Latest / Connecting the Dots / Microsoft’s Billion-Dollar Bet, Valve’s Hardware Comeback, and the Markets That Move It All

Transcript

Microsoft’s Billion-Dollar Bet, Valve’s Hardware Comeback, and the Markets That Move It All · Connecting the Dots

• 0:00
  Today's tech news moves fast, but understanding
• 0:03
  it shouldn't be hard. That's why Snarfl 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. Good morning, good afternoon, and welcome
• 0:17
  back to Connecting the Dots. I'm Alex. And I'm
• 0:20
  Morgan. It is Thursday, November 13th, 2025,
• 0:23
  and we are tracking some... Truly wild contrasts
• 0:27
  today. Absolutely. Whether it's the weather systems
• 0:30
  or, you know, the pressure cooker that is Silicon
• 0:32
  Valley. Speaking of atmosphere, if you're out
• 0:34
  here with us in Sacramento, you know what we're
• 0:36
  dealing with. Oh, yeah. An atmospheric river.
• 0:38
  We've got heavy rain, potential for flooding,
• 0:41
  and a wind advisory that's running all the way
• 0:43
  through 1 .00 p .m. today. Definitely a stay
• 0:47
  inside kind of day. It is. And what a stark contrast
• 0:50
  to the rest of the country. A huge one. Further
• 0:52
  east, they're looking at dozens of record high
• 0:54
  temperatures. both maximum and minimums. It's
• 0:57
  like the whole country is in a thermal vice grip.
• 0:59
  Wow. Now, let's let's talk balance sheets. The
• 1:03
  market saw a small pullback yesterday, but the
• 1:06
  overall optimism for the week is still really
• 1:08
  strong. It is the major indices took a slight
• 1:11
  dip, but the trend is holding. Yeah, the Dow
• 1:13
  is still up over 2 % since Monday, which I mean,
• 1:16
  that shows market optimism is really rooted in
• 1:18
  future growth. Mm -hmm and Bitcoin just keeps
• 1:20
  on defying gravity. It's showing real strength
• 1:23
  closed up nearly a percent at over a hundred
• 1:25
  and two thousand dollars The appetite for that
• 1:29
  kind of transformative tech and you know the
• 1:31
  assets that back it It's clearly not slowing
• 1:33
  down not at all now before we jump into the Morningstar
• 1:37
  report a quick word from a great sponsor old
• 1:40
  glory If you're starting to think about holiday
• 1:43
  shopping, maybe for that techie friend or someone
• 1:46
  in the family, Old Glory is definitely the place
• 1:49
  to check out. They have all this iconic music
• 1:52
  and sports fan merch that goes way beyond just
• 1:55
  a basic t -shirt. Really great quality gifts.
• 1:58
  And for you, our listeners, you can get 15 %
• 2:00
  off your entire purchase. All you have to do
• 2:02
  is use the promo code SNARFL at checkout. That's
• 2:05
  Old Glory, and you can find them over at oldglory
• 2:08
  .com. Now, let's connect the dots on some massive
• 2:11
  tech bets. And why they matter to your business.
• 2:13
  We're going to start, fittingly, right at the
• 2:16
  very top. Microsoft CEO Satya Nadella. He just
• 2:18
  secured a record $96 .5 million pay package for
• 2:23
  fiscal year 2025. That is a 22 % jump from last
• 2:28
  year. It's an incredible number. $96 .5 million.
• 2:31
  I mean, that figure is almost impossible to internalize.
• 2:34
  And we know exactly why it's that high. 84 .2
• 2:37
  million of that came directly from Stock Awards.
• 2:40
  Tied to Microsoft's just stellar market performance
• 2:43
  during this whole AI boom. Right, which drove
• 2:46
  their revenue up to an enormous $281 .7 billion.
• 2:50
  It's the ultimate validation really for their
• 2:52
  strategic direction. But to understand this...
• 2:55
  this level of success, you really have to go
• 2:57
  back in time a bit. You do. Back to 2019. You
• 3:01
  have to remember the intense skepticism when
• 3:03
  they first decided to invest a billion dollars
• 3:06
  in OpenAI. Which was, at the time, a non -profit
• 3:09
  venture. And I think this is the key insight
• 3:11
  for any leaders listening today. The risk was
• 3:14
  enormous. It was. Bill Gates himself was reportedly
• 3:17
  pretty wary of the whole thing. He apparently
• 3:19
  told Nadella he was going to burn this billion
• 3:20
  dollars. Burn this billion dollars? I mean that
• 3:22
  quote just perfectly encapsulates the kind of
• 3:25
  high stakes risk you have to be willing to take
• 3:26
  at that level. Exactly. It was a huge uncertain
• 3:30
  bet. on a technology that really hadn't matured
• 3:33
  yet. Now, their total investment has grown to
• 3:36
  over $13 billion. But that 27 % stake they have
• 3:39
  in OpenAI's restructured for -profit group, it's
• 3:43
  conservatively valued at $135 billion. Just staggering.
