Latest / Connecting the Dots / Nvidia's Fiber Optic Push, US Manufacturing Boost, and Anthropic's Mega Cloud Deal

Transcript

Nvidia's Fiber Optic Push, US Manufacturing Boost, and Anthropic's Mega Cloud Deal · Connecting the Dots

• 0:00
  Today's tech news moves fast, but understanding
• 0:03
  it shouldn't be hard. That's why Snarful Solutions
• 0:06
  Group brings you Connecting the Dots, where we
• 0:09
  break down what's happening in tech and why it
• 0:11
  matters to you. Now, let's dive in with your
• 0:14
  hosts. Imagine, like, writing a check for $200
• 0:18
  billion today. Just to guarantee that your company
• 0:20
  has a place to plug in its computers five years
• 0:22
  from now. Yeah, it sounds completely crazy. Right,
• 0:25
  but that is not a hypothetical scenario or some,
• 0:28
  you know, distant sci -fi concept. That is the
• 0:31
  sheer physical reality of the artificial intelligence
• 0:35
  arms race happening right beneath our feet. I'm
• 0:38
  Alex. And I'm Morgan. And welcome to Connecting
• 0:40
  the Dots, coming to you right here from the Snarfle
• 0:42
  Solutions Group in Sacramento, California. Yep,
• 0:45
  glad to be here. Today is Wednesday, May 6. 2026.
• 0:50
  And, you know, our goal here every day is to
• 0:52
  take the most complex tech news and figure out
• 0:55
  how it actually impacts your everyday life. Because
• 0:57
  it really does impact you. Exactly. We want to
• 1:00
  connect those dots so you can see the bigger
• 1:01
  picture without getting like completely lost
• 1:04
  in the weeds. Which is incredibly important today,
• 1:06
  I mean, because we are tracking a shift from
• 1:09
  abstract algorithms floating in the cloud to
• 1:11
  heavy, expensive physical infrastructure. Yeah.
• 1:15
  The stuff you can actually touch. Yeah, but first
• 1:17
  let's get you cut up on your morning for those
• 1:19
  of you traveling today Definitely keep an eye
• 1:22
  on the skies. Yeah, it's a bit rough out there.
• 1:24
  It really is there are currently 218 active us
• 1:27
  weather alerts with 93 of those sitting over
• 1:30
  land Wow, so you will definitely want to check
• 1:33
  weather .gov if you are heading out of state
• 1:35
  But if you are staying local with us in Sacramento
• 1:38
  today, you are looking at a gorgeous one Oh,
• 1:41
  absolutely. Perfect. Mostly sunny a perfect 77
• 1:44
  degrees Fahrenheit. We're just a you know a light
• 1:46
  three -model per -hour breeze. It is the kind
• 1:49
  of day that makes you want to skip the office
• 1:50
  entirely, to be honest. Don't tempt me. And honestly,
• 1:53
  the financial markets were looking just as sunny
• 1:55
  yesterday. They really were. Tuesday brought
• 1:58
  us some, like, really solid positive numbers
• 2:00
  across the board. The Dow jumped 261 .13 points.
• 2:04
  Which is a good bump. Yeah. That is a 0 .53 %
• 2:07
  increase, closing at 49 ,298 .25. The S &P 500
• 2:12
  followed suit. gaining 25 .60 points to close
• 2:15
  out at 7 ,259 .22. Nice to see that steady green.
• 2:19
  For sure. And for those of you tracking crypto,
• 2:22
  Bitcoin is just continuing its climb. It shot
• 2:24
  up 11 .03 .52 yesterday, closing at 80 ,927 .05.
• 2:30
  The momentum in the markets right now is undeniable.
• 2:33
  I mean, even if the sports world is a little
• 2:35
  less forgiving. Yeah. Rough transition, but accurate.
