Discussion on "STARGATE AI: WHO WILL WIN THE $500 BILLION BET ON THE FUTURE OF INTELLIGENCE?" took place at the Faena Forum on DAY 2, February 21, 2025, of the 3rd edition of #FIIPRIORITY Summit in #Miami
The race to build next-generation AI infrastructure has begun, and the Stargate AI project—backed by SoftBank, OpenAI, Oracle, and MGX—aims to invest up to $500 billion to reshape global AI capabilities. With AI-driven industries projected to contribute over $15.7 trillion to the global economy by 2030, this initiative could determine the next era of technological leadership—but as Stargate scales up, nimbler AI challengers like DeepSeek are moving fast, proving that innovation isn’t always about size. Can this megaproject outpace more agile rivals, or will its massive scope slow it down in an AI race that rewards speed?
MODERATOR: Richard Attias, Chairman of the Executive Committee, FII Institute, Founder & Chairman, RA&A
Thank you, my dear Masa, for being our host for the closing session of this amazing FII in Miami.
I don't think Masa needs any introduction, but please, since we have many people also watching,
Masa Yossison is the representative director, corporate officer, chairman and CEO of SoftBank,
a group, a strategic investment holding company focused on driving the AI transformation,
information revolution founded in 1981.
And SoftBank and its subsidiary works in multiple areas, of course, AI, robotics, telecoms,
internet services and clean energy and so forth and so on.
Let me start this conversation.
I know lunchtime is soon, so we will try to go as fast as we can.
But there is a lot to say when we have Masa with us.
So, my dear Masa, I know you for almost 40 years now.
Yeah. From the time I was working for IBM.
And you keep surprising me literally every day, all the time.
So what is your magic to bring and to keep together a diverse coalition of partners on
massive projects like the one you just have announced with President Trump, the Stargate?
So how did you, how were you able to do it?
Well, first of all, I always feel that I have not achieved anything yet.
I'm very ashamed of myself that I have wasted my time for so many years achieving so little.
But my dream is big.
My passion is big.
I'm a small guy physically, but my dream is big.
Come back about that.
Harvard Business Review just wrote something about you.
No, no.
So when your dream is big, you are always humble.
You're always hungry.
You always wake up in the middle of the night and thinking about the solutions.
So that's where I am now.
Still working with passion to make things happen.
And I think timing is really coming now.
Now is the time to go big.
Tell us a little bit more about Stargate and how do you think Stargate project will influence this AI trajectory?
Yeah, first of all, let me give you the big picture.
Stargate will provide more chip, more compute.
So how much more compute, how much more intelligence would there be?
Stargate will provide at least 10x chip, number of chip per year.
So let's say we provide 10x more chip.
Each chip will be about 10 times more compute capability per chip.
So 10 times 10.
And then each model introduced by OpenAI becomes about 10x more capable.
So 10 times 10 times 10.
1000x in one cycle.
The second cycle coming every 12 months to 18 months.
The second cycle comes 1000x times 1000x.
That's a million x.
The third cycle comes, that's another 1000x.
That's a billion x.
So now you know, already OpenAI is O3.
Model O3 passed the exam for PhD on every subject.
Mathematics, physics, medical, law and everything.
So I don't know how many of you can pass PhD on five subjects.
Even one subject is not easy.
Five subjects by one person is super difficult.
So I think in a way already, what is the definition of AGI?
Already it is so smart.
But just imagine, that intelligence passed the exam of PhD.
Becoming a million x in two cycles.
A billion x in three cycles.
And three cycles is coming in just several years.
So if you think, oh, JAI still has some mistakes in understanding, blah, blah.
Oh, I am smarter than you, you know.
Well, that thing you are looking from downward from yourself into JAI.
But that little thing is becoming a billion x smarter.
How dare can you say, oh, you are much little thing.
This is changing our common sense.
Drastically much more than you think.
Much quicker than you think.
And it's exponential.
It's not people, most people think linear.
The reality coming exponential.
And when you discover, wow, most of you too late.
So talking about this mega project action plan.
So as you know, AI driven industries are projected to contribute to 15.7
trillions to the global economy by literally tomorrow, by 2030.
So how does Stargate ensure its captures and drive this value
while maintaining a competitive edge?
Yeah, well, competition is definitely becoming more and more stronger.
So you have to go ahead.
You have to be brave.
You have to go all in.
So some people say after the deep seek syndrome, right?
Oh, you are overspending.
Masa, you are spending 500 billion dollars.
You're overspending.
You know, you can save so much more by less spending.
But I think you are looking at the wrong way.
How much, how much
percent of GDP will be replaced by a billion X smarter thing?
I would say at least five percent within 10 years.
And that five percent is nine trillion dollars.
Or if it's 10 percent, it's 18 trillion dollars.
So somewhere in between five to 10 percent
of today's GDP
will be replaced by this super intelligence.
Well, if that's the amount of return,
you shouldn't be scared of spending a few trillion dollars.
If return is nine to 18 trillion dollars per year.
Why should you save?
Why should you try to be efficient?
For what? I don't get it.
You know, just a little difference
make huge return on your market share.
As of today, OpenAI is roughly 80 percent market share
of every JAI, you know, produced.