• 3:48
  That massive reward comes directly from that
• 3:51
  counterintuitive high -risk move. So beyond the
• 3:55
  money, Nadella is offering some pretty specific
• 3:58
  advice for other leaders on how to navigate this
• 4:01
  AI era. He argues it's not just a tech shift.
• 4:04
  No, it's a business model shift and an organizational
• 4:06
  shift all at once. The real core challenge is
• 4:09
  that leaders have to learn the new production
• 4:12
  function. Exactly. So learning how work actually
• 4:14
  gets done when a tool like Copilot is writing
• 4:17
  code or when AI is generating knowledge, that's
• 4:20
  step one. Right. But the critical piece that
• 4:23
  he stresses is the need to unlearn old habits.
• 4:26
  You just can't cling to legacy model. You can't.
• 4:28
  OK, but that's why I kind of want to push back
• 4:30
  a little. I mean, that sounds great coming from
• 4:32
  a CEO who just banked 96 million on a long term
• 4:35
  bet. Sure. But for a professional who's managing
• 4:38
  quarterly budgets, how do you actually operationalize
• 4:42
  unlearning when? The old way that profitable
• 4:45
  legacy process is still bringing in revenue today.
• 4:49
  Well, that is the defining tension point of this
• 4:51
  whole transition, isn't it? It is. Nadella's
• 4:53
  view is that if you wait until that new production
• 4:55
  function is, you know, clearly dominant. You've
• 4:59
  already lost. You've already lost the game. Unlearning
• 5:02
  means you have to systematically defund, decommission,
• 5:05
  and shift resources away from those legacy processes.
• 5:09
  And you have to do it before they stop being
• 5:11
  profitable. That's the hard part. It requires
• 5:13
  treating your organization's adaptability as
• 5:15
  the most valuable thing your business has. It
• 5:17
  sounds like an act of faith, really, backed by
• 5:20
  some very strong strategic foresight. It is.
• 5:22
  Now let's pivot to our second story, which looks
• 5:25
  at the brutal flip side of Microsoft's AI strategy.
• 5:29
  We're talking about execution failure, a failure
• 5:32
  in the infrastructure war. This is the messy
• 5:34
  part of the story. Microsoft was, I mean, incredibly
• 5:38
  aggressive at first. Planning these massive AI
• 5:41
  training clusters for open AI, these gigantic
• 5:44
  data center campuses like Fairwater. They were
• 5:47
  clearly all in on building the physical compute
• 5:49
  power that this revolution needs. But then...
• 5:52
  They had what the industry is calling the big
• 5:55
  pause. Right. In Q2 of 2024, they suddenly just
• 5:59
  hit the brakes, froze over 3 .5 gigawatts of
• 6:03
  planned data center capacity that was slated
• 6:05
  all the way through 2028. Why? Why the sudden
• 6:08
  stop? It was a miscalculation of demand. They
• 6:10
  misjudged XPU cloud demand. OK, let's break that
• 6:13
  down. When we say XPU, what are we talking about?
• 6:15
  It's the industry catch all term for specialized
• 6:17
  accelerators. So think primarily of Nvidia GPUs.
• 6:21
  The chips that run and train these large AI models.
• 6:24
  Exactly. Microsoft just failed to predict the
• 6:26
  sheer explosive scale of demand for them. So
• 6:28
  they slow down building their own compute engine
• 6:30
  and the consequences were, well, they were immediate
• 6:33
  and they were dramatic. Absolutely. This allowed
• 6:35
  a competitor Oracle, to essentially come in and
• 6:39
  steal the lunch that was meant for Microsoft.
• 6:41
  So this sluggishness cost them the rumored hundred
• 6:44
  billion dollar Stargate contract with OpenAI.
• 6:47
  It did. Because Oracle executed so much faster.
