YouTube Transcript:
Global AI Investment
Skip watching entire videos - get the full transcript, search for keywords, and copy with one click.
Share:
Video Transcript
Available languages:
View:
I'm Ena freed Chief technology
correspondent at axios actually had a
dream last night I was hosting SNL so
this is probably as close as I'm going
to get um if you haven't subscribed to
my newsletter I write our daily AI plus
newsletter you can go to axios.com
newsletters sign up for mine Jims miks
all the good stuff um I'm really excited
uh that we're GNA have another great
conversation around where really are we
in this AI moment huge moment of change
I don't need to tell anyone that but
also you know I've been covering Silicon
Valley for 25 years tremendous amount of
change I've never seen the actual
technological change happen as fast but
then right on top of that we have yet
another hype layer that I'm very
familiar with and so it's really hard
when the technology change is this real
and at the same time there's so much
hype so we're going to have a great
discussion one last uh housekeeping note
if you haven't signed up for our last
event of the week um it's taking place
at the c3.ai booth uh which is close to
the Congress Center on the prominade
I'll be interviewing Eric Bolson and
Nina Singh and today we're gonna talk
about how investment in infrastructure
obviously a huge piece of this is
getting all that massive compute that
we're going to need and I'd like to
bring up my colleague John Otto to kick things
[Applause]
off good morning everyone uh day four of
Davos if you can believe believe it
thanks for for starting uh with us here
today I'm really excited for for the
program I'm John AO the GM of axios live
uh huge thanks to qualcom house our
venue partner um but programs like this
are not made possible without the
support of our Underwriters and and
partners and so uh like today in this
morning um mgx and mubadala Investment
Company have been a huge supporters of
us today so I'm excited to uh introduce
to the stage Chief investment officer of
Technology at mgx Osman come on up
sir and Deputy Chief Deputy Chief
strategy officer admad Investment
Company Mark Andi [Applause]
[Applause]
Mark all right um so let's start off big
picture here mubadala uh supported the
the launch of mgx and and really want to
understand a little bit more for the
audience here what's the relationship
like what's the the working relationship
with with both groups thank you very
very much joh and good morning everybody
pleasure to be here maybe a a quick
intro on mobad for those that don't know
mad we're Abu Dhabi based Global
Investment Company portfolio size 330
globally Diversified and the way we
invest we invest with top tier Partners
both directly and indirectly in teams
and sectors and geographies where we
have deep convictions things like
technology where we deployed capital in
semiconductor and software over the past
10 years and Healthcare Life Sciences
also in the in the past 10 years clean
energy and and and what happened over
the past uh couple of years with that
inflection point of gen the acceleration
a lot of thematics in the AI space we
decided to have a dedicated focused team
uh vehicle to look at the AI space and
go at the speed that that space actually
requires we have have built tremendously
fantastic relationships Partners access
depth of expertise across the 20 years
of capital deployment we've been doing
but that space in particular requires
dedication focus and active uh
deployment I'm sure that my colleague
Ali is going to speak a lot about mgx
and the cool stuff that we're doing but
aside from the the mgx uh story The
thematic is also driving a lot of
acceleration on the other thematic where
mubadala is involved so we continue to
push on both fronts dedicated the AI
ecosystem the other areas as well and
certainly benefit from that close
proximity and collaboration and ali uh
talk to us a little bit more mgx is
strategic AI investment goals for this
year um help the audience understand
that a little bit more and um so you
know I think Mark said it well we
basically about 18 months ago took a
step back and really wanted to redefine
what it meant to be an investor an AI
first investor or an AI investor uh in
an AI native environment um and we came
to the conclusion that you needed to
have the ability to invest across the
entire thematic and ecosystem that meant
you needed to invest across asset
classes acoss across risk categories you
needed to be able to bring scale uh to
the table that that is very important
given the capital needs of the space uh
and so we we built mgx to be an a first
and AI thematic fund with the ability to
invest across semiconductors across AI
infrastructure including data centers
comput as a service and eventually token
as a service and intelligence as a
service as well as the entire um entire
Tech stack right so thinking about
Frontier models obviously looking at
every layer as well as software
businesses that ultimately will be
transformed and so as we think about the
enablers of this technology you know
taking a 12 to 18 even 20 24mth view the
infrastructure and the infrastructure
buildout is going to be a critical piece
of the puzzle that's really where we're
focused today and in terms of like where
do you see the most opportunities for
