00:00:02
I believe connectivity is a human right.
00:00:05
I felt that this is a great opportunity
00:00:07
to really build and enable a country
00:00:09
like Saudi Arabia that has a incredible
00:00:11
potential and capability to really build
00:00:13
a digital champion and a digital hub for
00:00:15
the region.
00:00:16
This is the first time I landed in a
00:00:18
place where I feel I'm home.
00:00:21
Ladies and gentlemen, please welcome
00:00:25
Humane's Tar Amin.
00:00:27
[Music]
00:00:28
[Applause]
00:00:29
[Music]
00:00:32
[Applause]
00:00:36
I did. Good to see you.
00:00:38
So,
00:00:39
I've been spending a lot more time in
00:00:41
the Middle East in the region. It's
00:00:43
changing so dramatically year after
00:00:46
year. Um, and I'm going to be, I
00:00:48
announced on the podcast two weeks ago,
00:00:50
I'll be doing Foundry University in
00:00:52
Riyad, November 3rd, 4th, and 5th. And
00:00:55
I'm really excited to partner with such
00:00:58
a dynamic country. Maybe you could fill
00:01:01
us in on what's changed over the last 5
00:01:04
years in the region.
00:01:06
Well, thank you very much. First of all,
00:01:07
I must say it's the first time I did an
00:01:09
event that's only 45 minutes away from
00:01:11
my home. Um my flight from Riad uh
00:01:15
obviously was 23 hours but uh just to
00:01:17
give you some context um I was born in
00:01:21
Aman Jordan before coming to the US to
00:01:24
finish my studies and um you know
00:01:26
progress in my life and my career. Um,
00:01:29
so I moved from Tokyo to Saudi Arabia,
00:01:31
but this is the first time I ever worked
00:01:33
in the region. And um, I didn't know
00:01:36
what to expect, you know. Um, you know,
00:01:39
sometimes it def it depends on your
00:01:41
perspective on how you look at the
00:01:42
world. You know, I wanted to really
00:01:43
discover and understand everything I
00:01:45
used to hear under the massive
00:01:48
opportunity of transformation and the
00:01:50
diversification on the economy. You know
00:01:53
my first observation is really started
00:01:55
with discovery of how amazing the people
00:01:59
were in terms of hospitality,
00:02:02
welcom um I really felt like I'm back
00:02:05
home. It's it felt very very different
00:02:07
feeling to me. The second thing the
00:02:09
population is awfully young and hungry
00:02:12
and hungry for new things and new
00:02:14
advancements.
00:02:16
So um to me it was really a pleasant
00:02:19
surprise because what I was really
00:02:21
worried about and what everybody keeps
00:02:23
talking about um maybe TK have done
00:02:25
great things in his life talent and
00:02:28
opportunity may be a big challenge in in
00:02:30
his new venture and it was remarkable
00:02:34
you know um in every aspect whether you
00:02:36
look at at the government side in terms
00:02:38
of society the transformation
00:02:40
it has changed dramatically it's really
00:02:43
really a u an area that I feel and the
00:02:46
opportunity is remarkable and and the
00:02:49
embracement of the society towards
00:02:51
looking at the future in which digital
00:02:53
and AI is fundamental to their
00:02:55
transformation.
00:02:56
Tara, can you maybe explain what humane
00:02:58
is and how it came to be because it's it
00:03:01
was sort of an evolution, right?
00:03:02
Yeah, I mean it's it's uh you know for
00:03:04
for you to know um I'll tell you we take
00:03:08
things for granted in the US to a
00:03:10
certain extent. Let's assume you're a
00:03:12
startup here and you want to access
00:03:14
compute. It's as simple as obviously
00:03:16
going to hyperscaler and within 30
00:03:19
seconds you're ready to go. So um uh
00:03:22
this story is fascinating because my
00:03:24
first discovery let's say in the role I
00:03:26
was in at the time I was hired um uh to
00:03:29
really run a subsidiary of Aramco
00:03:32
uh a new company that was intended to
00:03:34
diversify out of its core business in
00:03:36
which digital AI was a really core
00:03:38
component of it. One of the biggest
00:03:40
surprises I had is lack of AI
00:03:44
infrastructure. I mean I did not know
00:03:47
that
00:03:48
um the ability for startups companies to
00:03:51
access AI infrastructure is a challenge.
