Latest / Connecting the Dots / OpenAI’s Chip Chase, Amazon’s Holiday Hiring Spree, and Wayve’s Big AI Bet

Transcript

OpenAI’s Chip Chase, Amazon’s Holiday Hiring Spree, and Wayve’s Big AI Bet · 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, brought
• 0:17
  to you by Snarfle Solutions Group. Now, before
• 0:19
  we unpack the biggest stories in tech today,
• 0:22
  just a quick shout out to our sponsor, Old Glory.
• 0:25
  They're that iconic music and sports fan merch
• 0:27
  store. So if you need vintage tour stuff or,
• 0:31
  you know, the latest jersey, they've got you
• 0:32
  covered. Use the promo code SNARFLE. That's S
• 0:36
  -N -A -R -F -U -L at checkout for 15 % off your
• 0:39
  entire purchase. Check them out at OldGlory .com.
• 0:43
  I'm your host, and with me is our expert, ready
• 0:45
  to connect the dots on today's news. Hey there.
• 0:48
  Happy Monday, October 13th, 2025, everyone. I
• 0:51
  feel like a big day for infrastructure talk,
• 0:53
  definitely. And we're also seeing some market
• 0:55
  signals that maybe suggest businesses are getting
• 0:57
  a bit nervous about the final quarter. Absolutely.
• 1:00
  So here in Sacramento, it's Columbus Day, maybe
• 1:01
  a slight chance of showers later, high around
• 1:03
  62, but nationally. Things are a bit more active.
• 1:06
  Yeah, it looks like a storm hitting the East
• 1:08
  Coast, some heavy rain risk down in the southwest
• 1:10
  and four corners, and even typhoon remnants hitting
• 1:14
  western Alaska. So lots happening weather -wise.
• 1:17
  Right. And speaking of choppy conditions, the
• 1:19
  financial markets saw a bit of a bounce Friday,
• 1:21
  but the week overall, well, it was down. The
• 1:24
  Dow Jones closed Friday up 0 .56 percent. That's
• 1:28
  about 255 points. But for the whole week, it
• 1:31
  lost 1 .76 percent. Yeah, similar story for the
• 1:34
  S &P 500. You gained 0 .40 % Friday, but dropped
• 1:37
  1 .26 % over the seven days. And Bitcoin, well,
• 1:41
  it was up just a tiny bit Friday, 0 .20%, but
• 1:44
  it took a real hit over the week, down more than
• 1:46
  7 .5%. So, you know, it definitely feels like
• 1:49
  investors are being cautious, even with all the
• 1:51
  buzz around AI. Speaking of careful moves, before
• 1:53
  we get into the main stories, we want to thank
• 1:55
  Pins and Aces for sponsoring this segment. They're
• 1:58
  the place to go for premium golf apparel and
• 2:00
  accessories, really top -notch stuff. If you're
• 2:01
  hitting the course or honestly just want some
• 2:03
  sharp gear, use that promo code SNARFL at checkout.
• 2:06
  It gets you 21 % off your entire purchase. Find
• 2:08
  them at pinsanddances .com. Okay, let's get into
• 2:12
  the Morning Star Tech Report. That's our Snarfle
• 2:15
  Tech article report where we connect the three
• 2:17
  biggest stories. And today, yeah, it really kicks
• 2:18
  off with this AI arms race and a huge bet on
• 2:21
  custom hardware. That's right. Story number one.
• 2:23
  OpenAI is doubling down on custom chips and they're
• 2:26
  partnering with Broadcom to do it. The headline
• 2:29
  is they're co -designing and producing their
• 2:31
  first in -house AI processors. Custom GPUs, basically.
• 2:34
  Okay, so this isn't just like buying a few extra
• 2:38
  chips to help out their Nvidia supply. This sounds
• 2:41
  Bigger, like a whole new strategy. Oh, much bigger.
• 2:44
  We're talking a deal potentially worth around
• 2:46
  $10 billion. But the number that really makes
• 2:49
  you pause is the scale. The plan is to roll out
• 2:52
  10 gigawatts worth of these custom chips. 10
• 2:56
  gigawatts. OK, help us wrap our heads around
• 2:58
  that. What does 10 gigawatts actually mean in
• 3:00
  the real world? It's staggering. Think about
• 3:03
  this. 10 gigawatts is roughly the power needed
• 3:05
  for over 8 million US homes continuously. or
• 3:08
  another way to look at it, it's like five times
• 3:10
  the peak electricity output of the Hoover Dam.