• 6:51
  I mean, think about this comparison. Okay. While
• 6:53
  Microsoft's projects were still in planning,
• 6:55
  maybe early construction after two years, Oracle
• 6:58
  operationalized their cluster in Abilene, Texas,
• 7:01
  within months of breaking ground. Months. What
• 7:04
  did Oracle do so much better? Was it just pouring
• 7:07
  concrete faster? It was operational innovation
• 7:10
  and supply chain mastery. How so? Oracle leaned
• 7:14
  really heavily into modular data center design,
• 7:17
  prefabricated units that let them bypass all
• 7:20
  the traditional slow construction methods. So
• 7:22
  they treated it like a high -speed manufacturing
• 7:24
  challenge. Exactly. Not a standard real estate
• 7:26
  venture. And Microsoft, using more traditional
• 7:29
  building methods, just couldn't match that speed.
• 7:31
  And that speed difference is just staggering
• 7:33
  when you look at the financial toll. It is. Stepping
• 7:36
  away from Stargate meant sacrificing an estimated
• 7:39
  $150 billion of potential gross profits over
• 7:42
  the contract. One operational failure basically
• 7:45
  funded Oracle's entire entry into the AI factory
• 7:49
  business. That number, $150 billion, that's the
• 7:52
  staggering cost of not unlearning your execution
• 7:55
  habits. So now Microsoft is facing this urgent
• 7:59
  capacity crunch, and they're forced to rent GPUs
• 8:02
  from what are called Neo Clouds. Yeah, let's
• 8:05
  define that term. We've seen it pop up in the
• 8:07
  sources. So what is a Neo Cloud, and why is this
• 8:10
  so damaging to Azure's business model? A Neo
• 8:13
  Cloud is essentially a smaller, more specialized
• 8:15
  cloud operator. They're often focused purely
• 8:17
  on offering high -end GPU compute. So they bought
• 8:21
  the chips and built the capacity when Microsoft
• 8:23
  didn't. That's right. So when Azure rents from
• 8:25
  them to resell that capacity, they're paying
• 8:27
  a near retail price for the power and hardware
• 8:29
  they should have built at wholesale cost. So
• 8:32
  they're literally paying their competitors supply
• 8:34
  chain margin. Exactly. It immediately slashes
• 8:37
  their profit per compute hour, maybe by 30 to
• 8:39
  40 percent. It's described as the worst option
• 8:42
  for their margins. It is because it forces the
• 8:44
  giant to act like a desperate customer in its
• 8:47
  own market. And if that wasn't complicated enough,
• 8:49
  there's a whole other layer with the hardware
• 8:51
  dynamic. Right. Microsoft has access to OpenAI's
• 8:54
  custom chip IP, the one developed with Broadcom.
• 8:57
  And given that OpenAI's chip trajectory looks
• 9:00
  strategically better than Microsoft's own...
• 9:03
  in -house chip. The Maya 100 chip. Yeah, Microsoft
• 9:07
  might very well end up using OpenAI's custom
• 9:10
  silicon to serve their models. They're deeply
• 9:12
  reliant on this partnership for their hardware,
• 9:15
  the intelligence, even as they struggle to build
• 9:17
  the infrastructure the body to support it. It's
• 9:20
  a real tightrope walk between being a dominant
• 9:23
  partner and a dependent customer. All because
• 9:25
  of those execution bottlenecks. Fascinating.
• 9:27
  So we've talked about strategic billions and
• 9:29
  execution failures. Let's shift gears and connect
• 9:32
  those ideas to a completely different sector.
• 9:35
  Gaming. Let's talk about Valve. Valve is back.
• 9:37
  They are officially resurrecting the steam machine.
• 9:40
  That little cube shaped gaming PC. and it's coming
• 9:43
  with a completely redesigned Steam controller.
• 9:46
  We're expecting this to arrive in early 2026.
• 9:49
  Now, the first Steam Machine attempt back in
• 9:50
  2015 was famously a flop. It was. So how does
• 9:55
  this fit our theme of learning and unlearning
• 9:57
  mistakes? Well, Valve's developers are very clear
• 10:00
  on this. The original attempt failed because
• 10:04
  getting PC games to run reliably on that Linux
• 10:07
  -based Steam OS was... Well, it was a nightmare
• 10:10
  for users. They had the hardware, but the software
• 10:12
  just wasn't ready. Exactly. But the success of
• 10:15
  the Steam Deck since then has proven the viability
• 10:18
  of SteamOS and Proton, their compatibility layer.
• 10:21
  So they solved the software problem first this
• 10:22
  time, which gives this new effort a much better
• 10:26
  shot. a real shot. And the foundation is solid.