00:03:54
I mean it took look how large Aramco is
00:03:56
took them 9 months from the process of
00:03:59
purchase order export control process
00:04:02
deployment installation. That means we
00:04:04
have hindered a company to really
00:04:06
achieve its objectives um by 9 months.
00:04:09
So that was the first opportunity that I
00:04:12
saw that we need to address in the
00:04:14
country. I tell you the story because
00:04:16
this is how the humane story started. So
00:04:18
I started meeting with several ministers
00:04:21
in the country. I told them I see
00:04:23
nothing but opportunity. I see something
00:04:25
that the country could really do and
00:04:27
participate at a global space. I think
00:04:29
we have an abundance of land, an
00:04:31
abundance of power, amazing connectivity
00:04:35
um and an opportunity to really
00:04:37
participate in the digital
00:04:39
infrastructure.
00:04:41
So um uh last year during my birthday, I
00:04:45
get a call and um uh it was like really
00:04:48
a strange number. So I answered it. It
00:04:51
says um it was from the royal court and
00:04:55
um it said uh his royal highness wants
00:04:57
to meet with you. I said okay what did I
00:05:00
do? You know it was a uh so anyway I we
00:05:04
all came in I said what's the topic? The
00:05:06
topic was uh AI brainstorming and um
00:05:10
this was the first time I had the
00:05:12
opportunity to meet his royal highness
00:05:14
and uh
00:05:15
Muhammad bin Salam.
00:05:16
Yeah. And let me tell you it was not
00:05:17
about brainstorming. It was about really
00:05:20
addressing what the country needs to do
00:05:24
to accelerate to address some of the um
00:05:28
fragmentation that exist. And the idea
00:05:31
and the concept of humane is in order
00:05:34
for us to really accelerate our
00:05:36
development, let us now bring um uh
00:05:39
public and private entity combines and
00:05:42
take AI investments, projects,
00:05:45
initiatives and put them one under one
00:05:47
umbrella company that is really focused
00:05:49
on the entire AI total value chain. So
00:05:53
um honestly in that meeting I I could
00:05:55
tell you it took about 10 minutes to
00:05:56
come to realization that the opportunity
00:05:58
is massive. But we need to really unite
00:06:01
the fragmented effort and put our energy
00:06:05
to address the basics
00:06:06
and and you were given a pretty large
00:06:09
balance sheet to do this with.
00:06:10
Yeah. I mean look um I mean to to start
00:06:13
with as you know uh building
00:06:15
infrastructure is not necessarily
00:06:18
uh a small task. So uh to do this and to
00:06:22
participate and to make an impact you
00:06:24
need to be very well funded. So um uh
00:06:27
you know even though we call humane as a
00:06:29
startup but it's actually a very
00:06:30
wellunded startup between our data
00:06:33
center segment our models team
00:06:35
applications um and also our ventures
00:06:38
that we will soon have in the United
00:06:39
States is u you know is a very wellunded
00:06:42
company that I'm feeling really
00:06:44
comfortable about the opportunities that
00:06:45
we could capture between capital people
00:06:48
talent and partnerships. So should we
00:06:50
think about this as there's you know we
00:06:52
live in a world as you said when when
00:06:55
you're back home here in the US there's
00:06:57
AWS there's GCP there's you know Azure
00:07:00
it's just simple straightforward but the
00:07:02
rest of the world a little bit more
00:07:03
complicated do you see humane being that
00:07:07
competitive alternative for the rest of
00:07:10
the world is that how it starts or
00:07:12
well I think I think look at it in
00:07:14
certain areas we partner really well and
00:07:17
all of them are colleagues and friends I
00:07:19
mean I've I've known the CEOs of these
00:07:20
companies really well through my tenure
00:07:22
in Japan and India. But I will tell you
00:07:26
I think many companies do not see the
00:07:29
Middle East the way I see it today and
00:07:31
they don't understand the requirements
00:07:32
and the needs of what we have to do to
00:07:35
build the depth and the talent. So I see
00:07:38
in certain areas great partnerships and
00:07:40
in certain other areas I think that we
00:07:43
will be a compelling alternative but it
00:07:45
doesn't mean that we don't partner. So
00:07:48
um uh you know certainly when you see
00:07:50
now the partnerships that we announced
00:07:51
with uh the likes of AWS and Google uh
00:07:55
clearly I want to see the companies
00:07:58
invest in addressing the lack of
00:08:01
infrastructure that I saw. So a portion
00:08:04
of what we do today will be built on
00:08:06
