WEBVTT	#Elemental Media Engine(TM) 2.17.1.0

00:00:04.629 --> 00:00:08.466
heating up again, this time
over satellite internet.

00:00:08.717 --> 00:00:13.638
Project Kuiper is Amazon's
plan to deliver internet

00:00:13.763 --> 00:00:19.811
from space using 3,236
small satellites

00:00:19.936 --> 00:00:21.563
in low Earth orbit.

00:00:21.771 --> 00:00:25.483
In many ways, it's a
response and a competition

00:00:25.608 --> 00:00:30.113
to Elon Musk and SpaceX
with its Starlink network.

00:00:30.238 --> 00:00:34.534
Amazon is known as the
everything company, and it's

00:00:34.659 --> 00:00:38.413
hard to have an everything
company without internet.

00:00:38.580 --> 00:00:44.044
So Amazon's fastest-growing
segment has been its AWS

00:00:44.169 --> 00:00:45.378
cloud service.

00:00:45.503 --> 00:00:48.590
And in support of that,
they've built out a

00:00:48.715 --> 00:00:51.843
tremendous amount of
internet infrastructure,

00:00:51.968 --> 00:00:55.513
whether it's data centers
or fiber.

00:00:55.638 --> 00:00:59.976
Space is actually a very
natural expansion of 1.

00:01:00.101 --> 00:01:02.687
their data business and 2.

00:01:02.812 --> 00:01:05.982
their consumer business,
providing goods and

00:01:06.107 --> 00:01:10.403
electronics and resources
to people around the world.

00:01:10.528 --> 00:01:14.491
SpaceX's Starlink already
has about 2,000 satellites

00:01:14.616 --> 00:01:19.162
in orbit, serving about
250,000 total subscribers.

00:01:19.287 --> 00:01:22.624
But the FCC has approved
SpaceX to launch a total of

00:01:22.749 --> 00:01:24.751
12,000 satellites.

00:01:24.918 --> 00:01:26.795
Amazon has yet to launch
any.

00:01:27.962 --> 00:01:32.008
Amazon first revealed
Project Kuiper in 2019, but

00:01:32.133 --> 00:01:34.552
in early April, a big
announcement cemented the

00:01:34.677 --> 00:01:36.137
project's momentum.

00:01:36.346 --> 00:01:41.559
Amazon recently signed a
multibillion-dollar contract

00:01:41.684 --> 00:01:45.188
in what is the largest
rocket deal in the

00:01:45.313 --> 00:01:50.860
commercial space industry's
history, for launches of its

00:01:50.985 --> 00:01:54.739
Kuiper satellites with
three different entities,

00:01:54.864 --> 00:01:59.327
Blue Origin, Jeff Bezos'
space company, United Launch

00:01:59.452 --> 00:02:02.288
Alliance, which is the
joint venture of Boeing and

00:02:02.413 --> 00:02:05.708
Lockheed Martin, and
Arianespace, which is the

00:02:05.834 --> 00:02:07.961
European rocket maker.

00:02:08.128 --> 00:02:11.798
The question now is, can the
e-commerce giant catch up to

00:02:11.923 --> 00:02:12.715
SpaceX?

00:02:13.466 --> 00:02:17.011
It is true that SpaceX has a
first mover advantage and

00:02:17.137 --> 00:02:21.015
has more satellites
launched than Amazon.

00:02:21.266 --> 00:02:25.353
The advantage that Amazon
has is they have the chance

00:02:25.478 --> 00:02:28.314
to learn from Starlink's
mistakes.

00:02:35.864 --> 00:02:38.825
Amazon's Project Kuiper is
named for the outermost

00:02:38.950 --> 00:02:40.451
asteroid belt in our solar
system.

00:02:41.369 --> 00:02:44.164
Amazon wants to launch its
satellites into low Earth

00:02:44.289 --> 00:02:49.377
orbit at about 360 to 390
miles above the Earth's

00:02:49.502 --> 00:02:53.256
surface, an orbit similar
to where SpaceX has its

00:02:53.381 --> 00:02:56.926
satellites. The basis for
all satellite internet

00:02:57.051 --> 00:03:00.680
networks involves three
components: a satellite

00:03:00.805 --> 00:03:04.893
dish, a ground station and
the satellites themselves.