• 2:40
  It was a pretty brutal Tuesday. if you were a
• 2:42
  Northern California baseball fan. Yeah, not our
• 2:45
  best showing. The SF Giants dropped their game
• 2:47
  to the Padres 5 -10. Walker Bueller ended up
• 2:50
  taking the win for San Diego. Good for him, bad
• 2:53
  for us. Right. And the Oakland A's had a really
• 2:55
  tough night too, losing 1 -9 to the Phillies,
• 2:59
  with Christopher Sanchez taking that win. Ouch.
• 3:01
  But hey, the Sacramento Kings had the night off,
• 3:03
  so at least our local basketball team couldn't
• 3:06
  break our hearts yesterday. I will take the small
• 3:08
  victories where we can get them. Seriously. And,
• 3:10
  speaking of starting your morning off with a
• 3:12
  victory, let's pivot to something that will actually
• 3:14
  give you the energy to tackle all this heavy
• 3:17
  news. Today's episode is brought to you by Strong
• 3:19
  Coffee, the premium instant coffee that tastes
• 3:21
  like a traditional brew but is packed with 15
• 3:24
  grams of protein, MCTs and adaptogens like ashwagandha.
• 3:29
  Get clean, sustained energy without the jitters.
• 3:32
  Use promo code SNARFL for 20 % off at strongcoffeecompany
• 3:36
  .com. Alright, let's get into the news because
• 3:38
  today we are looking at a trio of stories that
• 3:41
  paint a very continuous Causal picture of where
• 3:45
  the global economy is heading. Yeah, it's all
• 3:47
  connected It really is we are watching the artificial
• 3:50
  intelligence boom manifest in the real physical
• 3:52
  world. So let's unpack this Our first story looks
• 3:55
  at a massive financial move by the undisputed
• 3:58
  heavyweight champion of the AI era, which is
• 4:00
  Nvidia. Obviously. Yeah. They just announced
• 4:02
  a 500 million dollar equity linked investment
• 4:05
  in Corning. And, you know, if you have a smartphone
• 4:07
  in your pocket or Pyrex of baking dishes in your
• 4:10
  kitchen, you know Corning. Right. They're a hundred
• 4:12
  seventy five year old American glass company.
• 4:14
  Exactly. And the financial structure of this
• 4:17
  Nvidia deal is fascinating because they aren't
• 4:19
  just like buying stock on the open market. They're
• 4:22
  using something called warrants. And the market
• 4:25
  reaction alone tells you how significant this
• 4:27
  is. I mean, before the opening bell even rang,
• 4:30
  Corning shares soared 20 % in pre -market trading.
• 4:33
  Wow. Yeah, hitting $185 .59. NVIDIA popped over
• 4:37
  2 % as well. But you are right to focus on the
• 4:40
  warrants because it reveals NVIDIA's actual strategy
• 4:43
  here. So I want to make sure we break down the
• 4:45
  mechanics of this without getting bogged down
• 4:47
  in, like, Wall Street jargon. Nvidia received
• 4:50
  two different types of warrants, right? The first
• 4:53
  is a traditional warrant to buy up to 15 million
• 4:57
  shares of Corning at $180 a share. That makes
• 5:00
  sense to me. It's essentially an option to buy
• 5:02
  later at a set price. Yep, pretty standard. But
• 5:05
  the second one is weird. It is a pre -funded
• 5:07
  warrant to buy 3 million shares at a nominal
• 5:10
  price of literally a fraction of a penny. A tiny
• 5:13
  fraction. We're talking point zero zero zero
• 5:15
  zero one dollars per share. Why go through the
• 5:18
  headache of a pre -funded fractional penny warrant
• 5:20
  instead of just handing them a check or you know
• 5:22
  just buying shares? Well it is a highly specialized
• 5:25
  financial tool used to deploy capital instantly.
• 5:28
  Yeah. If Nvidia just went to the open market
• 5:31
  and sought three million shares, it wouldn't
• 5:33
  actually give Corning any new money to build
• 5:35
  factories, right? It just buys out existing shareholders.