things with our own technology and a
00:08:08
portion will be done also with
00:08:09
partnerships with the hyperscalers
00:08:11
and what uh give us the lay of the land
00:08:14
from the foundational model side the
00:08:16
opensis the anthropics the gros
00:08:19
how is that moving how do you work with
00:08:21
them how do you make sure that you know
00:08:23
everybody can get access to the latest
00:08:25
models
00:08:25
so so a couple of things we have done
00:08:27
and look um for people in this audience
00:08:29
that have participated on the build a
00:08:32
foundation model you It's not trivial
00:08:33
and it's not easy. One of the
00:08:36
acquisitions that has happened in humane
00:08:39
and this is where I was really really
00:08:41
surprised to see the depth of science
00:08:45
team I have today in humane. I mean this
00:08:47
is the first comment I get in my social
00:08:49
post to say yes you have capital but
00:08:52
there's no people no talent. So we ended
00:08:54
up building a foundation model and I'll
00:08:57
tell you the reason we did this. We
00:08:58
built it from the scratch. It was not
00:09:00
distilled of any open- source tool.
00:09:02
Well, for two reasons. I just wanted to
00:09:04
see I mean a let's venture into this to
00:09:07
understand the depth and the capability
00:09:08
of the organization. Second thing when
00:09:11
it comes to culture, language and biases
00:09:14
you know I felt that it's important for
00:09:16
us to really participate in this arena.
00:09:19
So um so we launched in uh Saudi Arabia
00:09:23
what we call humane chat. I just wanted
00:09:24
to see the reaction from uh the people
00:09:28
that uses this model and the whole
00:09:30
objective of it. Arabic first preference
00:09:32
is the way we train this model, not
00:09:34
English first preference.
00:09:36
And um
00:09:37
I'm sorry, just on that, you see huge
00:09:38
differences in how the foundational
00:09:40
models
00:09:41
process and digest data.
00:09:43
I think for us and any model um that is
00:09:47
used today
00:09:49
um quality of data is always much better
00:09:53
than I think just feeding at quantity of
00:09:55
data. Our model today has a proprietary
00:09:58
set of data in Arabic language that you
00:10:00
cannot find on the public internet. So
00:10:03
we had a proprietary data set that was
00:10:05
important and obviously within the
00:10:06
Arabic language the government today
00:10:08
uses all its correspondence,
00:10:10
transactions, translations is in Arabic.
00:10:13
So the preference of training first in
00:10:15
Arabic is very different than you
00:10:16
starting in English and then you add to
00:10:18
it the Arabic components. Um, so we
00:10:21
built that, we launched it, it actually
00:10:23
now became the number one uh app in the
00:10:25
app store in in the country. Um, it
00:10:28
rallied really an important thing for us
00:10:30
and and I'm I want to be really specific
00:10:32
that we didn't do this to say we're
00:10:34
better than OpenAI or we're better than
00:10:36
X or better than that. It was important
00:10:39
for us to train the team on
00:10:42
understanding how to build the entire
00:10:44
stack. It's really really critical and I
00:10:47
and I would tell you there's nothing
00:10:49
else I would do different than the last
00:10:52
nine months in building this foundation
00:10:54
model. We have um great partnerships
00:10:57
today with um uh you know discussions
00:11:00
with OpenAI uh Anthropic we use a lot
00:11:03
for our AI coding tool. So I I see this
00:11:06
is not an option of use this or that. I
00:11:09
think we are trying to formulate a
00:11:11
strategy on the model but I will tell
00:11:13
you where we differentiate hugely
00:11:15
differentiate. I think we have found the
00:11:18
answer that I hear everybody talking
00:11:20
about. I am I am a huge believer that AI
00:11:24
is not into a bubble and the reason I
00:11:26
believe in this I think we are one of
00:11:28
the few companies that have found true
00:11:31
value realization and this is really a
00:11:33
mind-blowing story. Um when I took on
00:11:37
the s role for him so I had an option to
00:11:39
say how do you run a large enterprise do
00:11:42
you run it the same old way where you
00:11:44
have legacy systems in play you have
00:11:46
hundreds of IT tools that do the job for
00:11:49
finance legal HR cyber security or you
00:11:52
do something different what we have
00:11:54
differentiated on while I'm excited
00:11:56
about the model is what we build on top
00:11:58
of the model so um we will launch in
00:12:00
October this year a platform we call
00:12:03
humane one it is truly the AI oper
00:12:05
operating system for the enterprise.