applications and industries how do you
look at at that from from an investment
standpoint I think one thing we have yet
to see um from an application
perspective is really um is is is really
where Enterprises are pursuing growth
and growth opportunities rather than
efficiency in deploying AI applications
so we're starting to see a lot of
enterprising we heard this across I mean
throughout the week a lot of Enterprises
have started to look at efficiencies and
optimizing the way they do things today
leveraging technology in order to be
more efficient the reality is we think
that the opportunity is actually how do
you bring value forward by looking at
growth how can you serve your customer
better how can you unlock customer Roi
as the enterprise we think that we're
very early days uh with respect to that
and and mark so while there is every bit
in ttention around uh what UA has to
offer there is also every bit of a a
global opportunity for for AI so in
terms of like Global presence and Global
Focus how does MO look with
that so we we're in many many different
geographies Let Me Maybe take them uh
west to east so the US North America the
us we are 40% of our portfolio is in the
US we've been building that portfolio
for quite a while it's been playing out
extremely well like I said we build
great relationship great businesses
they're performing extremely extremely
well we continue to perceive the us as a
great Market to attract Capital it's and
the reasons are pretty known uh the
depth of the market the quality of De
flow the partners that we have the the
capital markets that are extremely
attractive but also if you just look at
that drive on Innovation and technology
is fueling that productivity drive that
you see translating into a lot of great
businesses going going from local to
Global Etc so there's a lot of exciting
thematics in the US that we continue to
underwrite the consumer space the new
consumer space the the the the young
Generations coming Upstream uh healthc
care is Big energy is Big the
infrastructure as a whole is uh very
interesting and so forth so we continue
to look at the us as a prime Market if
we go a little bit more towards the east
Asia Asia as a as a whole also we have
more or less 40% of our portfolio in
Asia including the UAE UAE at the end of
the day is in Asia is a big growth
market for us and I'll speak a bit about
the UAE in a couple of minutes but Asia
is a place where we've been deploying
capital for the past 10 15 years we see
huge growth opportunity on the back of
Simply demographics just look at the
pace of growth population young
Generations coming online think about
all the reforms that have been happening
in call it India other places and how
it's attracting more and more business
FDI Capital formation Etc so in Asia and
Asia is Big we cannot speak about Asia
as one one country or one geography
India has been a bright spot for us
we've been doing quite a lot there and
we continue to uh deploy capital in
India China we have a capability that we
build there for the past 10 years um we
all know the challenges that are ongoing
and there's a lot of reforms coming but
we're extremely well positioned in China
to be able to scale at the right time
because we are early positioned whereas
a lot of our peers or other Capital
deployers have pulled out from China
there's a lot of other bright spots in
in Asia that we're also contemplating
but let me stop on the uee because uee
is a very very exciting grow story a lot
of you know how the U has developed over
the past 20 30 years and has put
everything in place that is needed to be
able to continue that forward leaning
journey of innovation of vibrant economy
and really focusing on the future
sectors so we in the U give or take at
20 25% have great presence but the UEI
is a place where we're significantly
deploying Capital as well to continue
help building the the economy the
country but also simply because it makes
sense we are coming up on time we're
going to get the hook but we're in Davos
let's pretend like we have a crystal
ball here I what's one big thing you're
most excited about for this
year I think from what we're seeing um
picking up on your point on applications
the evolution of the application layer
um I think will be extremely material in
2025 uh from some of the technologies
that we're already seeing being rolled
out over the course of the next two
months um it will transform the way we
work today and the embracing of that
technology and of the application layer
by the Enterprise again from a growth
standpoint rather than an efficiency
standpoint is really what I'm most
excited about in 2025
Mark certainly application and things
coming real but I'll take maybe a
different spin I think one thing that's
going to be very interesting is
partnership and collaboration and we all
hear that fragmentation isolationist and
certainly is going to have an impact but
throughout my entire week at Davos
partnership is critical not only to
navigate that complex world because it's
very clear that nobody can do it alone
and at the same time as these
partnership actually form and
enhance these Partnerships can help
shape a bit that World of Tomorrow in a
different way that these Partnerships
were were
in the past so very much looking forward
to it uh that's all the time we have
