the good of humanity? It serves no real purpose. AI's following in the footsteps of Bluetooth and "smart" appliances. Why do I need Bluetooth everything when I can just easily flick a switch myself? Do we need AI light switches so they can decide when to come on for me? lolololol

The net result will be over-dependence for the most trivial things. Subsequently, most humans will become even dumber by relying on AI for everything including wiping their posteriors.

 

[Puiu]

Rarely do we agree but most of this I do agree with. Overuse of AI already exists and it will continue to be overused. If someone figures out how to AI a bubble gum machine, does that REALLY matter for the good of humanity? It serves no real purpose. AI's following in the footsteps of Bluetooth and "smart" appliances. Why do I need Bluetooth everything when I can just easily flick a switch myself? Do we need AI light switches so they can decide when to come on for me? lolololol

The net result will be over-dependence for the most trivial things. Subsequently, most humans will become even dumber by relying on AI for everything including wiping their posteriors.

I'm not sure about bubble gum machines, but vending machines with AI do exist. And their primary function is dynamically adjusting the pricing based on demand and time of day.

 

[DirkMeister]

Most of my life I've always been adamant to take the RED Pill, but now in today's world the BLUE Pill is starting to look so much better.

 

[Rocky4040]

So he say's
The age when humans are the highest life form on Earth will end. For better or for worse, it will happen, and it can't be stopped," Son said.

Yes, it can be stopped all it takes is someone with some common sense to actually stop it from happening by not allowing this to continue and end it before it is too late for us to recover from this.

But this will not happen because there are too many greedy little dogs wanting a piece of the AI pie and try and make trillions from it. They are aware of what could happen and yet they are too stupid to pull the plug and say hey man this is not worth risking the human race for money.

I am at an age that by the time this could happen I will likely be too old to care about it, but I do feel sorry for the younger generation that will have to live through it. I may be overreacting to it, but the guy himself is predicting this is most likely how things are going to turn out and yet he is all gong hoe on making it happen. Only fools move forward fully knowing what can happen.

if it does happen the only hope I have is it happens to them first and they are replaced or wiped out but we all know it will be the common folk first. I have never been a supporter of the tin foil hat crowd but maybe they on to something here or maybe not. Either way by the time it happens I will be 73 years old so for myself I am not caring to much as I will be either dead or drooling on my shirt and not having a care in the world...lol

 

[Northy]

Confucius dude says literally all the money in the world will have to be spent on AI, and that's not a bubble.

 

[bigreddyde]

Tagging has already been taken over by AI so that's out. The same with "filters".

AI can only "reliably" tag or filter things that are similar to what it has seen. It is well known that training AI with output from AI results in no new information.