00:12:07
Imagine the um the era I believe Windows
00:12:11
was embedded in 1981. We all got taught
00:12:13
that you use icons on the desktop. If
00:12:16
you want to take a vacation, you go to
00:12:17
success factor or other HR tool. If you
00:12:20
want finance, go to Oracle, SAP. If you
00:12:23
want something else, there's another
00:12:24
tool, another application, another icon.
00:12:27
Well, now we change this enterprise
00:12:29
world. We're really intent driven
00:12:31
system, multi- aent orchestration system
00:12:33
and its impact is unreal. what has
00:12:36
happened in driving true value
00:12:38
realization.
00:12:40
Um, at least in my company I could tell
00:12:42
you it is remarkable the efficiencies
00:12:45
now that we have derived out of changing
00:12:48
these legacy systems and maybe the
00:12:51
biggest challenge that I had had nothing
00:12:53
to do with the technology nothing and
00:12:55
this is maybe my own personal opinion
00:12:57
about the struggles that companies going
00:12:59
to go through as they embrace AI into
00:13:01
their operation. And it's the mindset,
00:13:02
mentality, culture, organization. It's
00:13:05
underestimated effort. So, um, so yeah,
00:13:08
between the model of what we build, but
00:13:10
I I'm more excited about what we have on
00:13:12
top of the model.
00:13:13
Let me ask you a point of question.
00:13:15
You're in Saudi, there's an enormous
00:13:17
amount of energy, almost a surfate of
00:13:19
energy.
00:13:21
Do you have this unfair advantage where
00:13:22
if you need a vast amount of energy to
00:13:26
throw at a compute problem, do you
00:13:28
effectively get it for free or do you
00:13:30
have to pay for it as well? No, of
00:13:31
course not. I mean, look, um, uh, even
00:13:34
though it might seem like I I wish I
00:13:36
could get it for free. I that would that
00:13:38
would change my business model
00:13:39
dramatically. Um, so the
00:13:44
when I took this role, I think one of
00:13:46
the key things I said, look, what can we
00:13:48
really compete and what can we do to
00:13:50
offer to the world something that the
00:13:52
world desperately needs today. I think
00:13:55
everybody in this room maybe realize we
00:13:58
are really just at the beginning of what
00:14:00
I would call the AI inferencing world.
00:14:03
While the model training will continue
00:14:05
to evolve, but now as we move into
00:14:07
accelerated compute, I think the demand
00:14:10
is going to become much larger than it
00:14:12
exists today. So how do we and and
00:14:14
what's the challenge? The challenge is
00:14:16
really simple power. You know, how do
00:14:18
you really find power? So that's where I
00:14:20
think Saudi Arabia has a big role to uh
00:14:22
to participate. And I was bullish enough
00:14:25
to tell the world I said outside of the
00:14:26
United States, outside of China, I
00:14:28
really think Saudi Arabia has a good
00:14:30
shot to be the third largest country in
00:14:33
infrastructure. So um we have to go
00:14:36
through processes to secure power from
00:14:38
the uh ministry of energy through the
00:14:41
local electric company. I am treated as
00:14:44
fairly as any other entity that comes to
00:14:47
the country. my rates, my tariff is
00:14:49
equivalent to what Google would get, to
00:14:51
what AWS get. But however, the energy
00:14:54
generation that exists in Saudi Arabia
00:14:56
is just remarkable. I mean, I think um
00:14:59
what I told many people um maybe Saudi
00:15:03
Arabia today led the world in energy
00:15:05
exports via oil. we should look at an
00:15:07
opportunity to lead the world through
00:15:09
energy exports via tokens, you know, and
00:15:11
but that's really a a key area that I
00:15:14
think we could differentiate
00:15:16
is the energy infrastructure.