Mark Ali thank you so much for the
conversation today and a huge thanks
again for the support mgx and mubadala
Investment Company so uh Eno will be up
next for our next conversation thanks
for joining thank you very much thank
come on up
Christian so as I said in my
introductory remarks you know one of the
things everyone is trying to figure out
here in Davos is just just where are we
with this AI Revolution you know I I
don't know if you heard me backstage but
I was saying you know there's real
change here and it's super fast and I
think businesses are struggling to keep
up and I'm sure you guys aren't
contributing to the hype but I think
there's also a lot of hype on top um I
thought maybe a good place to start is
to talk about what are businesses doing
today with AI with you what's working
let's start with what's working what are
businesses doing today with generative
where where are you seeing the most
success the most value driven that sort of
of
thing I absolutely I mean look u j and
so first of all I guess it's very
important Ena to understand 2023 was the
year of you know we had the first ideas
of what generative AI can do for
businesses not for you know triple
consumers or in our private life 2024
was then really the adoption of the
first use cases so that companies feel
the value uh but you know these are use
cases in uh take for example HR yeah you
white thousands of chob profiles chob
postings every year of course can be
taken over we closed which is good
thousands of contracts by the end of the
year with customers 800 Pages completely
now automated refre checks compliance
checks data security over regulation in
Europe checked by you know Ai and then
of course
sh yeah see yeah you know a completely
new topic and uh now and look these are
the use cases where customers say you
know what that is why I absolutely
willing to also lean in and adopt this
new AI use cases now of course when you
now I mean Grandma next next week
becomes you know 95 uh when I ask my
kids to write a poem they ask judge gbd
and get a good result the problem what
is still is and this is I guess what we
now have to solve in 20 25 to not only
feed into the hype is really you know
these llm modules they have for good
reasons no access to business data but
in businesses you need 100% accurate
results I mean the poem for Grandma can
be 95% good and 5% okay that's is
absolutely fine in business you need
100% accurate results and where many
companies still struggle with is I have
my co-pilot here SCP has its Jewel and
then we have this and that but you know
how do I bring this together and because
only when I have unstructured structured
data together SCP non SCP data together
I really understand better with AI My
Consumer Trends the consumer wishes I
can thr grow by really detecting earlier
is this Adidas sneaker now just you know
more in the in the demand or less and
then I can adjust my supply chain and
here we come and here we come to oh and
here we come to the the agent story and
I guess that is now the next thing where
also technology companies like SCP data
bricks and others need to team up in
order how can we really build this data
foundation for AI well and and you hit
on a really important thing and one of
the things I've been asking which is if
the if the llms aren't 100% accurate and
right now we live in a world where
they're generally returning information
to a human being those contracts
probably someone's giving it a final
read hopefully but when we talk about
agents we're talking about more autonomy
uh is the technology ready to be given
agentic capabilities and if so what
types of Agents cuz we've also been
having a conversation people mean really
different things when they say agents
yeah um look it's a good question I mean
I can share with you two agent agentic
AI capabilities we are going to launch
this year I mean first yeah when
customers using our software to package
their services their products to price
it hopefully you know in the right way
uh to position it at a wi point of time
so there there's a sales agent you know
helping you with you know a lot of
recommendations on what is the best
price how to package this up how to
bundle it best to really hit the
consumer at the right point in time
that's is a sales agent now the sales
agent would be great if the sales agent
can talk to the supply chain agent
because is this product on stock can I
deliver on time so that the sales agent
is not making any kind of promises you
know which you know the supply chain
can't fulfill and that is now one of the
scenarios what we are planning because
we see there is clear value in it and of
course is there then still Jewel on top
who has to orchestrate that that
absolutely and that's why we are
investing not in an llm but SCP is
investing in a knowledge graph in really
you know having to contextualize the
data yeah that you can set it in context
that is very very important and then we
we need to bring it to adoption yeah and
uh and that's why we really focus on a
few use cases not on a 100 but on a few
push this out and then I I feel this
will be a big next step in the evolution
of AI and talk about what you're hearing
from customers because I think you know
nobody's saying I don't think we need AI
or correct me if I'm wrong I doubt
anyone saying oh we don't need AI but