• 5:38
  Oh, right, of course. On the flip side, if Corning
• 5:41
  simply issued three million brand new shares
• 5:43
  to Nvidia tomorrow, it would instantly dilute
• 5:46
  the value for every other investor. Which people
• 5:48
  hate. Exactly. And it would trigger a massive
• 5:51
  wave of regulatory filings and immediate tax
• 5:53
  burdens. So the pre -funded warrant is what,
• 5:56
  like a loophole? Think of it more like an advance
• 5:58
  payment on a contract. By pricing the warrant
• 6:01
  at a fraction of a penny, Nvidia is handing Corning
• 6:04
  nearly half a billion dollars in upfront cash
• 6:07
  today. Wow. But they're legally deferring the
• 6:10
  actual issuance of those three million shares
• 6:12
  until later It allows corning to take that cash
• 6:15
  and immediately start pouring concrete and buying
• 6:18
  machinery tomorrow morning Okay, that makes sense
• 6:20
  doing without blowing up their current stock
• 6:22
  structure precise. Okay, that clarifies the how
• 6:25
  but it brings up a much bigger why When I was
• 6:28
  reading through our source material I was just
• 6:30
  trying to figure out why a company that designs
• 6:32
  the most advanced silicon chips on the planet
• 6:35
  is suddenly funneling a half billion dollars
• 6:38
  into an old -school glass manufacturer. It does
• 6:41
  seem disconnected at first. Right. In the early
• 6:44
  days of this tech boom, people used the analogy
• 6:47
  that Nvidia was getting rich by selling the pickaxes
• 6:49
  to the gold miners. But this is entirely different.
• 6:53
  How so? It's more like, in the AI gold rush,
• 6:56
  instead of just selling the pickaxes, Nvidia
• 6:58
  is now buying up the factories that make the
• 7:00
  metal for the pickaxes. Oh, that is a perfect
• 7:02
  parallel, yeah. They are looking at the macro
• 7:04
  environment and realizing they have a massive
• 7:07
  bottleneck on their hands. Supply chain issues.
• 7:09
  Huge ones. The hyperscalers, companies like Microsoft,
• 7:12
  Alphabet, Meta, they are projected to spend an
• 7:15
  astonishing $700 billion this year alone on data
• 7:18
  centers. That number is just hard to comprehend.
• 7:21
  It really is. And NVIDIA's GPUs are the brains
• 7:23
  of those data centers. But, you know, a brain
• 7:25
  is useless if it cannot communicate with the
• 7:27
  rest of the body. Right. Nvidia realizes that
• 7:30
  if they don't actively lock down the raw materials
• 7:32
  to build the nervous system that connects their
• 7:34
  chips together, the entire $700 billion boom
• 7:38
  will just grind to a halt. Moving on to our next
• 7:40
  story, let's trace where that $500 million cash
• 7:43
  injection is actually going. Because it isn't
• 7:45
  just sitting in an investment account yielding
• 7:47
  interest. No, it's being put to work immediately.
• 7:50
  Exactly. It is sparking a massive physical build
• 7:53
  out right now in the United States. This is where
• 7:56
  we see the technology fundamentally change the
• 7:58
  real world landscape. Yeah, the partnership between
• 8:01
  Nvidia and Corning is funding three brand new
• 8:04
  advanced manufacturing plants going up in North
• 8:07
  Carolina and Texas, which is huge. We are looking
• 8:10
  at the creation of over 3000 high paying domestic
• 8:13
  US jobs. Corning is officially targeting a 10x
• 8:17
  increase in its US -based optical manufacturing
• 8:19
  capacity. Ten times. Ten times. Yeah. And they
• 8:22
  are expanding their domestic fiber production
• 8:24
  by more than 50%. Plus, Meta recently threw in
• 8:27
  $6 billion to help Corning build a massive plant
• 8:31
  in Hickory, North Carolina. It is a full -scale
• 8:34
  industrial mobilization at this point. But wait,
• 8:37
  let me push back on that a little bit. because
• 8:40
  we have laid millions of miles of copper wiring
• 8:44
  over the last 50 years to build the global internet.