00:15:17
Let's talk a little bit about the talent
00:15:18
pool there. The this was shocking to me
00:15:21
when I got there, but so many of the uh
00:15:23
young people even up to Gen X, our
00:15:26
generation, uh have been educated in the
00:15:28
west. In fact, the woman who we are in
00:15:31
partnership with uh with Sonabble um has
00:15:34
a MBA from Stanford and Oxford and all
00:15:38
these incredible schools and they've all
00:15:40
come back after these incredible
00:15:42
scholarships that are given uh by the
00:15:44
country. talk about the talent that has
00:15:47
come back to Saudi and the application
00:15:50
level level and Golden Visas as well
00:15:53
just how open they are to having
00:15:55
companies set up shop there and there's
00:15:57
a bit of a competition is there not
00:15:59
between you know DOA Dubai Abu Dhabi
00:16:03
Riyad Bahrain there's a really great
00:16:06
competition that's emerging for talent
00:16:08
and you've got the entire country all
00:16:10
the nationals very engaged in starting
00:16:13
companies
00:16:14
yeah So um I mean this is such an
00:16:16
important point because I think this is
00:16:19
um a key fact that people don't realize
00:16:22
yet.
00:16:23
I didn't Yeah.
00:16:24
Yeah. The if you look at it, you know,
00:16:25
30 years ago, the investment and the
00:16:28
strategy for Saudi Arabia to send many
00:16:31
of their citizens abroad across the
00:16:33
world to pursue degrees is really paying
00:16:36
off um you know in a in a really large
00:16:39
way. Um again back into the example I
00:16:42
mentioned about my um my models team I
00:16:46
have 40 PhD students uh not students 40
00:16:49
PhD scientists today in my team I
00:16:52
sometimes honestly when I sit with them
00:16:54
I said I'm really maybe I don't have
00:16:56
enough credentials to sit with you guys
00:16:58
in the room but they're graduated from
00:17:01
uh yes some are for PhD from Stanford
00:17:03
MIT Harvard Oxford you know from
00:17:07
anywhere that you want they're really
00:17:09
have obtained now the foundation
00:17:11
knowledge to really take to that next
00:17:14
step. Now, when you talk about
00:17:15
competition, I mean, the beauty
00:17:18
when they graduated, was it that they
00:17:20
were thinking about staying in the US,
00:17:22
getting an OPT, staying, or was it this
00:17:24
pull to go back and say, "Okay, we're
00:17:26
now trained. Let's go back and help our
00:17:29
mother country." What What is the
00:17:30
motivation of this?
00:17:31
I think there's a couple things that
00:17:32
happened. I mean, I I mean, this is my
00:17:34
own view and I talked to my colleagues
00:17:36
in Saudi Arabia. I really think that
00:17:39
they had a sense of responsibility to
00:17:43
take in the build process of what
00:17:45
tomorrow is going to look like. Uh I
00:17:47
would have done the same if I was in
00:17:48
their shoes. I would have went back and
00:17:50
really helped into this uh this build
00:17:52
process. So um they learned a lot and
00:17:55
mostly from I would say United States
00:17:57
and uh UK and now they all came back
00:18:01
looking at the opportunity because the
00:18:02
inflection point has happened. If you
00:18:05
look at the fast-paced transformation,
00:18:09
it's an opportunity I don't think they
00:18:10
want to miss. It's really changing
00:18:12
rapidly.
00:18:12
How much of that maybe is a desire to
00:18:15
work in the era of MBS and maybe you can
00:18:17
use that as a jumping off point to tell
00:18:19
us what is it like to work with this
00:18:20
person that's young and beloved?
00:18:22
So, so a couple of things. I mean, look,
00:18:24
I I'll tell you my uh my honest opinion.
00:18:27
I mean first of all the first meeting
00:18:29
that we had with him um key observation
00:18:32
and takeaway his vision blew my mind. I
00:18:35
don't think I have seen someone with
00:18:37
such a bold vision towards what the
00:18:39
future looks like. His responsibility to
00:18:42
the citizens of Saudi and the uh
00:18:44
diversification and the transformation
00:18:46
initiatives that are happening is second
00:18:48
to none. It's really really incredible.