And OpenAI has 400 million active users per week now,
which is 2X compared to several months ago.
2X. It's growing so fast.
But you were talking about humility.
So let's try to be humble and
and how does Target plan to stay ahead?
I would say of more adaptive AI initiatives.
What would be your strategy to always be ahead?
Two things. OK.
So OpenAI, they continue to innovate with a model.
Amazing engineers.
Amazing leadership by Sam.
OK, so it's an amazing group.
Their creativity of this innovation is amazing.
And then on top of that, we are trying to provide
the compute with a massive investment.
So my responsibility is to go all in, right?
Invest, provide the compute to Sam, to OpenAI
so that their innovative mind
can be utilized by this massive power of compute.
You know, President Trump has always this great
America's first.
So let me ask you a very direct question.
This Target project, do you think it will become the backbone
of a global AI infrastructure
or it could create something more closed,
an ecosystem which prioritizes US dominance?
Well, the center of innovation is definitely happening from the US.
The center of innovation is coming from Silicon Valley.
That's the fact that all the chip,
all the intelligence, innovation is happening there.
So that's already a center.
The NVIDIA has more than 90% market share,
95% market share of all the AI's compute.
The OpenAI and all the GAFAs,
they're providing most of the intelligence.
So that's the fact.
And now with President Trump,
compared to the last administration who tried to limit
the power of AI, Europe tried to limit
the power of AI.
I think it's wrong.
OK, this is happening no matter what.
And China has already shown their passion,
their capability.
So this is not the time to slow down.
This is time to go all in.
Otherwise, somebody else will do it.
So it's not about
America trying to take away the other country's opportunity.
The other countries like Europe or maybe some other country
try to limit their own future by themselves.
I think that's wrong.
Before criticizing, why do you give up your own future by yourself?
Let me, you know,
knowing that we will have this great conversation yesterday,
I say, oh, what are the last news about myself?
And I discovered an amazing article, which I will share with you,
because I know you don't read a lot from the Harvard Business Review.
You don't read the media.
So let me just tell you the summary of the Harvard Business Review
that was published a few days ago about you,
because you became now a case study in Harvard.
So Masayoshi Son, the billionaire founder and CEO of SoftBank,
Japanese media technology conglomerate is often cast as a dreamer,
financial engineer and speculator.
But this is what they say as a lesson for leaders.
Turn adversity to advantage.
Find a way to persist.
Bridge east and west.
Seek mentor and mentees.
Go big or go home.
Don't let love outweigh logic.
Beware mercenaries and focus on the future.
Do you agree with this definition about how you were conducting your life?
Yeah, I'm a dreamer.
I'm a gambler.
And very often I bet wrong horse, like we work.
But, you know, I learn.
I learn from mistakes.
And I would like to continue
my spirit of challenging.
So at least if I make a mistake by challenging too much,
I don't regret.
I would tell, oh, I was not smart enough, but I'm still challenging.
If I regret by not challenging,
that's when I really get, you know, upset about myself.
So at least, you know, I would like to continue this.
Someday I will make a hit.
Talking about dreamer,
we know another dreamer who is also a doer, who, you know, very well.
He's right hand is the Crown Prince Mohammed bin Salman.
How were you able to convince him to invest years ago?
I remember so much with you and creating this great partnership with you.
No, I didn't convince him.
He already had the vision.
He just needed to find somebody who shared the same vision.
Another dreamer. Another dreamer. OK.
But, you know, he's so smart.
He has a big vision.
And I just happened to be at the right timing
and get to know and we didn't have to
spend our effort to convince, as I said.
He already had his own vision.
We just met at the right time.
Masa, my last question.
But I haven't still given him enough return.
I still owe him, you know.
I still owe him.
Masa, last question.
So many people are trying to enter in this AI.
I would not say industry.
I would say this AI world.
What would be the advice to give to a young entrepreneur
who would like to potentially shift from the current sector
where he or she is working and to enter in the AI?
What should be the first things to do and the keys to enter in the AI industry?
Well, I have five grandkids, grandkids now,
and my daughters ask,
what should I tell to my kids?
You know, they are six years old, 10 years old.
Should I let them use, you know,
charge GPT, right?
So I said, no, no, don't let them use.
Let them live with it.
Breathe with it.
You know, dream with it.
So that it becomes part of their life.
And so don't try to limit
or how many times you can use.
Don't use it for homework, blah, blah.
Don't limit any of those.
They are native, native kids with AI.
So let them live with it.
Use as much as possible, as aggressively as possible.
And be part of that.
Let them, AI, be part of their own life.
That's what I'm telling.
And I truly believe that should be the case.
It's extension of our imagination.
It's extension, billion times extension of our brain.
Why not use it?
You know, there are still some part of the world,
some country, some village with no electricity.
The electricity was invented hundreds of years ago.
And even today, there is a small village
with no access to the electricity.
What a difference in life.
So if you have access to AI,
this is a billion times more intelligence.
You have to use it.
This is how we will close this fantastic
third edition of FII in Miami.
Thank you, Masa, for being with us.
A round of applause for Masayoshi Son.
And ladies and gentlemen,
we hope to see you last week of October in Riyadh.
Have a wonderful afternoon. Thank you.
Please, a round of applause for Masayoshi Son.
Thank you.