• 10:29
  Let's talk about the specs on this cube. They
• 10:30
  are swinging for the fences in terms of performance.
• 10:33
  They really are. The new Steam Machine is running
• 10:35
  that Linux -based Steam OS. It has a semi -custom
• 10:38
  AMD Zen 4 CPU and an RDNA 3 GPU. And Valve is
• 10:43
  claiming it's six times more powerful than the
• 10:45
  Steam Deck. Targeting smooth 4K resolution at
• 10:47
  60 frames per second using AMD's FSR upscaling.
• 10:51
  And that compact cube design, it's about six
• 10:53
  inches on all sides. That's pretty bold for a
• 10:55
  4K capable PC. That design is driven entirely
• 10:58
  by thermal requirements. Right. The cube shape
• 11:00
  is necessary because it's housing a massive internal
• 11:04
  heatsink and a single custom 120 millimeter fan.
• 11:08
  So it's a deliberate choice for efficiency and
• 11:12
  for quiet operation in a living room. Exactly.
• 11:14
  It's built to be powerful without sounding like
• 11:17
  a jet engine is taking off. And the controller
• 11:20
  overhaul is just as important. The original one
• 11:22
  was notoriously polarizing. This new one incorporates
• 11:26
  all the best lessons from the Steam Deck Controller.
• 11:29
  It keeps the essential features, the dual trackpads,
• 11:31
  the four grip buttons. But the critical update
• 11:34
  is the dual parallel joysticks. What about stick
• 11:37
  drift, the plague of every modern controller?
• 11:40
  Valve is aiming to virtually eliminate it. They're
• 11:43
  using this sophisticated sensor tech tunneling
• 11:46
  magnet or resistance to detect stick movement.
• 11:48
  We don't need to get bogged down in the acronym.
• 11:50
  No, but the benefit is clear. This significantly
• 11:53
  reduces stick drift, allows for a smaller dead
• 11:55
  zone so your controls are much tighter and they
• 11:57
  won't wear out easily. Plus there's that quirky
• 11:59
  addition of capacitive grip sense. Yeah, the
• 12:02
  controller knows when you're physically holding
• 12:04
  it. Practical applications could be, what, instant
• 12:07
  wake up, seamless profile switching. All of that,
• 12:11
  and Add to it a solid 35 plus hours of estimated
• 12:15
  battery life, and Valve is signaling they are
• 12:17
  very serious about this second act. A budget
• 12:20
  friendly 4K PC alternative for the living room
• 12:23
  in 2026. They learned their lesson from 2015
• 12:26
  and they are ready to execute on a much more
• 12:28
  stable platform. All right, before we wrap up
• 12:30
  today's Connecting of the Dots, we want to thank
• 12:32
  another sponsor, Scony. Scony makes premium golf
• 12:35
  shoes. They specialize in performance and style
• 12:38
  on the course. A fantastic gift option for the
• 12:40
  golfer in your life, especially with the holidays
• 12:42
  coming up. And just like with Old Glory, you
• 12:45
  can get 15 % off your entire purchase using promo
• 12:47
  code SNARFL at checkout. That's SCONI, S -K -O
• 12:51
  -N -I dot com. So, to quickly recap our Morningstar
• 12:54
  report. We saw Satya Nadella's massive payoff,
• 12:57
  which validated his risky, long -term strategic
• 13:00
  AI bet. And proved the immense value of organizational
• 13:03
  adaptability and, you know, foresight. But that
• 13:06
  success is currently being hobbled by critical
• 13:08
  infrastructure execution failures. A $150 billion
• 13:12
  mistake that forced Microsoft to pay premium
• 13:15
  margins to its competitors. And handed significant
• 13:19
  market share over to Oracle. And finally, we
• 13:21
  looked at Valve, a company applying the hard
• 13:24
  -won lessons from their Steam Deck success. For
• 13:26
  a second... and much smarter attempt at entering
• 13:29
  the living room PC market. It all comes down
• 13:31
  to strategic foresight married with flawless
• 13:34
  execution. Or the consequences of failing to
• 13:37
  unlearn old habits. We're here to help it snarfle,
• 13:39
  so reach out on our website if you have any more
• 13:41
  questions. That's a wrap on today's top stories.
• 13:48
  Remember, tech doesn't have to be complicated,
• 13:50
  and we're here to help connect the dots. If you've
• 13:53
  got questions or want to learn more, visit us
• 13:56
  at snarful .com. Thanks for listening, and we'll
• 13:59
  catch you next time.