Latest / Connecting the Dots / OpenAI’s Trillion-Dollar Power Play and Tension with Microsoft

Transcript

OpenAI’s Trillion-Dollar Power Play and Tension with Microsoft · 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. Welcome back to Connecting the Dots. I'm
• 0:17
  Alex. And I'm Morgan. Thanks for joining us.
• 0:19
  It is Monday, October 27th, 2025 here in Sacramento.
• 0:23
  And the weather's looking pretty good today,
• 0:25
  becoming sunny, expecting a high around 70 degrees.
• 0:28
  Nice. Nationally, though, a bit more action.
• 0:31
  Snow's finally ending out in the Cascades and
• 0:33
  Rockies. Yep. And Southern California is getting
• 0:35
  its first real Santa Ana wind event of the season.
• 0:39
  Plus, unfortunately, Hurricane Melissa is really
• 0:42
  hitting Jamaica hard with destructive winds and
• 0:45
  flooding. Yeah, sending thoughts out there. OK,
• 0:47
  let's look at the markets quickly. Things were.
• 0:50
  Well, generally positive to start the week. The
• 0:53
  Dow Jones saw a little bump up about 37 points.
• 0:56
  So plus point zero eight percent in the S &P
• 1:00
  500. Pretty similar story. Finished up about
• 1:02
  point zero six percent gaining just over four
• 1:04
  points. But you know, if you want a sign of where
• 1:06
  the big money confidence is, look at Bitcoin.
• 1:08
  Another strong day. Right. It jumped over six
• 1:11
  hundred and twenty dollars. It's more than half
• 1:13
  a percent. Closing Monday at one hundred fifteen
• 1:15
  thousand one hundred eighty one dollars and twenty
• 1:17
  eight cents. Still climbing. That definitely
• 1:20
  reflects some of the optimism around the massive
• 1:23
  tech infrastructure build out we're going to
• 1:24
  talk about today. Absolutely. But before we dive
• 1:27
  into that, a quick word about our sponsor. Yeah,
• 1:29
  we want to thank Old Glory. They're an iconic
• 1:31
  store for music and sports fan merchandise. Seriously,
• 1:35
  great stuff. And with the holidays creeping up,
• 1:38
  maybe you've got that techie friend who's impossible
• 1:40
  to shop for or honestly, anyone who loves bands
• 1:44
  or teams. Check them out. Definitely. And if
• 1:47
  you use the promo code SNARFL, that's F -N -A
• 1:50
  -R -F -U -L at checkout, you'll get 15 % off
• 1:53
  your whole order. Great deal. The website is
• 1:55
  OldGlory .com. Again, promo code SNARFL for 15
• 1:59
  % off at OldGlory .com. Okay, so let's get into
• 2:03
  it. We've been digging through our internal Snarfl
• 2:05
  Tech article report, we call it the Morningstar,
• 2:07
  and today it really boils down to two major interconnected
• 2:10
  themes. Right. First, OpenAI's almost unbelievable
• 2:13
  spending spree and deal -making style. And second,
• 2:16
  how that maybe connects to the frame relationship
• 2:20
  between OpenAI and its biggest backer, Microsoft.
• 2:23
  It seems like they're, well, growing apart. Let's
• 2:25
  tackle that spending spree first. Story one.
• 2:28
  Sam Altman's unconventional $1 .5 trillion deal
• 2:33
  spree. Unconventional almost feels like an understatement
• 2:36
  here. It really is. So Sam Altman, OpenAI CEO,
• 2:40
  has orchestrated these massive multi -year deals
• 2:42
  for computing infrastructure. We're talking potentially
• 2:45
  up to $1 .5 trillion combined. With all the big
• 2:49
  players, NVIDIA, Oracle, AMD, Broadcom, you name
• 2:53
  it. But the shocking part... how he did it. Yeah,
• 2:57
  this is what got me. He largely skipped the usual
• 3:00
  process, right? Like, the armies of external
• 3:02
  bankers and lawyers you'd expect for deals this
• 3:04
  size. Exactly. Instead, he relied on a really
• 3:06
  small inner circle. We're talking President Greg
• 3:08
  Brockman, CFO Sarah Fryer, and Peter Hochel,
• 3:11
  who handles their infrastructure financing. Just
• 3:13
  a handful of people. For potentially $1 .5 trillion?
