Nvidia and Georgia Tech announce first AI supercomputer for students

Nvidia
and
the
Georgia
Institute
of
Technology
announced
Wednesday
the
first
artificial
intelligence
supercomputer
designed
for
student
use.

The
chipmaker’s
graphics
processing
units
are
at
the
heart
of
the
large
language
models
created
by
OpenAI,


Alphabet, Meta and
a
growing
crop
of
heavily-funded
startups
all
battling
for
a
slice
of
the
generative
AI
pie.
Nvidia
has
seen
its
revenue
soar
over
the
past
year,
and
its market
cap has
surpassed
those
of
Alphabet
and Amazon.

The
move,
a
way
to
democratize
access
to
supercomputing
resources
generally
earmarked
for
tech
giants
and
connected
startups,
underscores
the
importance
of
training
the
next-generation
workforce
on
AI.

At
first,
only
Georgia
Tech’s
undergraduate
students
will
be
able
to
use
the
computing
cluster,
which
is
fueled
by
Nvidia’s
enterprise
AI
software
and
a “virtual
gateway”
developed
by
Penguin
Solutions.

So
far,
students
have
used
the
supercomputer
in
just
one
class
—
foundations
of
machine
learning
—
since
the
beginning
of
the
semester,
but
Georgia
Tech
plans
to
continue
scaling
usage.
By
spring
2025,
all
of
its
undergraduate
and
graduate
students
will
have
access,
Arijit
Raychowdhury,
chair
of
Georgia
Tech’s
School
of
Electrical
and
Computer
Engineering,
told
CNBC
in
an
interview.

The
news
follows

OpenAI’s
decision
in
January
to
enter
into
its
first
partnership
with
a
higher
education
institution.
Arizona
State
University
received
full
access
to
ChatGPT
Enterprise
in
February,
with
plans
to
use
it
for
coursework,
tutoring,
research
and
more.

Georgia
Tech’s
supercomputer
runs
on
20
Nvidia
HGX
H100
systems,
which
house
160
of
Nvidia’s
H100
GPUs,
which
are
high
in
demand
across
the
tech
industry.
For
reference,
it
would
take
one
of
these
160
GPUs
a
single
second
to
come
up
with
a
multiplication
function
that
would
take
50,000
students
22
years,
according
to
a
Georgia
Tech
release.

Students
will
use
the
supercomputer
for
course
projects
related
to
computer
vision,
large
language
models,
robotics,
supply
chain
management,
chemical
or
biomedical
engineering,
creativity
and
design
pursuits
related
to
generative
AI,
and
also
separate
entrepreneurial
ventures
of
their
choice,
Raychowdhury
said.

“The
idea
here
is
for
us
to
be
able
to
set
up
these
computational
resources
…
working
closely
with
Nvidia
and
Penguin
Solutions,
and
essentially
making
this
a
sandbox
where
students
will
be
able
to
do
their
own
work,”
Raychowdhury
told
CNBC
in
an
interview.