00:03:05.143 --> 00:03:08.062
The goal is to provide
connectivity to remote parts

00:03:08.188 --> 00:03:11.149
of the planet, not served
by traditional internet

00:03:11.274 --> 00:03:13.735
options like cable and DSL.

00:03:13.860 --> 00:03:17.030
Amazon has been fairly
tight-lipped about the

00:03:17.155 --> 00:03:19.240
design of its satellites.

00:03:19.365 --> 00:03:22.952
However, they have stated
that these satellites will

00:03:23.077 --> 00:03:27.207
be much bigger than those
being launched by SpaceX.

00:03:28.166 --> 00:03:31.336
The difference or the
reason for this is Amazon

00:03:31.461 --> 00:03:33.588
needs spacecraft that are
going to be much more

00:03:33.713 --> 00:03:36.382
powerful because they're
going to have fewer

00:03:36.507 --> 00:03:39.427
spacecraft in the orbit
that they chose.

00:03:39.552 --> 00:03:43.348
And in order to deliver a
powerful signal to the

00:03:43.473 --> 00:03:47.018
antenna or user terminal in
the ground, the spacecraft

00:03:47.143 --> 00:03:48.811
has to do all the work.

00:03:49.187 --> 00:03:53.858
In April, CNBC's Andrew Ross
Sorkin got an exclusive,

00:03:53.983 --> 00:03:57.528
first tour inside Amazon's
Project Kuiper Facility in

00:03:57.654 --> 00:03:59.239
Redmond, Washington.

00:03:59.822 --> 00:04:03.451
The two buildings that are
here in Redmond, one is

00:04:03.576 --> 00:04:06.537
focused mostly on R & D, to
create Kuiper Space

00:04:06.663 --> 00:04:10.375
satellites, and base
stations and customer

00:04:10.500 --> 00:04:12.794
terminals, the satellite
dishes you'll put in your

00:04:12.919 --> 00:04:17.173
house. And then the room
we're in here is going to be

00:04:17.298 --> 00:04:20.760
the manufacturing line, at
least the first of the

00:04:20.885 --> 00:04:24.055
manufacturing lines for the
actual satellites

00:04:24.180 --> 00:04:25.390
themselves.

00:04:25.640 --> 00:04:28.893
An estimated 37% of the
world's population has still

00:04:29.018 --> 00:04:32.522
never used the internet,
with 96% of those

00:04:32.647 --> 00:04:35.400
individuals living in
developing countries.

00:04:35.525 --> 00:04:39.487
Many people, we grew up with
connectivity over modems and

00:04:39.612 --> 00:04:42.532
things like that that all
ran over copper, that was

00:04:42.657 --> 00:04:44.575
subsidized here in the U.S.

00:04:44.701 --> 00:04:48.413
in the forties and fifties
and sixties, and nobody's

00:04:48.538 --> 00:04:52.041
upgrading that in rural
America and nobody is

00:04:52.166 --> 00:04:54.377
upgrading that in
sub-Saharan Africa.

00:04:54.502 --> 00:04:57.046
And so what's happening is
people are leapfrogging to

00:04:57.171 --> 00:05:00.591
mobile, but it's leaving
this gap in the home.

00:05:00.717 --> 00:05:04.637
And Kuiper, if successful,
is about bridging that.

00:05:05.179 --> 00:05:10.518
The satellite communications
market is one that's valued

00:05:10.643 --> 00:05:13.896
at a few tens of billions
of dollars.

00:05:14.022 --> 00:05:18.651
It's not uncommon for tech
giants to support digital

00:05:18.776 --> 00:05:22.905
infrastructure, satellite
internet being one of those.

00:05:23.031 --> 00:05:27.577
Facebook invested in a
satellite project, or

00:05:27.702 --> 00:05:30.997
purchase capacity on a
satellite that,

00:05:31.122 --> 00:05:34.667
unfortunately it blew up,
but they spent considerable

00:05:34.792 --> 00:05:40.089
money looking at space as a
way to extend their own

00:05:40.214 --> 00:05:42.925
service, to bring more
people onto Facebook.