• 3:16
  I mean, that sounds risky, skipping all the usual
• 3:18
  financial due diligence. Well, yeah, from a traditional...
• 3:20
  finance perspective absolutely but according
• 3:22
  sources close to the deals the thinking was completely
• 3:25
  different it was tech first meaning meaning the
• 3:28
  absolute priority was locking down the supply
• 3:30
  of chips and computing power secure that first
• 3:33
  worry about the precise financial details later
• 3:36
  that was a mantra Financial details coming later.
• 3:39
  So it was driven by the sheer need for computing
• 3:41
  power. Immense need. They needed to stimulate
• 3:43
  chip manufacturing, get as many chips developed
• 3:45
  and built as possible, like yesterday. Altman's
• 3:48
  goal is apparently aiming to secure up to one
• 3:51
  gigawatt of power per week. One gigawatt a week.
• 3:54
  OK, put that in perspective for us. Think of
• 3:56
  a small nuclear power plant that's roughly 1
• 3:58
  gigawatt. He wants to add that much compute capacity
• 4:01
  every single week. It's just an insane scale.
• 4:04
  The technical need just overrode everything else.
• 4:06
  And this sort of unconventional deal making,
• 4:09
  it started earlier, didn't it, like with CoreWeave?
• 4:11
  That's a key example. The initial CoreWeave deal
• 4:13
  was for $11 .9 billion in compute. But instead
• 4:17
  of just cash, OpenAI actually received $350 million
• 4:22
  worth of shares in CoreWeave. Okay, that deals
• 4:24
  now expanded to over 22 billion dollars and guess
• 4:27
  what those quarry of shares open AI got they've
• 4:30
  tripled in value since then Wow, so they're not
• 4:32
  just buying compute They're essentially becoming
• 4:34
  investment partners with their suppliers tying
• 4:37
  their fates together. Precisely. It forces everyone
• 4:40
  to make it work It's a fascinating maybe slightly
• 4:42
  terrifying model. Okay, so they're locking down
• 4:45
  supply in these unusual ways, but they're also
• 4:49
  clearly trying not to rely on just one supplier,
• 4:52
  especially the big one, NVIDIA. Which brings
• 4:56
  us to story two, the AMD Alliance betting on
• 5:00
  chip competition. Exactly. NVIDIA became the
• 5:02
  world's most valuable company because they dominate
• 5:05
  high -end AI chips. That gives them enormous
• 5:08
  pricing power, right? Dictating terms. So Altman
• 5:10
  and OpenAI are actively trying to build up competitors.
• 5:14
  diversify. And the biggest move there is this
• 5:16
  major multi -year deal with AMD. This is huge
• 5:19
  for AMD. What are the specifics? The deal involves
• 5:22
  deploying AMD chips that require, get this, up
• 5:25
  to six gigawatts of power, so a massive order.
• 5:28
  Six gigawatts. Okay, but here's the kicker, the
• 5:30
  really interesting part. As part of this deal,
• 5:32
  OpenAI received warrants, basically options,
• 5:35
  to purchase up to 10 % of AMD stock for just
• 5:39
  one cent a share. One cent? So, wait. OpenAI
• 5:43
  gets the right to buy a huge chunk of AMD for
• 5:46
  practically nothing. Yep. So think about what
• 5:48
  that does. Suddenly, OpenAI has a massive financial
• 5:52
  incentive to make sure AMD not only competes
• 5:54
  within video, but actually succeeds and grows
• 5:57
  significantly. They become AMD's biggest cheerleader.
• 5:59
  essentially. Because if AMD stock goes up, OpenAI
• 6:03
  makes a fortune. Exactly. It's a direct play
• 6:05
  to foster competition. And why does this matter
• 6:08
  to everyone else? Well, the hope is it creates
• 6:11
  long -term resilience in the chip supply chain.
• 6:13
  Right. Less dependence on one company, protects
• 6:15
  against geopolitical risks, maybe natural disasters.