• 3:12
  Yeah. Just for these chips. Wow. Okay, so why
• 3:15
  build that much? And why Broadcom specifically?
• 3:18
  Why not just keep pouring money into NVIDIA?
• 3:20
  Well, it's about a few things. Diversification
• 3:22
  is one. Not putting all your eggs in the NVIDIA
• 3:25
  basket. Cost control is another big one. And
• 3:28
  optimization. Amidius chips are powerful, sure,
• 3:30
  but their general purpose. By working with Broadcom,
• 3:33
  OpenAI can design chips specifically tailored
• 3:36
  to their AI models. That could mean faster training,
• 3:39
  faster inference, and probably cheaper operation
• 3:42
  in the long run compared to just buying off the
• 3:44
  shelf. Right. Tailor -made efficiency. Makes
• 3:46
  sense. Is there a specific technical edge Broadcom
• 3:50
  brings here beyond just helping design the chip?
• 3:52
  Yeah, there is. And it's actually pretty interesting.
• 3:55
  Besides co -designing the chip itself, Broadcom
• 3:57
  is going to supply all the networking gear for
• 3:59
  these new systems. And they're using high -speed
• 4:01
  Ethernet. Now, this is key because it's a direct
• 4:04
  challenge to NVIDIA's own networking tech, InfiniBand.
• 4:07
  InfiniBand has kind of been the default, almost
• 4:09
  mandatory choice for these massive AI clusters.
• 4:12
  But OpenAI going with Ethernet, well, it suggests
• 4:15
  they want a more open, non -proprietary standard
• 4:18
  for how their systems talk to each other. Hmm,
• 4:20
  interesting. An architectural power play, almost.
• 4:23
  But wait, if they're worried about relying too
• 4:25
  much on Nvidia, didn't they also just announce
• 4:27
  that massive $100 billion deal with Nvidia and
• 4:30
  another big 6 gigawatt deal with AMD? That seems...complicated.
• 4:33
  Three different suppliers. It is complicated.
• 4:36
  It's definitely a hedging strategy, you know,
• 4:37
  trying to avoid... a single point of failure
• 4:40
  if one supplier has issues. But yeah, the complexity
• 4:42
  is enormous. Think about it. You've got resources
• 4:45
  running on NVIDIA's architecture, AMD's architecture,
• 4:48
  and now your own custom Broadcom architecture,
• 4:51
  making all of that work together efficiently,
• 4:53
  globally. It's a huge integration challenge.
• 4:56
  Sam Altman calls it critical infrastructure building.
• 4:58
  But the risk, the risk is that they're betting
• 5:00
  billions. They can actually unify all these different
• 5:03
  compute pieces effectively by 2029. There's a
• 5:05
  real chance for things to get inefficient. The
• 5:07
  AI infrastructure arms race, indeed. And the
• 5:11
  entry fee is measured in gigawatts and fragmentation
• 5:15
  risk, it sounds like. OK, let's. pull on that
• 5:19
  thread of needing massive infrastructure. That
• 5:21
  also requires people, right? To build it, manage
• 5:23
  it, move the products it enables. Our second
• 5:26
  story shifts to the labor market, and oh, it's
• 5:28
  sending some pretty cautious signals for the
• 5:30
  holidays. Yeah, this focuses on Amazon. Their
• 5:33
  holiday hiring plans are always a bit of a bellwether
• 5:35
  for the retail sector. Yeah. So Amazon announced
• 5:37
  they plan to hire 250 ,000 workers for the holiday
• 5:40
  season. That includes full -time, part -time,
• 5:42
  and seasonal roles. 250 ,000? I mean, that still
• 5:45
  sounds like a lot of jobs. Why is that being
• 5:47
  framed as cautious? Because it's flat. It's the
• 5:50
  exact same number they targeted in 2023 and the
• 5:53
  same number again in 2024. So while Amazon holding
• 5:56
  steady is notable, it actually makes them stand
• 5:58
  out. The broader forecast for holiday hiring
• 6:01
  this year is being described by analysts as,
• 6:04
  well, bleak. Bleak. That's a strong word. How
• 6:06
  bleak are we talking? Potentially quite bleak.