what I hear a lot is I don't really know
I mean they they really don't know how
to use it they've spent the last two
years they came to Davos two years ago
oh my God Chad GP is going to do it they
had chat jpt right their terrible
speeches all that the next year they
were launching lots of Pilots all over
the place proof of concept I know they
want to get returns but I also think you
know the same company that was trying to
sell them you know a chatbot last year
is trying to sell them on agents this
year they haven't even gotten through
with the last thing H how are
businesses what are they telling you
they need from you I mean yeah
absolutely now I mean look we also have
to ground ourselves a little bit in
reality and the reality is I mean 80% of
our customers are not the large
Enterprises who are not having a big it
Department who are not having 100 data
scientists so what I told our teams is I
mean we are SC we are running with our
software the business so we need to
embed it we need to offer that out of
the box meaning we have 400,000
customers 40,000 customers giving us the
right to train our modules with their
anonymized data of course and then we
plug it in yeah so we have all the
predictions all the automation it's
coming with the software because often
times customers just don't have you know
the team also the Investments yeah to
take a piece of AI and bring it into the
context of the business and this is
where we are sitting at the Nexus and of
course we need to take make use of that
we five years back we developed an iot
scenario here machine learning scenario
there and then you have a piece of for
software there customers really
struggling with the adoption so that's
very important that it comes embedded
and then the second piece is you know
what I mentioned before is also on the
data side I mean we also have to say
there's a lot of Legacy still out there
there are a lot of data silos I mean
there's always this imagination oh I
take my sep nonp structured unstructured
in a data Lake magic happens and somehow
I have a meaningful data layer for the
company that is not the reality and then
for good reasons the llm modules they
can't read your financials because this
would be super bad I would be
uncompliant if all of my end users
suddenly could read a group pnl of sap
so this is something what we have to
solve yeah and this is where we work
with our customers on hey how can we
build a better data Foundation how can
we harmonize the generative AI use cases
with your business data and these are
the challenges what custom customers
today facing and that's why we also
working a lot under the hood on these
things I often think tech companies
learn the most from what they're doing
internally cuz you know whether it's dog
fooding their own technology or using
others technology and I'd love to hear
you mentioned one use case uh you know
contracts and some of that stuff I know
you've used it some in marketing um I
assume you're using it for coding for HR
where where are you getting the most
value what are some things that might be
unexpected where saap's internal use of
AI has been interesting yeah I I mean I
get my weekly report and uh about the
effectiveness of our AI on our business
I mean we have to prove sep once's sep
very important to also have credibility
and in development I mean now it's
budget time I mean if our development
managers like it or don't like it they
have a 30% you know efficiency gain in
there in producing more code with the
same team and it's absolutely doable we
were I mean a year back with our AI
pilot for up up where it was probably 5
to 10% now we absolutely see the the
code which we produce or AI prod
producing can absolutely lead to 30 so
they have to show you they're doing and
we and we see it and honestly to make a
a typical developer excited about these
tools take some time but you can feel
now also the excitement yeah that they
have to you know deliver now more output
but they also have the tools in their
hand so that is really software
development and honestly when you look
at the workforce of SCP be it in HR be
it in Cloud delivery be it in sales in
marketing I mean we are doing a massive
enablement program now I mean you know
we cannot you know hire you know 100,000
you know new employees so you need to
invest into your Workforce and you have
to think through what is the profile of
a developer now and how will it look
like in three years from now and the
same in my eyes you have to do in order
to equip your Workforce for this uh you
know next age of AI where hasn't a AI
been as effective as you thought what
are some areas where you've tried to use
AI where either it hasn't worked as well
or you said you know I'm sure you tested
out things that didn't didn't pan out
the way you thought I have to say um
what is sometimes more an obstacle also
inside SCP is um data protection and my
lovely data Protection Officer yeah so
um you know I can accept Okay we can't
do that but what is then the solution
yeah so on how to use data yeah for
testing Etc and then we also see
Regional differences yeah what is
allowed in Europe what is allowed in the
US or what is allowed in Asia and we are
making investment decisions also R&D
wise based on you know what is where do
we feel we are over regulated and where