• 8:47
  Right, copper is everywhere. The entire world
• 8:49
  already runs on copper infrastructure. Why are
• 8:53
  we suddenly ripping the guts out of data centers
• 8:55
  just because AI chatbots are getting smarter?
• 8:57
  That's a fair point. I mean, why can't we just
• 8:59
  lay thicker copper cables? It feels like an unnecessary,
• 9:02
  incredibly expensive sci -fi upgrade or something.
• 9:05
  It is a great question, but it really just comes
• 9:07
  down to the unforgiving laws of physics. Traditional
• 9:10
  data centers use copper wire to transmit data
• 9:14
  via electrical pulses. Okay. And copper is fantastic
• 9:17
  for normal web traffic, you know, loading a web
• 9:19
  page, streaming a video, sending an email. But
• 9:22
  copper has inherent physical limits regarding
• 9:25
  electrical resistance. So cop's out. Exactly.
• 9:27
  When you try to push the sheer volume of data
• 9:29
  required for artificial intelligence through
• 9:32
  copper, it generates immense friction. Friction
• 9:35
  meaning heat. Massive amounts of heat and significant
• 9:38
  latency too. In modern AI workloads, you have
• 9:43
  thousands of NVIDIA GPUs that have to perform
• 9:46
  complex math in parallel. Right. They have to
• 9:49
  talk to each other almost instantly to train
• 9:51
  a single model. If you use copper, the electrical
• 9:54
  signal degrades over distance and the cables
• 9:56
  literally run too hot to function at that density.
• 10:00
  So pushing AI data through traditional copper
• 10:03
  wire. It's essentially like trying to force a
• 10:05
  fire hose of water through a plastic straw. Yeah,
• 10:08
  that's exactly it. The friction generates massive
• 10:10
  heat, the pressure builds up, and everything
• 10:12
  just slows down. The physics just fail. So the
• 10:15
  industry is being forced to transition to Corning's
• 10:18
  optical glass fibers. The technology is called
• 10:21
  co -packaged optics. Co -packaged optics, got
• 10:23
  it. Instead of using electrical pulses, you are
• 10:25
  translating the data into lasers. If copper is
• 10:28
  the fire hose through a straw, optical glass
• 10:30
  is like flashing a laser pointer across an empty
• 10:32
  room. Wow, okay. There's zero physical resistance,
• 10:36
  zero heat generated by the transmission, and
• 10:38
  it arrives instantly. Which gives a lot of context
• 10:40
  to that quote from Nvidia CEO Jensen Wong. Oh,
• 10:43
  right, the speed of light comment. Yeah. He specifically
• 10:46
  said, they are building an infrastructure where
• 10:49
  intelligence moves at the speed of light while
• 10:51
  advancing the tradition of made in America, because
• 10:54
  they are literally using light to move the data.
• 10:56
  And that Made in America piece is the crucial
• 10:59
  takeaway for, you know, the working professionals
• 11:01
  and families listening right now. It really is.
• 11:04
  For decades, the tech industry felt completely
• 11:06
  invisible. It was just code written in the Silicon
• 11:09
  Valley office, hosted on a server somewhere out
• 11:12
  of sight. Just floating in the cloud. Right.
• 11:14
  But AI is demanding physical space. We are talking
• 11:18
  about thousands of construction workers pouring
• 11:20
  concrete. We are talking about specialized glass
• 11:23
  manufacturing, complex cooling systems. Real
• 11:26
  jobs. Exactly. Hiring thousands of skilled tradespeople,
• 11:30
  engineers and factory workers in places like
• 11:32
  Texas and North Carolina. The AI boom is actively
• 11:36
  revitalizing physical domestic supply chains.