• 6:09
  Some reports suggest that retailers overall might
• 6:11
  add fewer than 500 ,000 seasonal jobs this year.
• 6:14
  Total. Across the whole sector. If that turns
• 6:17
  out to be true, it would be the lowest number
• 6:19
  of seasonal hires since way back in 2009, right
• 6:22
  after the financial crisis. We're seeing other
• 6:25
  big retailers either staying flat like Amazon,
• 6:27
  or just not announcing their plans at all, which
• 6:30
  usually isn't a great sign. Wow. So even though
• 6:33
  Amazon is bumping pay, they're investing over
• 6:35
  a billion dollars more, getting seasonal workers
• 6:38
  to an average of $19 an hour, permanent folks
• 6:40
  to $23. Even with that investment, The rest of
• 6:45
  the market seems to be pulling back hard. It
• 6:47
  feels... different this year. Like companies
• 6:49
  are genuinely worried about Q4 spending. That
• 6:51
  seems to be exactly the signal, yes. Companies
• 6:54
  don't appear to be betting on a big holiday spending
• 6:56
  surge. Analysts are pointing to things like worries
• 6:58
  about tariffs, inflation sticking around, and
• 7:01
  just general anxiety about the job market weakening.
• 7:03
  Now e -commerce itself is still expected to grow.
• 7:06
  Online spending in the US for November and December
• 7:08
  is projected at around $253 .4 billion. That's
• 7:12
  up about 5 .3 % from last year. Okay, so people
• 7:15
  are still buying online, maybe even more so,
• 7:17
  but the overall hiring needed to support all
• 7:19
  retail, including brick and mortar, is flat or
• 7:22
  shrinking everywhere except Amazon. It really
• 7:26
  sends a nervous message about consumer confidence,
• 7:28
  doesn't it? Like, CEOs are bracing for a slowdown,
• 7:32
  not celebrating growth. Caution is definitely
• 7:34
  the word for the Q4 labor market this year. Alright,
• 7:36
  let's pivot one more time for our final Morningstar
• 7:39
  story. We've talked hardware, we've talked people.
• 7:42
  Now let's talk about the software driving the
• 7:44
  future of transportation. Specifically, self
• 7:48
  -driving tech and a London -based startup called
• 7:50
  Wave. Right, WAVE. They're making some serious
• 7:53
  waves, no pun intended. They're reportedly in
• 7:55
  talks to raise a huge amount of money. We're
• 7:57
  hearing numbers between $1 billion and $2 billion,
• 8:01
  potentially from giants like Microsoft and SoftBank.
• 8:04
  And a funding round like that could value the
• 8:06
  company somewhere around $8 billion. $8 billion.
• 8:09
  OK, they must have something pretty special going
• 8:11
  on. We followed self -driving for a while. What's
• 8:13
  WAVE's angle? What makes them different? Their
• 8:15
  core technology is something they call embodied
• 8:18
  AI. So think about the big players. like Waymo.
• 8:22
  They often rely heavily on creating these incredibly
• 8:24
  detailed 3D maps of everywhere the car might
• 8:27
  drive. LiDAR is usually key there. Waves approach
• 8:31
  is different. They focus more on software that
• 8:33
  learns primarily from watching real world traffic
• 8:35
  using cameras and machine learning. Ah, okay.
• 8:38
  So less about mapping every single curb and lane
• 8:42
  line beforehand and more about teaching the AI
• 8:44
  to understand the flow and intent within the
• 8:47
  environment, like how a human learns to drive.
• 8:50
  Exactly. In theory, this could remove the need
• 8:52
  for those constantly updated, super expensive,
• 8:55
  high -definition maps, especially across large
• 8:58
  areas. And their software is designed to be vehicle
• 9:00
  -agnostic, meaning it could potentially work
• 9:02
  in different types of cars or even delivery vans.
• 9:04
  That sounds really flexible, but if they're not
• 9:07
  using those HD maps, doesn't that remove a layer
• 9:09
  of safety redundancy, especially for tricky city
• 9:11
  driving? How do they handle that? That's really
• 9:14
  the big debate in the self -driving world right
• 9:15
  now. Wave, and obviously its investors, believe
• 9:19
  their deep learning approach can actually handle
• 9:21
  the unpredictable nature of cities better than
• 9:24
  static maps. And the money seems to be agreeing.