is actually also more freedom to drive
Innovation that is a constant battle
decision uh discussion we always have
also inside sap now of course when HR
and finance make up two or 3% of your
pel the impact is a bit less
nevertheless the CFO needs AI for
steering you have very volatile markets
maybe tariffs are coming inflation then
you have you know interest rates
changing and that is where AI can help
but of course when you look at the pure
cost I would say also for our customers
it's typically supply chain it's
typically Asset Management I mean this
is where a lot of the CeX and the Opex
is going in and obviously customers also
tend to invest exactly in that areas so
the business itself has got to be also
changing the way you sell you know one
of the things I was talking to you know
some folks at Microsoft and Mark
bennyhoff and I were actually talking on
stage you know historically a lot of
enterprise software software is a
service per user per month pretty pretty
well understood model doesn't really
work in the day where you know we're
going to have agents one human being
could be controlling you know 10 100
agents uh how do you imagine sap will be
sold differently a couple years from now
and priced and what other Beyond just
pricing because that's a packaging thing
you've you know there was a time we
bought software in a box and it was a
one-time license and you tried to sell
an upgrade so my guess is you can
navigate that one but how else do you
see your business
changing I mean commercially it's
absolutely a fair question I mean my
investors asking me the same question um
do you really want a price per user when
you want to make this process you know
lot more productive and will'll be there
actually less users at the end and I
feel it was a very good decision that we
right from the start said for Jew for
our AI use cases it's consumption it's
output based because I would say it's
also then a hard a tough ask also some
in this um you know environment to say
oh I charge you a 30% pre premium on top
of your user price and it was also very
good for us to learn to have really an
educated discussion with the customer
about the value of these AI use cases
and price them outcome based consumption
based and so that was for us absolutely
the right decision so we are already on
outcome output based pricing of course
our Solutions the software is still
subscription based I would say the
tendency there is also to come to more
outcome based but you also have to see
many cios yeah they all love to hear
about outcome based and but you also
need to have Security in your it budget
yeah when you you know overshoot your it
budget by 20 or 30% your job is also at
risk so I would say look I mean outcome
based is absolutely the way to go
consumption on AI if this will now
radically change on the software side I
don't know we only have about a minute
left what aren't we talking enough about
here in Davos what is missing from the
conversation I mean you said it right
you know I mean at the end you know I
mean I I would love to say buy a piece
of AI and technology and you know get
the value in l i mean that is not not in
the Enterprise base how it works you
need to understand how will my business
will run in the future how is my
business model changes how how do I want
to automate a certain business process
and then drive the change from the end
user coming and then apply the AI in the
right way and so I I guess this is
something what also the technology
companies need to take care of there's
always an end user there's always a
change needed on the business side and I
get guess this honesty we also have to
give our customers because otherwise we
are playing into a hype and we are here
with aend AI and our customers are here
and I always remind our teams the Gap
there is always a certain Gap you know
until customers can adopt but the Gap
needs to be well managed and cannot be
you know like a two or three year time
frame and that I guess is very important
that also here in daros we talk about
how to change the business how to change
my Workforce in order to apply AI with
the wide output all right well I want to
give you one suggestion for an agent sap
is very involved in the San Jose Sharks
they play at sap Center don't go there
my founder gives me a hard time with
they are not playing really well yeah uh
they do have a great young player mlin
celebrini is great anyway we're gonna
have to leave it there hopefully we'll
have better news on both the AI and
hockey fronts next year Kristen Klein
from sap [Applause]
[Applause] [Music]
Click on any text or timestamp to jump to that moment in the video
Share:
Most transcripts ready in under 5 seconds
One-Click Copy125+ LanguagesSearch ContentJump to Timestamps
Paste YouTube URL
Enter any YouTube video link to get the full transcript
Transcript Extraction Form
Most transcripts ready in under 5 seconds
Get Our Chrome Extension
Get transcripts instantly without leaving YouTube. Install our Chrome extension for one-click access to any video's transcript directly on the watch page.
Works with YouTube, Coursera, Udemy and more educational platforms
Get Instant Transcripts: Just Edit the Domain in Your Address Bar!
YouTube
←
→
↻
https://www.youtube.com/watch?v=UF8uR6Z6KLc
YoutubeToText
←
→
↻
https://youtubetotext.net/watch?v=UF8uR6Z6KLc