• 11:39
  It grounds the whole concept. When you ask an
• 11:41
  AI a question on your phone, the answer isn't
• 11:44
  just magic. Not at all. It is physically traveling
• 11:46
  on a beam of light through American -made glass
• 11:49
  manufactured in North Carolina, powered by a
• 11:52
  chip designed by Nvidia. But, you know, all of
• 11:54
  that physical infrastructure, the land, the concrete,
• 11:58
  the glass fiber, the power grid connections,
• 12:01
  it takes years to build. Which is a problem when
• 12:04
  you need it yesterday. Right. And that physical
• 12:06
  delay brings us to the software side of the equation.
• 12:09
  Let's look closer at the companies desperately
• 12:11
  trying to rent this physical infrastructure before
• 12:14
  the concrete is even poured. They're panicking
• 12:16
  a bit. Oh, big time. Our third story today focuses
• 12:20
  on Anthropic. They are the AI startup behind
• 12:24
  the Claude models and the financial commitments
• 12:26
  they're making right now are just mind -bending.
• 12:29
  There really are. Anthropic has reportedly agreed
• 12:31
  to pay Google a staggering $200 billion over
• 12:35
  five years just for access to cloud servers and
• 12:38
  chips. And they're committing to infrastructure
• 12:40
  that doesn't even exist yet. Right. They're specifically
• 12:42
  trying to lock down Google's TPU capacity that
• 12:46
  isn't coming online until 2027. Unbelievable.
• 12:49
  And that is just one deal. They're also spending
• 12:52
  $100 billion on Amazon's AWS cloud in a multi
• 12:57
  -cloud strategy. We should probably define what
• 12:59
  they're actually buying there. You mentioned
• 13:01
  TPUs, which stands for Tensor Processing Unit.
• 13:04
  How does that differ from the NVIDIA GPUs we
• 13:06
  were just talking about? Well, a standard computer
• 13:08
  processor, like a CPU, is like a fast sports
• 13:11
  car making one delivery at a time. It does things
• 13:14
  sequentially. A GPU, which Nvidia dominates,
• 13:18
  is like a massive fleet of delivery trucks. It
• 13:20
  handles lots of tasks at once. But Google custom
• 13:24
  build the TPU specifically for neural network
• 13:27
  math. So it's even more specialized. Yeah, it
• 13:29
  is like a high speed freight train track built
• 13:32
  exclusively to deliver AI data in massive parallel
• 13:35
  batches. Anthropic knows that if they want their
• 13:38
  Claude model to survive, they need an exclusive
• 13:40
  diet of that massive parallel processing. But
• 13:42
  the financial risk here seems terrifying. I mean,
• 13:45
  to put this in perspective, Anthropic's annualized
• 13:47
  revenue run rate hit $30 billion in 2026. Which
• 13:51
  is incredible. It is an incredibly successful
• 13:53
  business. Yeah. But they are turning around and
• 13:56
  taking on hundreds of billions in long -term
• 13:58
  debt just to secure their computing power. Why
• 14:01
  would a company with that much cash flow mortgage
• 14:03
  their entire future? just to reserve server space.
• 14:07
  Because they are looking at the physical supply
• 14:09
  chain bottlenecks we just discussed. They know
• 14:12
  that laying glass fiber, building substations,
• 14:15
  and securing land permits takes five to ten years.
• 14:18
  Oh, so it's a race against the clock. It reminds
• 14:21
  me of, like, Booking every hotel room in the
• 14:23
  city five years in advance because you know a
• 14:26
  massive festival is coming That's a great way
• 14:28
  to look at it. They don't want to risk their
• 14:29
  engineers and their data, you know sleeping on
• 14:32
  the street They're willing to pay an astronomical
• 14:34
  premium just to lock in the real estate now and
• 14:37
  that real estate hoarding is creating a massive
• 14:39
  backlog Right now, there is an estimated two
• 14:42
  trillion dollar revenue backlog across Amazon,
• 14:45
  Google, Microsoft and Oracle. Two trillion. Yeah.
