Latest / Connecting the Dots / $30 Billion Rounds and the Race for AI Supremacy

Transcript

$30 Billion Rounds and the Race for AI Supremacy · 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. Good morning, everyone. You are listening
• 0:17
  to Connecting the Dots. I'm Alex. And I'm Morgan.
• 0:20
  And we are coming to you live from the Snarful
• 0:22
  Solutions Group headquarters. Right here in sunny
• 0:27
  Sacramento. It is Friday February 13th 2026 Friday
• 0:32
  the 13th the day where you know logic usually
• 0:36
  takes a backseat to superstition Did you manage
• 0:39
  to get to the studio without? walking under any
• 0:41
  ladders? I successfully navigated the obstacle
• 0:44
  course. No ladders, no black cats, and I was,
• 0:47
  well, very careful with my shaving mirror this
• 0:49
  morning. Good. Though, honestly, looking at the
• 0:51
  tech headlines we have to get through today,
• 0:54
  it feels like a few major CEOs might be feeling
• 0:57
  a little superstitious, or maybe just paranoid.
• 0:59
  Paranoid is a strong word. Yeah. But when you
• 1:01
  see the amount of money flying around in these
• 1:03
  reports, hypervigilant is probably the minimum
• 1:07
  requirement to keep your job. Absolutely. Before
• 1:09
  we get into the billions and, you know, the robot
• 1:12
  wars, let's make sure the actual physical world
• 1:14
  is still behaving. If you're tuning in from the
• 1:16
  southern plains, you probably don't need me to
• 1:18
  tell you to grab an umbrella. An umbrella might
• 1:20
  not even cut it. We're looking at a pretty intense
• 1:22
  storm system. It's bringing heavy rainfall and
• 1:25
  some serious flood threats to the south today.
• 1:28
  It is going to be a wet one. And if you're joining
• 1:30
  us from way up north in Alaska, specifically
• 1:32
  the west coast or central interior, you're dealing
• 1:34
  with the opposite problem. There's a massive
• 1:36
  bearing system moving in. Yeah, we're talking
• 1:38
  snow, high winds and blizzard conditions. So
• 1:41
  stay inside if you can. Meanwhile, here in Sacramento,
• 1:44
  we are living in a bubble of absolute calm. It
• 1:47
  is mostly sunny, looking at a high of 64 degrees.
• 1:50
  It's basically patio weather in February. I am
• 1:54
  definitely taking lunch outside. Now shifting
• 1:57
  gears to the financial markets, it was a bit
• 1:59
  of a rough ride yesterday for the traditional
• 2:00
  investors. Rough is putting it politely. The
• 2:03
  Dow closed down 1 .43 % yesterday, sitting at
• 2:07
  49 ,451. Wow. The S &P 500 took a hit too, closing
• 2:12
  down nearly 1 .8%. It seems like Wall Street
• 2:15
  is catching his breath or... I don't know, maybe
• 2:18
  anticipating some volatility. But interestingly
• 2:20
  enough, while the traditional markets are sweating,
• 2:23
  the crypto world is, well, it's doing its own
• 2:26
  thing. Bitcoin is actually the bright spot this
• 2:29
  morning, trading up about 1 .59%, hovering right
• 2:32
  around $67 ,250. It's that classic decoupling
• 2:37
  we see sometimes. When the institutional money
• 2:39
  gets nervous about, you know, interest rates
• 2:42
  or earnings, the digital assets sometimes act.
• 2:45
  that alternative safe haven. Or at least a volatile
• 2:48
  safe haven. Exactly, a very volatile safe haven.
• 2:52
  Speaking of volatility and needing the energy
• 2:54
  to track all these moving parts, I need to make
• 2:56
  sure I'm fully caffeinated. And you know what
• 2:59
  that means. I have a feeling it involves our
• 3:01
  friends at Strong Coffee. You know it. Listen,
• 3:03
  if you are still drinking that watery diner stuff
• 3:06
  or the instant coffee that tastes like battery
• 3:07
  acid, you have to stop. You deserve better. You
• 3:10
  really do. Strong Coffee is premium instant coffee
• 3:14
  that actually tastes like a traditional brew.
• 3:16
  It really does. But the key differentiator isn't
• 3:19
  just the taste. It's what's in it. It's the formulation.