00:05:43.051 --> 00:05:47.096
And Google has financed
Fiber as well to bring more

00:05:47.221 --> 00:05:50.016
people online for their
service. What Amazon is

00:05:50.141 --> 00:05:54.937
doing is part of a
continued trend of tech

00:05:55.063 --> 00:05:58.316
giants building out
infrastructure that can

00:05:58.441 --> 00:06:03.863
ultimately support their
own core services, which for

00:06:03.988 --> 00:06:06.157
Amazon would be e-commerce.

00:06:10.036 --> 00:06:13.498
Although the payout may be
huge, building out such

00:06:13.623 --> 00:06:17.877
large satellite internet
networks is very expensive.

00:06:18.211 --> 00:06:21.339
Both SpaceX and Amazon have
said that they expect to

00:06:21.464 --> 00:06:24.801
spend over $10 billion to
develop their respective

00:06:24.926 --> 00:06:26.511
constellations.

00:06:26.844 --> 00:06:30.264
For almost any other
company. That would be a

00:06:30.390 --> 00:06:33.142
mammoth hurdle to get over.

00:06:33.351 --> 00:06:38.940
But Amazon spent roughly
$60 billion on capital

00:06:39.065 --> 00:06:41.943
expenditures in 2021 alone.

00:06:42.068 --> 00:06:46.114
So compared to the average
company, this is a cost

00:06:46.239 --> 00:06:50.701
Amazon could actually get
its arms around and it

00:06:50.827 --> 00:06:55.998
doesn't threaten capsizing
the program, provided they

00:06:56.124 --> 00:06:59.961
can keep it within that
cost estimate.

00:07:00.086 --> 00:07:03.631
Trying to build this kind of
global satellite internet

00:07:03.756 --> 00:07:09.762
network is much less of an
existential threat to the

00:07:09.887 --> 00:07:13.933
existence of the business
of Amazon versus the

00:07:14.058 --> 00:07:19.939
business of SpaceX, because
Amazon does so many other

00:07:20.064 --> 00:07:25.486
things. Whereas for SpaceX,
this is their largest other

00:07:25.611 --> 00:07:28.406
project alongside it's
building its Starship

00:07:28.531 --> 00:07:29.365
rocket.

00:07:29.949 --> 00:07:33.828
SpaceX charges $110 per
month for its baseline

00:07:33.953 --> 00:07:38.624
internet service, and
customers have to pay $599

00:07:38.749 --> 00:07:40.251
for a Starlink satellite
dish.

00:07:41.085 --> 00:07:43.921
Amazon has not yet revealed
how much its internet

00:07:44.046 --> 00:07:47.175
service will cost, but
Kuiper will follow a similar

00:07:47.300 --> 00:07:49.969
business model, with
customers paying for a

00:07:50.094 --> 00:07:52.722
subscription and a
satellite dish.

00:07:53.055 --> 00:07:58.060
Amazon recently disclosed
that its Kuiper satellite

00:07:58.186 --> 00:08:02.315
antennas, which consumers
would use here on the

00:08:02.440 --> 00:08:08.029
ground, are under a cost of
$500 apiece.

00:08:08.196 --> 00:08:13.576
By comparison, SpaceX,
which has been selling its

00:08:13.701 --> 00:08:19.624
Starlink hardware for about
$600 apiece, has disclosed

00:08:19.749 --> 00:08:24.921
last year that it cost the
company around $1,300

00:08:25.046 --> 00:08:28.799
dollars for each of its
antenna, meaning that it's

00:08:28.925 --> 00:08:34.472
subsidizing roughly half of
the price of those Starlink

00:08:34.597 --> 00:08:38.768
hardware to deliver the
service to customers on the

00:08:38.893 --> 00:08:39.769
ground.

00:08:40.311 --> 00:08:44.190
Several constellations,
Starlink not being the

00:08:44.315 --> 00:08:49.654
exception, have tried first
with the spacecraft in hopes

00:08:49.779 --> 00:08:52.198
that they would solve the
antenna later.

00:08:52.323 --> 00:08:55.326
And what ultimately results
from this, or what has

00:08:55.451 --> 00:08:59.705
resulted in the past, is
the antenna has remained too

00:08:59.830 --> 00:09:04.710
costly for the service to
ever be truly widespread.