• 6:18
  And crucially, if AMD becomes a stronger competitor,
• 6:21
  it should help drive down the cost of these super
• 6:24
  expensive AI chips for the entire ecosystem.
• 6:27
  every AI startup, every university researcher.
• 6:30
  That makes sense. It's also a huge vote of confidence,
• 6:32
  isn't it? All these massive deals, including,
• 6:34
  by the way, a hundred billion dollar investment
• 6:36
  and chip deal with Nvidia, too. Yeah. It pushes
• 6:39
  back against the idea this is all just an AI
• 6:41
  hype bubble. Definitely signals belief in a sustained
• 6:45
  build out. Though you got to love the cynical
• 6:47
  analyst take. Stacey Razgan over at Bernstein,
• 6:50
  he quipped that Altman basically has the power
• 6:53
  to crash the global economy for a decade if OpenAI
• 6:57
  actually fails after making these bets. Well,
• 6:59
  let's hope not. But it highlights the stakes.
• 7:02
  OK, so Altman shaking up finance, shaking up
• 7:04
  the chip market. But the ground might also be
• 7:07
  shifting under his biggest partnership. That's
• 7:09
  story three, Microsoft and OpenAI. growing apart
• 7:13
  on ethics in Office performance. Yeah, this is
• 7:15
  fascinating. Microsoft has poured something like
• 7:17
  $13 billion into OpenAI. It was meant to be this
• 7:21
  super tight, almost exclusive partnership, but
• 7:23
  we're seeing clear cracks forming on two fronts,
• 7:26
  really. Let's start with the practical one, the
• 7:28
  one that might actually affect how your work
• 7:30
  software performs. We're talking Office 365.
• 7:32
  Exactly. Microsoft is ending its exclusive use
• 7:35
  of OpenAI models for powering features in Word,
• 7:37
  Excel, PowerPoint, all those tools. Okay, so
• 7:40
  who are they using instead? or in addition? They're
• 7:42
  bringing in models from a major OpenAI rival,
• 7:45
  Anthropic. Anthropic, okay, why? Performance,
• 7:48
  plain and simple. Microsoft's internal tests
• 7:51
  apparently showed that Anthropic's model, specifically
• 7:54
  one called Claude Sonnet 4, just did a better
• 7:57
  job on certain tasks than OpenAI's models. Like
• 8:00
  what kind of tasks? Things like complex visual
• 8:03
  design generation within PowerPoint, or sophisticated
• 8:06
  automation in Excel spreadsheets. For those specific
• 8:10
  jobs, Claude Sonnet 4 apparently came out ahead.
• 8:13
  Interesting. So they're prioritizing best tool
• 8:15
  for the job over partnership loyalty. Totally.
• 8:18
  And get this for alliance complexity. Microsoft
• 8:20
  isn't buying access to Anthropic's models directly.
• 8:24
  They're buying it through Amazon Web Services
• 8:26
  AWS. Wait, what? So Microsoft is paying its Cloud
• 8:30
  Arch rival Amazon to get access to models from
• 8:33
  its AI partner's rival, Anthropic. You got it.
• 8:36
  It shows just how tangled these relationships
• 8:38
  are becoming and maybe how desperate Microsoft
• 8:40
  is for the absolute best performance in its core
• 8:42
  products, no matter where it comes from. That
• 8:44
  is something. OK, so that's the practical rift.
• 8:46
  What about the philosophical one, the ethical
• 8:48
  divide? This one's much more public. It centers
• 8:50
  around what some people are calling the sex bot
• 8:52
  capability, allowing AI chat bots to generate
• 8:55
  erotic content. Ah, right. I saw headlines about
• 8:58
  this. Microsoft's AI boss, Mustafa Suleiman,
• 9:02
  came out very strongly against it. He publicly
• 9:04
  stated Microsoft will not follow OpenAI or Elon
• 9:07
  Musk's XAI in offering, quote, simulated erotica,
• 9:12
  calling that kind of capability very dangerous.