00:18:51
Actually I
00:18:53
was delayed yesterday. I was supposed to
00:18:54
be here, but he the man doesn't sleep
00:18:57
really. I mean, he's he's on all the
00:18:59
time uh thinking about initiatives that
00:19:03
needs to be done and fasttracking them.
00:19:05
So, um everybody now is really rallied.
00:19:09
If you go into Saudi Arabia today, one
00:19:11
thing that is surprising, you take a
00:19:13
taxi cab or you go to a hotel or you
00:19:16
talk to a government official, everybody
00:19:19
is intertwined with the vision 2030
00:19:22
mission. everybody. So this is something
00:19:24
that rallied the society around an idea
00:19:27
and the idea has incredible merits
00:19:29
especially I would tell you
00:19:31
and is it both the old and the young
00:19:33
it's I mean what I'm seeing now across
00:19:35
the board even when I talk to the older
00:19:38
generation or the younger generation
00:19:41
this mentality shift transformation has
00:19:43
happened now I'm I'm not representing
00:19:45
everybody I'm telling you the segments
00:19:47
I'm interacting with um I I see nothing
00:19:50
but optimism but you guys know very well
00:19:53
better than me.
00:19:53
This is such a key point. That was what
00:19:55
I took away from when I started spending
00:19:57
time there over 3 years. I saw the
00:19:59
change every year. The enthusiasm level
00:20:01
reminded me of New York and San
00:20:03
Francisco in the '9s, early 2000s. The
00:20:07
opportunity to build anything. The
00:20:09
culture was on fire. You've got movies
00:20:12
and cinemas coming back and and all this
00:20:14
great stuff happening. And they're so
00:20:16
engaged and they want to learn how to
00:20:18
build businesses. And they've got this
00:20:20
real, you know, the 2030 mark has
00:20:22
inspired everybody to think, well, what
00:20:24
can what's possible? We we have this
00:20:26
incredible energy business. Sure. But as
00:20:28
they explained it to me, we can take
00:20:30
that energy business and over the next
00:20:32
5, 10, 20 years, we can transform this
00:20:35
entire region and be a global leader.
00:20:37
And it's really inspiring.
00:20:38
And and uh to add to your point, look
00:20:40
back into your question about
00:20:42
competition. You know, this is really
00:20:43
interesting. You see um there's a lot I
00:20:47
have learned and I owe a lot for what I
00:20:49
learned in the US. Quite a bit a lot. Um
00:20:53
I learned innovation here. When I moved
00:20:55
to India, I learned scale. When I moved
00:20:57
to Japan, I learned precision and
00:20:58
quality. You come to Saudi Arabia and
00:21:00
I'm learning optimism and vision truly.
00:21:03
And um hopeful,
00:21:05
you know, I I find it to be really
00:21:08
remarkable. I mean, that's I'm I'm
00:21:10
living there almost full-time. Um and
00:21:13
what I have done if you come to humane
00:21:15
office today in my opinion it is no
00:21:18
different than a Silicon Valley office
00:21:21
not at all. Oh yeah it.
