Oracle’s Larry Ellison thinks every government will want to build a ‘sovereign’ AI cloud in the future

Every
tech
company
is
talking
up
its
AI
opportunity.
Oracle
is
no
exception.
But
during
an
earnings
call
in
March,
Oracle’s
Larry
Ellison
laid
out
a
future
market
opportunity
focused
on
a
major
customer
that
investors
may
think
about
less
often
that
Fortune
500
companies.

The
Oracle
founder,
former
CEO
and
current
chairman
and
chief
technology
officer,
sees
national
and
state
government
applications
being
run
on
platforms
like
Oracle
Cloud
Infrastructure
to
a
much
greater
degree
than
today,
and
indicated
that
it’s
starting
to
happen
in
a
variety
of
ways.

“We
talk
about,
you
know,
winning
business
with
companies.
For
the
first
time,
we’re
beginning
to
win
business
for
countries,”
Ellison
said.
“We
have
a
number
of
countries
where
we’re
negotiating
sovereign
regions
with
the
national
government.”

Major
tech
companies
vying
for
massive
government
contracts
in
the
cloud
are
nothing
new.
Microsoft
and
Amazon
had
a
lengthy
battle
over
a
cloud
deal
with
the
Department
of
Defense,
and
both
those
AI
players
as
well
as
Oracle
and
Google
ended
up
all
in
on
a
$9
billion
DoD
contract
in
2022.

But
Ellison
went
further
in
his
prediction
when
speaking
with
analysts

on
the
recent
earnings
call,
saying
“Every
government,
pretty
much
every
government,
is
going
to
want
a
sovereign
cloud
and
a
dedicated
region
for
that
government.”

Oracle,
which
works
with
Nvidia
and
Microsoft
on
generative
AI
capabilities,
has
already
helped
use
cloud
tech
to
cut
red
tape
for
countries.
One
example
Ellison
gave
was
Albania.
It
is

trying
to
ascend
to
the
European
Union
with
the
help
of
chatGPT,
with
the
generative
AI
helping
to
decipher
and
summarize
its
laws
and
aid
the
country
in
what
it
needs
to
change
in
order
to
be
compliant
with
E.U.
regulations.

“It
took
Serbia
eight
years
to
harmonize
their
laws
to
be
able
to
join
the
E.U.,”
Ellison
said.
“Albania
is
facing
the
same
thing,
but
with
generative
AI,
we
can
read
the
entire
corpus
of
the
Albanian
laws
and actually
harmonize
their
laws
with
the
EU
in
probably
more
like
18
months
to
two
years.”

Some
analysts
are
skeptical
of
Ellison’s
talk
as
being
anything
more
than
typical
C-suite
rallying
for
a
key
business
unit.
Oracle
shares
are
up
about
21%
YTD,
but
Barclays
analyst
Raimo
Lenschow
expressed
concern
about
lower
OCI
growth
during
its
latest
earnings,
which
could
“worry
investors,
as
this
is
the
main
investment
story.”

A
version
of
future
featuring
cloud
services
and
artificial
intelligence-powered
solutions
can
make
government
more
efficient.
Ellison
said
for
starters,
redundancy
is
a
focus
for
government,
in
the
case
of
disaster
and
disaster
recovery.
But
it’s
also
moving
into
health
care
information
and
internet
access
projects.

Countries
including
Serbia
are
standardizing
on
Oracle
Cloud
Infrastructure
and
using
generative
AI
for
processes
like
automating
health
care.
Deals
related
to
delivery
of
internet
services
in
partnership
with
Elon
Musk’s
Starlink
to
remote
areas
are
taking
place
in
Kenya
and
Rwanda,
where
OCI
and
Starlink
are
mapping
rural
farms
to
see
which
crops
are
growing
in
what
area,
and
if
they
are
getting
enough
nutrients
like
nitrogen
and
water.

“These
maps
are
AI-assisted,
help
them
plan
their
agricultural
output
and
predict
their
agricultural
output,
predict
markets,
the
logistics
of
the
agricultural
output,
doing
all
of
all
of
those
things
as
next-generation
national
applications,”
Ellison
said.

Food
security,
rural
school
and
rural
hospital
internet
access,
are
other
examples
of
what
Ellison
said
are
among
the
“all
sorts
of
interesting
new
AI
applications
out
there
that
you’ve
probably
never
heard
of
before,
at
least
I
hadn’t
heard
of
before
until
these
last
12
months
now
that
we’ve
worked
on
and
we’re
now
in
the
process
of
delivering.”

He
also
mentioned
automation
of
vaccination
programs,
and
other
healthcare
program
“across
the
board.”