• 3:23
  It's packed with 15 grams of protein, MCTs, and
• 3:26
  adaptogens like ashwagandha. Ashwagandha is the
• 3:29
  game changer for me personally. It just sort
• 3:31
  of levels out the caffeine hit so you get clean,
• 3:34
  sustained energy. Right. No jitters, no crash.
• 3:36
  Just pure focus. If you want to upgrade your
• 3:39
  morning routine, head over to strongcoffeecompany
• 3:41
  .com. And because we love you guys, use the promo
• 3:44
  code SNARFUL, that's S -N -A -R -F -U -L, for
• 3:48
  20 % off your entire purchase. Caffeine secured.
• 3:52
  Let's get into the news. We have a theme emerging
• 3:59
  today and it feels like we are moving from the,
• 4:02
  I don't know, the experimental phase of AI into
• 4:05
  something much heavier. Oh yeah. It feels like
• 4:07
  the industry is putting on its steel toed boots.
• 4:10
  That is a great analogy. We are witnessing the
• 4:12
  industrialization of artificial intelligence.
• 4:15
  We aren't playing in the sandbox anymore. We
• 4:16
  are pouring concrete and nowhere is that more
• 4:19
  obvious than with the story we have to start
• 4:21
  with. Anthropic. We've been tracking the AI funding
• 4:24
  race for a while, but Anthropic just dropped
• 4:27
  a hammer. They've closed their series G funding
• 4:29
  round. Break down the numbers for us because
• 4:31
  they are getting hard to wrap your head around.
• 4:33
  They are eye watering numbers. So Anthropic has
• 4:36
  raised $30 billion in fresh cash. $30 billion
• 4:40
  with a B. With a B, correct. Yeah. And that cash
• 4:43
  injection pushes their post -money valuation
• 4:46
  to $380 billion. Whoa. And just to give you some
• 4:51
  context on the speed of this, that is double
• 4:53
  what the company was worth just last September.
• 4:56
  So they doubled their value in under six months.
• 4:59
  That firmly plants them as the second most valuable
• 5:02
  private tech startup in the world, right? Right
• 5:04
  behind OpenAI, which is hovering around $500
• 5:06
  billion. And they've pulled ahead of SpaceX and
• 5:09
  Elon Musk's XAI. It is becoming a very, very
• 5:13
  crowded podium at the top. OK, I want to pause
• 5:15
  on the why here. Because $30 billion is the GDP
• 5:18
  of a small nation. Why does a software company
• 5:21
  need that much liquid cash right now? I mean,
• 5:24
  I know salaries in the Bay Area are high, but
• 5:26
  they aren't that high. This is the critical shift
• 5:28
  in understanding the current landscape. We aren't
• 5:30
  really talking about software companies anymore.
• 5:32
  We're talking about infrastructure companies.
• 5:34
  The cost of training these next generation models,
• 5:36
  what we call a compute, is astronomical. So this
• 5:40
  money isn't for people. It's for hardware. Exactly.
• 5:43
  It's for buying tens of thousands of GPUs from
• 5:46
  Nvidia. It's for building massive data centers.
• 5:49
  It's for the electricity bills to cool those
• 5:51
  data centers. To stay at the table, you have
• 5:54
  to spend tens of billions on capital expenditures.
• 5:57
  So if you don't build a factory, you can't build
• 6:00
  the product. You got it. And who is footing this
• 6:02
  bill? The round was led by GIC from Singapore
• 6:05
  and COA2. But here's the interesting and somewhat
• 6:09
  circular part of this whole economy. Okay, it
• 6:12
  includes a portion of that massive 15 billion
• 6:16
  dollar commitment from Microsoft and Nvidia that
• 6:18
  we heard about previously wait Let me get this
• 6:20
  straight Nvidia invests in anthropic and then
• 6:23
  anthropic takes that money and turns around to
• 6:25
  buy chips from Nvidia. It's a very tight feedback,
• 6:28
  Luke, yes. It sounds a bit like money laundering
• 6:30
  if you don't look closely, but it's actually
• 6:32
  strategic alignment. Right. Nvidia wants to ensure
• 6:34
  there are customers for its chips, so it funds
• 6:36
  the customers. Right. But Anthropic has to do
• 6:39
  it because they are trying to keep pace not just
• 6:41
  with OpenAI, but with Google. And Google has
• 6:44
  their own war chest. War chest is an understatement.