• 9:27
  NVIDIA, which already works with them, Wave's
• 9:29
  next -gen system, runs on NVIDIA's DoDrive AGX
• 9:32
  Thor platform. NVIDIA is apparently talking about
• 9:35
  putting in another $500 million. Okay. And we
• 9:37
  know Uber has been testing Wave tech in London.
• 9:40
  But who's actually planning to put this in cars
• 9:42
  people can buy? Well, the most interesting consumer
• 9:45
  angle right now is probably Nissan. They're planning
• 9:47
  to incorporate Wave's tech into a Level 2 autonomous
• 9:50
  system designed for urban driving. Level 2. Remind
• 9:53
  us what that means again for the listener. Right.
• 9:55
  Level two is advanced driver assistance. The
• 9:58
  car can handle steering, accelerating, braking
• 10:00
  under certain conditions. But crucially, the
• 10:03
  human driver must stay fully engaged, watching
• 10:05
  the road, ready to take over instantly. Think
• 10:08
  hands off, eyes on. Got it. So Nissan is using
• 10:12
  this for city driving assistance. When might
• 10:14
  we see that? Nissan's using a setup with 11 cameras
• 10:17
  five radars and actually one LiDAR sensor to
• 10:20
  for this system They're aiming to roll it out
• 10:22
  in Japan initially targeting their 2027 fiscal
• 10:25
  year So yeah, this huge potential funding for
• 10:28
  wave really underscores that the self -driving
• 10:30
  race is speeding up But the technologies are
• 10:32
  splitting you've got the map heavy approach and
• 10:35
  now this big bet on AI learning its way through
• 10:38
  the real world Fascinating. It really shows the
• 10:40
  autonomous future isn't just one path. It's more
• 10:42
  like a complex intersection with different ideas
• 10:45
  competing. For sure. And right now, a lot of
• 10:47
  the momentum and the money seems to be flowing
• 10:49
  towards that AI centric path. OK, before we wrap
• 10:52
  up the Morningstar report, one final mention
• 10:54
  for our sponsor, SCONI. They make premium golf
• 10:57
  shoes, super stylish, comfortable, high performance.
• 10:59
  If you need new shoes for the course, check them
• 11:01
  out. Use that promo code SNARFUL at checkout
• 11:05
  for 15 % off your whole purchase. All right,
• 11:12
  so let's quickly recap the big threads today.
• 11:14
  First, that AI infrastructure arms race is definitely
• 11:17
  accelerating. OpenAI's deal with Broadcom for
• 11:20
  10 gigawatts of cut -thumb chips highlights this
• 11:23
  push for control for diversification, but also
• 11:26
  introduces that risk of fragmented computing
• 11:28
  power. Then we saw the labor market flashing
• 11:30
  warning signs. Amazon's flat holiday hiring is
• 11:34
  a standout only because the rest of the retail
• 11:36
  forecast looks potentially like the worst since
• 11:38
  2009. It signals real caution about consumer
• 11:41
  spending in Q4. And finally the self -driving
• 11:43
  race is evolving. Wade's potential billion dollar
• 11:46
  plus funding shows a major bet on embodied AI
• 11:49
  teaching cars to learn like humans Rather than
• 11:51
  relying solely on pre -built maps driving forward
• 11:54
  those level two systems for city use Yeah that
• 11:57
  connecting thread here seems to be how these
• 12:00
  massive tech ambitions Custom ships autonomous
• 12:03
  cars are bumping up against the realities of
• 12:05
  the current economy the next year or so will
• 12:08
  really show if these multi -billion dollar bets
• 12:11
  can thrive if consumers stay cautious. So for
• 12:13
  you listening, maybe the question is, which approach
• 12:16
  wins out first? OpenAI's complex, fragmented
• 12:19
  chip strategy, or the market's cautious, perhaps
• 12:22
  fragmented approach to hiring and spending? That's
• 12:24
  a great thought to leave things on. We're here
• 12:26
  to help at Snarfl, so reach out on our website
• 12:28
  if you have any more questions. That's a wrap
• 12:33
  on today's top stories. Remember, tech doesn't
• 12:36
  have to be complicated, and we're here to help
• 12:38
  connect the dots. If you've got questions or
• 12:41
  want to learn more, visit us at snarful .com.
• 12:45
  Thanks for listening, and we'll catch you next
• 12:47
  time.