“We’re
living
in
a
world
where
like
data
and
information
is
the
gold
of
the
future,”
said
Dan
Gardner,
CEO
of
digital
strategy
agency
Code
and
Theory.
“If
the
government
can
get
access
and
action
on
that
their
data
faster,
why
would
we
want
to
slow
that
down?
We
want
that
to
be
as
efficient
as
possible.
A
lot
of
that
is
like
mundane
human
resources,
that
maybe
those
people
could
be
doing
something
else
that
is
way
more
valuable.”

Cloud
and
generative
AI
applications
allowing
countries
to
give
rural
areas
internet
access
could
increase
educational
opportunities
and
create
more
economic
value.
It
could
also
allow
citizens
to
have
more
insight
into
government
processes,
said
Tapan
Parikh,
Cornell
University
associate
professor.
“One
thing
technology’s
always
been
good
at
is
potentially
making
bureaucracies
more
efficient,
or
at
least
more
transparent
internally,”
he
said.

‘Black
Mirror’
governments

But
the
push
to
move
more
government
processes
to
the
cloud
is
also
opening
the
door
to
new
risks,
especially
as
countries
trust
newly
developed
generative
AI
systems.
While
they
may
make
processes
faster
than
ever,
there
are
bound
to
be
mistakes
as
the
technology
develops
and
could
make
citizen
data
accessible
to
cyber
criminals.

“We
shouldn’t
use
these
technologies
as
an
excuse
to
not
maintain
oversight
and
control
over
political
processes,”
Parikh
said.
“Certainly,
I
think
that’s
a
very
important
thing,
particularly
when
you’re
dealing
with
countries
that
may
not
have
the
same
kind
of
governance
capacity.”

Oracle
did
not
respond
to
a
request
for
additional
comment
on
Ellison’s
earnings
call
discussion.

“There’s
the
‘Black
Mirror’
bad
side
of
it:
Big
Brother,
data
wars,
AI
warfare
and
all
that
stuff,”
Garder
said.
“As
far
as
like
removing
red
tape
and
being
more
efficient
and
getting
better
use
out
of
crops
across
the
country,
that’s
incredible.
That’s
the
multiplier
of
humanity
that
could
really
improve
because
of
AI.”

AI
raises
a
host
of
concerns.

Gardner
pointed
to
the
proliferation
of
more
generative
content
in
an
election
year
around
the
world
and
all
the
issues
related
to
tech-enabled
interference.
“Maybe
it’s
not
like
chips
on
the
ground.
But
it’s
data
security,
authentication
of
who
you
are,
who
governments
are,
what
content
you’re
viewing,
all
the
connection
points
between
financial
systems,
and
AI
governance.
Using
AI
as
a
tool
of
destruction
is
quite
scary.” 

“No
big
government
in
the
world
can
afford
to
move
all
of
their
services
and
especially
critical
ones
like
defense,
taxes,
health
care,
completely
into
the
cloud
and
into
the
hands
of
gen
AI,” said
Simone
Bohnenberger,
chief
product
officer
at
cloud
company
Phrase.
“It’s
just
not
in
the
realm
of,
I
think
it’s
not
responsible
to
do
that.
The
potential
risks
outweigh
the
benefits
of
doing
that.”

OpenAI,
which
created
ChatGPT,
is
mostly
trained
on
existing
content
on
the
internet.
That
could
pose
a
problem,
especially
when
text
from
lesser
known
languages
like
Albanian
need
to
be
analyzed,
Bohnenberger
said.

“If
you
look
at
the
World
Wide
Web
or
the
internet,
the
vast
majority
of
content
there’s
English,
I
think
a
quarter
of
the
content
is
English,
followed
by
Chinese,”
she
said.
“Albanian
is
a
minority.
It’s
very
questionable
for
me
how
well
that
actually
works
for
a
small
country
like
Albania
and
like
an
outlier
language,
because
there’s
just
not
much
data
you
can
train
a
model
on.
And
if
you
don’t
have
much
data,
then
the
outputs
will
be
very
messy.”

Then
there’s
security
and
data
risks
with
allowing
foreign
companies
access
to
citizen
data,
Parikh
said.
Even
the
U.S.,
with
all
its
resources,
has
been
vulnerable
to
data
hacks,
including
a
recent
February
incident
with
contractor

CGI
Federal
which
exposed
personally
identifiable
information
on
employees. The
recent
battle
between
the
U.S.
and
China
over
TikTok
is
an
example
of
how
control
of
sensitive
consumer
data
can
be
interjected
into
geopolitics.
“I
think
certainly
that’s
a
concern
going
forward
for
countries
who
are
working
with
vendors
from
different
countries,”
Parikh
said.