• 6:48
  Google alone is planning to spend $185 billion
• 6:52
  on capital expenditures this year. That number
• 6:55
  is staggering. $185 billion. It really drives
• 6:59
  home that this isn't a startup game anymore.
• 7:02
  Two guys in a garage cannot compete with a $185
• 7:05
  billion CapEx budget. No, they can't. This is
• 7:08
  now an industrial arms race. If you don't have
• 7:11
  the capital to burn, you are effectively out
• 7:13
  of the frontier model business. You can build
• 7:16
  apps on top of these models, but you can't build
• 7:18
  the models themselves. But here is the pushback
• 7:21
  I hear from people outside the tech bubble. Raising
• 7:23
  money is one thing. That's just investors betting
• 7:26
  on you. Actually making money. Selling a product.
• 7:28
  That's the real test. Is Anthropic actually a
• 7:31
  business or is it just a money pit? And that
• 7:33
  brings us to our second dot. And this is arguably
• 7:36
  more significant than the valuation. We're moving
• 7:39
  from hype to revenue. According to the Morningstar
• 7:42
  report, Anthropic isn't just burning cash. They
• 7:45
  are bringing it in at a rate that is, well, frankly,
• 7:47
  it's historic. What are the actual sales numbers?
• 7:50
  Anthropic's annual run rate revenue has hit $14
• 7:54
  billion. $14 billion. A year. Yes. And the growth
• 7:58
  curve is vertical. They have been growing over
• 8:01
  10x annually for the past three years. I want
• 8:03
  to contextualize that for you. In the sauce world
• 8:05
  software as a service, getting to $100 million
• 8:08
  in revenue is usually a reason to pop champagne.
• 8:11
  Getting to a billion usually takes a decade.
• 8:13
  Hitting $14 billion this fast. That's just unprecedented.
• 8:17
  It's exponential in the truest sense. And what's
• 8:20
  interesting is where that money is coming from.
• 8:21
  A huge chunk is from a specific tool called Claude
• 8:25
  Code. That single tool has a run rate of over
• 8:28
  two and a half billion dollars. That's fascinating
• 8:30
  because I feel like when the general public thinks
• 8:32
  of AI, they think of chachibiki. They think of
• 8:36
  consumer stuff writing poems, planning vacation,
• 8:39
  maybe helping with a wedding speech. But Anthropic
• 8:42
  seems to be playing a different game. That is
• 8:44
  the strategic differentiator. OpenAI became a
• 8:47
  household name with consumers. Anthropic has
• 8:49
  carved out a massive nooch in the enterprise
• 8:52
  sector. They are the suit -and -tie AI. The serious
• 8:55
  option for serious business. Exactly. Their CFO,
• 8:59
  Krishna Rao, said explicitly, Claude is becoming
• 9:02
  critical to how businesses work. They aren't
• 9:04
  trying to be a fun chatbot friend. They want
• 9:06
  to be the underlying infrastructure of your company.
• 9:09
  The engine inside the car. Right. And are they
• 9:11
  profitable yet? with 14 billion coming in. Not
• 9:14
  yet. Which is typical for this stage of massive
• 9:17
  infrastructure build out. But they have a timeline.
• 9:20
  They're aiming to be profitable by 2028. And
• 9:23
  the reports say they're actually quite frugal
• 9:25
  compared to their competitors. They are. Okay.
• 9:27
  Calling a company that just raised 30 billion
• 9:30
  dollars frugal feels like a stretch. That's like
• 9:32
  calling a billionaire budget conscious because
• 9:34
  they didn't buy a third island. It sounds absurd.
• 9:37
  I know. But Relatively speaking, they are focused.
• 9:40
  They aren't spending as much on flashy viral
• 9:43
  stuff like video generation tools. Things that
• 9:46
  burn a ton of compute for what is essentially
• 9:48
  marketing demo reels. Anthropic is focusing on
• 9:52
  coding, legal analysis, heavy enterprise lifting.