• 9:16
  Okay, strong words. How did OpenAI respond? Sam
• 9:19
  Altman. Baltman defended their decision. He emphasized
• 9:22
  that using chat GPT for this requires strict
• 9:24
  age verification first, and he framed it as treating
• 9:27
  adults like adults, saying OpenAI is, quote,
• 9:30
  not the elected moral police of the world. So
• 9:32
  a clear difference in philosophy there. Microsoft
• 9:34
  taking a more cautious, maybe paternalistic stance
• 9:37
  and OpenAI championing user freedom within limits.
• 9:40
  Exactly. And what this divergence shows you,
• 9:42
  the listener, is that as this AI race matures,
• 9:45
  even the biggest players are making very different
• 9:46
  choices about what matters most. Is it performance?
• 9:49
  Is it ethical boundaries? Is it user freedom?
• 9:52
  And those choices directly influence the tools
• 9:54
  that end up on your computer or phone. Fascinating
• 9:57
  stuff. Really is. The whole landscape is shifting
• 9:59
  under our feet. Okay, before we wrap things up
• 10:02
  and connect all these dots, let's quickly thank
• 10:04
  our other fantastic sponsor. Yes, huge thanks
• 10:08
  to Scony. They make premium golf shoes. If you're
• 10:11
  a golfer looking to step up your game or just
• 10:13
  your style on the course, you should definitely
• 10:15
  check out Scony. Top quality stuff. And just
• 10:18
  like with Old Glory, we've got a deal for you.
• 10:20
  Use the promo code SNARFUL. Again, S -N -A -R
• 10:23
  -F -U -L at checkout on their site. That'll get
• 10:26
  you 15 % off your entire purchase. The website
• 10:29
  is scony .com. That's S -K -O -N -I dot Senn
• 10:33
  -I -S -A -R -E -S dot com. Promo code SNARFUL.
• 10:35
  Alright, so let's quickly recap the big takeaways
• 10:37
  from the Morningstar today. First, we saw Sam
• 10:39
  Altman making these absolutely monumental $1
• 10:42
  .5 trillion infrastructure deals, but doing it
• 10:45
  his own way, bypassing traditional finance, focusing
• 10:47
  purely on the technical need first. Second, OpenAI
• 10:50
  is using its immense buying power, like with
• 10:53
  that AMD warrant deal, to actively try and reshape
• 10:55
  the chip market, fostering competition against
• 10:58
  NVIDIA. And third, the bedrock partnership between
• 11:01
  OpenAI and Microsoft is visibly changing. Microsoft
• 11:05
  is looking elsewhere, Anthropic via AWS, for
• 11:08
  better performance in office, and they're publicly
• 11:10
  disagreeing on fundamental ethical questions
• 11:13
  about AI capabilities. It really feels like a
• 11:15
  new phase is beginning. Yeah. And, you know,
• 11:18
  connecting this back, I remember Altman saying
• 11:19
  recently in an interview that he realized he
• 11:22
  was wrong about previously being against vertical
• 11:26
  integration and... you know, controlling your
• 11:28
  whole supply chain. Right. He said traditional
• 11:30
  economic theory, which usually favors specialization,
• 11:34
  just didn't apply to open AI situation because
• 11:36
  their need for resources for compute was just
• 11:38
  so unique and massive. So that kind of makes
• 11:40
  you wonder, doesn't it? If the sheer scale and
• 11:42
  demands of AI forced open AI to basically throw
• 11:46
  out one long held economic principle. Yeah, it
• 11:49
  leads to a pretty provocative thought for you,
• 11:51
  our listener. What other established rules of
• 11:54
  business or traditional economic assumptions
• 11:56
  do you think this AI revolution might completely
• 11:58
  overturn next? Something to chew on. Definitely
• 12:01
  something to think about. A lot happening out
• 12:02
  there. We're here to help it's narful, so reach
• 12:04
  out on our website if you have any more questions.
• 12:10
  That's a wrap on today's top stories. Remember,
• 12:14
  tech doesn't have to be complicated, and we're
• 12:16
  here to help connect the dots. If you've got
• 12:18
  questions or want to learn more, visit us at
• 12:21
  snarful .com. Thanks for listening, and we'll
• 12:24
  catch you next time.