• 14:48
  These cloud providers have guaranteed money lined
• 14:50
  up for years, entirely driven by mega deals from
• 14:54
  giants like Anthropic. This is where I want to
• 14:56
  talk about the everyday listener. If the tech
• 14:58
  titans are hoarding all the electricity, the
• 15:01
  networking bandwidth and the data center space
• 15:03
  years in advance. Yeah. Where does that leave
• 15:07
  everyone else? In a tough spot. Like if I'm a
• 15:10
  mid -sized enterprise trying to upgrade my own
• 15:12
  internal cloud systems or a startup trying to
• 15:15
  launch a new app, what happens to me? You get
• 15:17
  squeezed out. It is the harsh reality of constrained
• 15:21
  supply. When the big cloud providers are looking
• 15:24
  at a $2 trillion backlog from the Titans, they
• 15:27
  have zero incentive to cater to small buyers.
• 15:30
  So smaller companies just can't buy what they
• 15:32
  need. They will face significantly longer wait
• 15:34
  times to get the cloud space they request. They
• 15:37
  are going to run into physical RAM shortages.
• 15:40
  And crucially, smaller businesses will have absolutely
• 15:43
  no pricing leverage. Hardware vendors and cloud
• 15:46
  operators are prioritizing the contracts that
• 15:48
  lock in the biggest, longest demand streams.
• 15:51
  So if your startup only needs a small amount
• 15:53
  of compute, you get pushed to the back of the
• 15:55
  line and you will pay a massive premium for whatever
• 15:57
  scraps of capacity are left over. That paints
• 15:59
  a really sobering picture of the tech ecosystem.
• 16:03
  I mean, we are seeing a complete restructuring
• 16:05
  of who gets access to the tools of the future.
• 16:08
  driven entirely by who has the deepest pockets
• 16:10
  today to monopolize that physical supply chain.
• 16:13
  Before we wrap up, a quick thanks to Old Glory,
• 16:16
  an iconic music and sports fan merch store. Perfect
• 16:19
  for the techie in your life. Use promo code SNARFUL
• 16:23
  for 15 % off at oldglory .com. So if we step
• 16:27
  back and look at the whole board today, the narrative
• 16:29
  is incredibly clear. Very clear. The AI boom
• 16:32
  has left the theoretical realm of software and
• 16:34
  has become a heavy physical reality. Yeah, we
• 16:38
  started with Nvidia deploying a 500 million dollar
• 16:40
  financial warrant to secure their own supply
• 16:42
  chain. They are looking past the chips and securing
• 16:45
  the optical glass required to connect them. Right,
• 16:47
  looking at the big picture. And that capital
• 16:49
  is flowing directly into the resurgence of U
• 16:51
  .S. manufacturing, with Corning building massive
• 16:54
  plants in North Carolina and Texas and creating
• 16:57
  thousands of jobs just to replace old copper
• 16:59
  wires with high -speed glass. And because that
• 17:02
  physical infrastructure requires, you know, pouring
• 17:05
  concrete and laying fiber, it moves slowly. Super
• 17:08
  slowly. And that physical delay has triggered
• 17:10
  a panic on the software side, forcing companies
• 17:13
  like Anthropic to drop hundreds of billions of
• 17:15
  dollars to hoard cloud capacity years before
• 17:18
  it even exists. It really changes how you view
• 17:21
  the devices we use every day, which leaves us
• 17:24
  with a final thought to mull over today. Let's
• 17:26
  hear it. As the tech titans drop billions to
• 17:29
  buy up the glass fiber highways of the future
• 17:31
  before they're even built, how will everyday
• 17:34
  startups and developers find room to drive their
• 17:37
  own innovations? That will be the defining economic
• 17:39
  battle of the next decade. We're here to help
• 17:41
  it snarfle, so reach out on our website if you
• 17:43
  have any more questions. That's a wrap on today's
• 17:49
  top stories. Remember, tech doesn't have to be
• 17:52
  complicated, and we're here to help connect the
• 17:54
  dots. If you've got questions or want to learn
• 17:57
  more, visit us at snarful .com. Thanks for listening,
• 18:01
  and we'll catch you next time.