• 9:56
  They are spending their money on the boring stuff
• 9:57
  that makes money. So for the listener driving
• 10:00
  to work right now or the manager trying to figure
• 10:02
  out their budget for next quarter, what does
• 10:04
  this signal? It signals that AI is graduating.
• 10:07
  It's moving from being a novelty to being a utility
• 10:10
  if Anthropic is hitting these revenue numbers.
• 10:13
  It means companies are paying for it and relying
• 10:15
  on it daily It's becoming like electricity or
• 10:17
  the internet you just pay the bill because you
• 10:19
  can't operate without it That is a huge shift.
• 10:22
  We are moving from look at this cool trick to
• 10:25
  I literally cannot do my job without this. Precisely.
• 10:29
  OK, so Anthropic is the disciplined, enterprise
• 10:32
  -focused contender. But the reigning champ, OpenAI,
• 10:36
  isn't just going to sit there and let them take
• 10:38
  the crown. They aren't exactly known for being
• 10:40
  passive. Absolutely not. And that leads us to
• 10:42
  the third part of our report. OpenAI has launched
• 10:45
  a counteroffensive, and the internal language
• 10:47
  they are using is intense. I read this in the
• 10:51
  report, a code red. That sounds like something
• 10:53
  out of a submarine movie. Are they manning the
• 10:56
  battle stations? It feels like it. Sam Altman,
• 10:59
  the CEO, declared this code red to rally the
• 11:01
  staff. The message is clear. Do not fall behind
• 11:04
  Google and Anthropic. The lead they had two years
• 11:07
  ago, it's gone. It's a dead heat now. It sounds
• 11:10
  like the pressure cooker is really on over there.
• 11:13
  And to back up this code red, they have just
• 11:14
  released GPT 5 .2. I feel like I need a spreadsheet
• 11:18
  just to keep up with the version numbers at this
• 11:20
  point. What is special about 5 .2? They're calling
• 11:23
  it the most capable model for professional knowledge
• 11:27
  work. And they have the benchmarks to try and
• 11:29
  prove it. It scored a 70 .9 % on the GDP VAL
• 11:33
  test. For the non -engineers listening, and for
• 11:35
  me, is 70 .9 % actually good. It's very good.
• 11:38
  It's a significant step up in reasoning capability.
• 11:40
  OK. But there was one specific stat from their
• 11:43
  internal testing that really jumped out at me.
• 11:45
  It's going to make some people nervous, specifically
• 11:47
  if they work in finance. The investment banking
• 11:49
  stat? I saw that. Yes. In internal tests, this
• 11:52
  model outperformed junior investment banking
• 11:54
  analysts on spreadsheet modeling by 9 .3%. Ouch.
• 11:59
  Investment banking has always been this safe,
• 12:01
  high -paying job. You grind for two years on
• 12:03
  Excel, you make your money, you sacrifice your
• 12:06
  sleep, and now a robot is beating you by nearly
• 12:09
  10%. It's a sobering statistic. It shows that
• 12:13
  AI isn't just writing text anymore, it's doing
• 12:15
  complex financial modeling, error checking, and
• 12:18
  projection. That is a definition of light collar
• 12:21
  automation. Right. If your job is purely moving
• 12:24
  numbers around in a spreadsheet, the machine
• 12:26
  is getting very, very good at it. Very good.
• 12:28
  That really challenges the idea of what an entry
• 12:31
  -level job is going to look like in five years.
• 12:33
  But there is a twist in this OpenAI story that
• 12:35
  I think is even bigger than the benchmark scores.
• 12:38
  It's about the hardware. Yes. We talked earlier
• 12:40
  about how everyone is buying NVIDIA chips, but
• 12:43
  OpenAI is doing something different here, aren't
• 12:45
  they? This is a crucial detail that might get
• 12:47
  lost in the noise, but it's massive for the industry.
• 12:50
  OpenAI revealed a new model variant called GPT
• 12:53
  5 .3 Codex Spark. And this model isn't running
• 12:57
  on NVIDIA. It's running on chips from Cerebras.
• 13:00
  Cerebras. We've heard that name floating around
• 13:03
  as an NVIDIA rival. Why is this significant?
• 13:05
  Because for the last five years, NVIDIA has had
• 13:08
  a near -monopoly. If you wanted to do top -tier
• 13:11
  AI, you paid the NVIDIA tax. You waited in line
• 13:14
  for their chips. You begged for allocation. Right.