00:09:04.835 --> 00:09:08.881
Another area where Amazon
has a head start is its AWS

00:09:09.006 --> 00:09:09.840
ground station
infrastructure.

00:09:11.008 --> 00:09:15.513
The biggest advantage for
Amazon is that they can link

00:09:15.638 --> 00:09:17.932
their constellation
directly back to the

00:09:18.057 --> 00:09:20.101
internet using their own
infrastructure.

00:09:21.352 --> 00:09:25.606
So if you have an Amazon
Kuiper terminal on your

00:09:25.731 --> 00:09:29.026
house, it's going to link
to the satellite and then it

00:09:29.151 --> 00:09:32.321
needs to link back to the
internet somewhere else.

00:09:32.446 --> 00:09:37.827
For Starlink, they have to
set up those gateways back

00:09:37.952 --> 00:09:43.874
to the internet using other
service providers, either

00:09:44.000 --> 00:09:48.087
other cloud network service
providers, or find ways to

00:09:48.212 --> 00:09:51.173
stitch Starlink back into
the internet.

00:09:51.757 --> 00:09:54.427
One of those service
providers is Google.

00:09:54.552 --> 00:09:58.681
In 2021, SpaceX signed a
deal that allows the company

00:09:58.806 --> 00:10:01.517
to install ground stations
at Google data centers

00:10:01.642 --> 00:10:02.852
around the world.

00:10:04.020 --> 00:10:06.147
AWS has a network backbone.

00:10:06.272 --> 00:10:07.898
It's close to the internet.

00:10:08.024 --> 00:10:09.483
Lots of people are hosting
on it.

00:10:09.609 --> 00:10:13.613
So if we can interconnect
our base stations to that

00:10:14.071 --> 00:10:18.200
AWS network, we're
automatically advantaged in

00:10:18.326 --> 00:10:21.495
that case for customers to
get better bandwidth.

00:10:21.621 --> 00:10:24.790
Billing is something that
we do a lot and we're riding

00:10:24.915 --> 00:10:27.668
on the shoulders of all
that building

00:10:27.793 --> 00:10:31.297
infrastructure.
Manufacturing is a big, hard

00:10:31.422 --> 00:10:35.718
component of this. We have
to build over time tens of

00:10:35.843 --> 00:10:39.263
millions of customer
terminals, but in the device

00:10:39.388 --> 00:10:42.391
business, we build tens of
millions of things every

00:10:42.516 --> 00:10:47.063
month. And so it gives us
the opportunity to use the

00:10:47.188 --> 00:10:50.441
expertise of that design
and manufacturing and bring

00:10:50.566 --> 00:10:51.275
it over here.

00:10:54.111 --> 00:10:57.698
Even with its deep pockets
and extensive experience

00:10:57.823 --> 00:11:01.327
producing consumer
electronic devices at scale,

00:11:01.452 --> 00:11:05.122
Amazon still faces a number
of challenges in making its

00:11:05.247 --> 00:11:06.540
satellite internet a
reality.

00:11:07.708 --> 00:11:12.296
One of the biggest hurdles
to getting such a network up

00:11:12.421 --> 00:11:14.382
and running, especially
when we were talking about

00:11:14.507 --> 00:11:18.427
this number of satellites,
is getting them into orbit.

00:11:18.552 --> 00:11:21.931
To get its constellation
into space, Amazon has

00:11:22.056 --> 00:11:25.685
contracts for 38 launches
with United Launch Alliance,

00:11:25.810 --> 00:11:29.772
18 launches with
Arianespace and 12 launches

00:11:29.897 --> 00:11:33.401
with Blue Origin, with an
option for as many as 15

00:11:33.526 --> 00:11:35.986
additional launches with
the private venture that's

00:11:36.112 --> 00:11:39.115
owned by Amazon founder
Jeff Bezos.

00:11:39.490 --> 00:11:43.035
Each of the rockets that
Amazon is relying on to

00:11:43.160 --> 00:11:47.540
launch Kuiper are at least
two years behind schedule.

00:11:47.665 --> 00:11:51.961
And on top of that, the
challenge most new rockets

00:11:52.086 --> 00:11:56.424
face is not just getting to
the launch pad for the first

00:11:56.549 --> 00:12:01.261
time, but scaling to launch
to three, five, ten times a