00:21:22
So I encourage you come visit see how
00:21:24
you mean from the basics of the basics
00:21:27
how the building layout is how the open
00:21:30
office design is to remove barriers and
00:21:33
thoughts mentalities and I want now
00:21:35
really my team has to become product
00:21:37
creator not just a reseller. So that's
00:21:40
very important. Tark, as you are
00:21:43
building your infrastructure, there's
00:21:45
supply chain and strategic partnerships
00:21:47
and relationships that I'm sure emerge
00:21:49
here in the United States, but also in
00:21:51
China. And Saudi Arabia seems to be in
00:21:54
this really kind of interesting position
00:21:56
as a large energy supplier, as a large
00:21:58
partner in capital and now in building
00:22:01
that could create tension between the
00:22:02
rivalry, the global rivalry between the
00:22:04
US and China. How do you think about
00:22:06
managing each of those two markets and
00:22:08
how you establish relationships and
00:22:10
where do you align yourself? So if you
00:22:11
if you see um how humane was launched
00:22:14
and and I wish I could take a small
00:22:15
credit for this but u I've asked for one
00:22:19
thing before President Trump came to
00:22:21
visit um Saudi Arabia we should you know
00:22:24
please launch the company 3 p.m. core
00:22:27
because we wanted to see ensure that the
00:22:30
alignment that we are going after we're
00:22:32
going after really where the innovation
00:22:34
is the talent that exists the
00:22:36
infrastructure that exists and let's be
00:22:38
really realistic and truthful today the
00:22:41
US is leading especially on the
00:22:43
semiconductor side and we don't want to
00:22:45
miss this opportunity so if you see our
00:22:47
partnership that we have done with AMD
00:22:49
with um a startup called Grock uh Nvidia
00:22:53
Qualcomm it shows you clearly our
00:22:56
commitment to this relationship and
00:22:58
partnership and it's really deep not
00:23:01
only on the silicon side but the same
00:23:03
thing you're going to hear very soon on
00:23:04
the AI and the software application and
00:23:07
that's why humane to succeed really
00:23:10
needs deeper engagement relationship I
00:23:13
spend a lot of time with um with
00:23:16
startups in the US and um you know and
00:23:19
as long as to be very honest with you
00:23:21
the hope and the optimism I have and and
00:23:24
David and the team has been doing really
00:23:26
a great job with this is our ability to
00:23:28
say we understand very very well the
00:23:31
concerns that one might have. But if you
00:23:33
see about my partners that we have
00:23:35
selected from software layer in the
00:23:37
cloud to manage tenant management to the
00:23:39
security on the data center to make sure
00:23:41
that these servers are secured. We will
00:23:44
do everything that is required with the
00:23:46
optimism that humane will be thought
00:23:48
through as a trusted supplier for um for
00:23:50
the US. David Saxs, how important is the
00:23:55
American and the kingdom's relationship
00:23:58
globally and for humanity in terms of
00:24:00
the the president's uh agenda and his
00:24:04
prioritization? Well, it's been a
00:24:06
critical relationship for the US and for
00:24:09
Saudi Arabia since I think 1945, if not
00:24:12
before when uh the the founder of Saudi
00:24:16
Arabia, the king Iban Sad met with uh
00:24:19
our president, our king so to speak, FDR
00:24:22
on a battleship and they hammered out
00:24:25
the foundation of the modern world which
00:24:27
is the US would provide security for the
00:24:30
region in exchange for the steady flow
00:24:31
of crude and that you know that was I
00:24:34
FDR did that on his way back from Yalta
00:24:36
and people don't know as much about that
00:24:38
as they do the Yaltta meeting but that
00:24:39
was a very important understanding and
00:24:41
then the relationship's evolved over the
00:24:42
last 80 years but what I can tell you we
00:24:44
I went with the president on the trip in
00:24:46
May to the Middle East and number one
00:24:48
like you said the business culture in
00:24:50
Saudi Arabia is very Americanized many
00:24:53
of the the the Saudi elite have studied
00:24:56
in the US second they want to have a
00:24:59
good relationship and a and a
00:25:00
partnership with the United States
00:25:02
there's nothing competitive at all all
00:25:04
about that relationship. And third, when
00:25:06
it comes to high-tech and AI, they want
00:25:08
to be part of our technology ecosystem.
00:25:10
When I got back to Washington, I was
00:25:12
really surprised at how controversial it
00:25:14
was that we wanted to do business with
00:25:17
the Gulf States and Saudi Arabia in
00:25:19
particular. The way I see it is very
00:25:21
simple. Saudi Arabia is going to have
00:25:23
data centers. Of course, uh, every
00:25:26
sovereign country that can afford them
00:25:28
is going to have data centers. Is that
00:25:30
going to be American technology or
00:25:32
Chinese technology? It's basically going
00:25:33
to be our uh companies or it's going to
00:25:36
be Huawei.
00:25:37
It's binary.
00:25:37
It's binary. And why would we want to
00:25:39
push any country into the arms of
00:25:41
Huawei?
00:25:42
Especially when their preference is to
00:25:43
work with America. Well, I think it's
00:25:45
especially for two reasons. One is the
00:25:47
proximity that that region has to 4
00:25:50
billion people. Because if you think
00:25:52
about building data centers, ultimately
00:25:53
inferences, there's a certain window of
00:25:55
time and you have to be under several
00:25:58
hundred milliseconds. But if you do that
00:26:00
and you draw that radius, you're
00:26:02
counting half the world's population. We
00:26:04
we can't strategically ignore that.
00:26:07
Otherwise, you'll be forced to do it
00:26:08
with somebody else.
00:26:09
We'll let talk in a second, but just I
00:26:10
think that these restrictions on the
00:26:12
region, they were placed in October of
00:26:14
2023. And at the time by previous
00:26:17
administration, at the time that
00:26:19
happened, it was justified on the
00:26:20
grounds that the US was the only game in
00:26:23
town. We were the only ones who could
00:26:24
really make advanced semiconductors. So
00:26:26
therefore, we could impose whatever
00:26:28
restrictions we wanted. Nobody would
00:26:30
have a choice. But since then, if you're
00:26:31
reading the headlines over the past few
00:26:33
months, it's all been about Huawei,
00:26:36
Cambercon, uh, SMIC. China is rapidly
00:26:40
advancing. Um, Dylan Patel from, um,
00:26:42
semi analysis just had a report. By next
00:26:45
year, China is going to be making
00:26:46
millions of chips. Admittedly, they're
00:26:48
not as good as American chips. But if we
00:26:51
deny the rest of the world the ability
00:26:52
to participate in the American tech
00:26:54
stack, then they will participate in the
00:26:56
Chinese tech stack. And I think the
00:26:58
question of what what we sell to China
00:27:00
will always be a complicated question
00:27:01
for obvious reasons. But when it comes
00:27:03
to the Middle East and the rest of the
00:27:04
world, I think it should be an easy
00:27:06
question that as long as these countries
00:27:08
are abiding by our security requirements
00:27:11
and they want to be partners and allies
00:27:13
of the United States, we should allow
00:27:15
them into the American tech ecosystem
00:27:17
because otherwise we're just creating a
00:27:19
Huawei belt and road. Maybe uh one one
00:27:22
uh important thing. Um we got approval
00:27:26
last year and and this is a really
00:27:28
really important case study for you.
00:27:30
When we looked at our business model,
00:27:31
actually the number we use is exactly
00:27:33
4.4 billion.
00:27:35
4.4. Sure.
00:27:36
We could reach within the tolerance of
00:27:39
fiber latency that inferencing would
00:27:41
become really still um ultra responsive.
00:27:45
Um we got approval after a few months on
00:27:49
um a startup in the US and this is a
00:27:51
classic example. I mean you could really
00:27:53
validate what I'm going to tell you now.
00:27:54
We supported this US startup. look at
00:27:56
where they where they are now today and
00:27:58
what their valuation post engagement
00:28:00
with us. So we picked um at that time
00:28:03
for inferencing Grock because I thought
00:28:05
it was really an interesting startup
00:28:06
that could help us democratize the cost
00:28:09
of inferencing out of um the data center
00:28:12
that we have deployed this in. Um we
00:28:15
have about 19,000 uh of of their chips
00:28:18
deployed overnight.
00:28:21
130 countries are using now this
00:28:24
inferencing cluster. 5% of the traffic
00:28:26
from Saudi, the rest is from abroad.
00:28:28
Why? Because we're able to offer
00:28:30
differentiated inferencing costs and we
00:28:33
offer to the world. Now keep in mind
00:28:35
what we also did for security
00:28:37
guarantees. We said Grock cloud, you
00:28:40
manage this. So now I don't have to
00:28:42
worry about KYC requirement.
00:28:45
You follow US rules. I what I care about
00:28:48
is I have revenue that I participate in.
00:28:50
And it was really a win-win for both of
00:28:52
us. And I hope we'll repeat the same
00:28:54
thing with uh with the deployment that
00:28:56
we're doing with Nvidia and uh AMD and
00:28:59
then something really really exciting
00:29:00
that we're going to do with Qualcomm on
00:29:01
the edge as well.
00:29:02
Thank you so much for joining us and see
00:29:04
you at Mamadori.
00:29:05
Thank you very much. Thank you.
00:29:07
Appreciate it. Thank you. Thank you very
00:29:08
much. Thanks. Thank you.
00:29:10
I got a great schwarma place for us.
00:29:12
Thank you. We're in town.