• 13:17
  OpenAI running a flagship model on Cerebras chips?
• 13:21
  That sends a signal to the entire market that
• 13:24
  the monopoly might be cracking. That changes
• 13:26
  the leverage completely. If OpenAI can say, hey,
• 13:29
  we don't need your H100s, we can use these other
• 13:30
  guys, the economics of AI could change. It's
• 13:33
  like when Apple stopped using Intel chips and
• 13:35
  made their own. Exactly. It breaks the bottleneck.
• 13:38
  Cerebras chips are designed differently. They're
• 13:40
  massive wafer scale chips designed specifically
• 13:43
  for this kind of work. If they prove viable at
• 13:46
  this scale, it's a seismic shift for the hardware
• 13:48
  supply chain. It means competition is finally
• 13:50
  arriving. Finally. And speaking of seismic shifts,
• 13:53
  we have to touch on the meta moment in this report,
• 13:56
  the self -building code. This is the part that
• 13:58
  feels like science fiction. OpenAI revealed that
• 14:01
  the vast majority of their coding tool, Codex,
• 14:04
  is now built by Codex. So the AI is writing the
• 14:07
  code for the AI. It's a recursive loop, a robot
• 14:10
  building a better version of itself. It is. And
• 14:13
  Codex now has over a million weekly active users,
• 14:17
  and it's essentially self -improving. That's
• 14:19
  the definition of exponential acceleration. When
• 14:22
  humans write code, we get tired, we make typos,
• 14:24
  we need coffee breaks, we go home to our families.
• 14:27
  When the AI writes the code, it just keeps going.
• 14:29
  And the cycle of improvement gets faster and
• 14:31
  faster. And that explains why we are seeing version
• 14:34
  5 .2, 5 .3 so quickly. The tools are building
• 14:38
  the tools. It's incredibly impressive, but also
• 14:40
  a little dizzying. It feels like the train is
• 14:43
  laying its own tracks while moving at 200 miles
• 14:45
  an hour. That is a very accurate description
• 14:48
  of the current state of Silicon Valley. So let's
• 14:51
  recap the battlefield here. You've got Anthropic
• 14:53
  raising a war chest to dominate the enterprise
• 14:56
  with a frugal safety first approach. You've got
• 14:59
  open AI hitting the panic button with a code
• 15:01
  red, releasing models that beat bankers at their
• 15:04
  own game and pivoting to new hardware to break
• 15:07
  the supply chain bottleneck. And meanwhile, the
• 15:10
  AI is literally coding itself. It's a lot to
• 15:13
  take in. But it shows that 2026 is going to be
• 15:17
  a defining year. We aren't in the experimentation
• 15:20
  phase anymore. We are in the deployment and dominance
• 15:23
  phase. Right. These companies are building the
• 15:25
  operating system for the next decade of the global
• 15:28
  economy. Deployment and dominance sounds ominous,
• 15:31
  but accurate. Well before we let our listeners
• 15:35
  go back to their hopefully still human staff
• 15:37
  jobs We have one more important connection to
• 15:39
  make and that is with your feet That's right
• 15:42
  our friends at scony scony is the premium golf
• 15:45
  shoe store that is changing the game Look, if
• 15:47
  you are walking 18 holes or just walking around
• 15:50
  the office pretending to work while you listen
• 15:51
  to us. Which we know you do. Yeah. We won't tell
• 15:54
  your boss. You need comfort and style. Skone
• 15:56
  shoes are incredibly comfortable. They look great
• 15:58
  on and off the course. I actually wear mine to
• 16:00
  the grocery store. They really bridge that gap
• 16:02
  between athletic wear and casual office wear
• 16:05
  perfectly. You don't look like you're wearing
• 16:07
  cleats, but you get the support. If you want
• 16:10
  to treat your feet, head over to theskone .com.
• 16:12
  and use the promo code SNARFL at checkout for
• 16:15
  15 % off your entire purchase. A great deal for
• 16:18
  a great shoe. All right, let's bring it all home.
• 16:20
  We covered a mountain of cash and tech today.
• 16:23
  What are the three things our listeners need
• 16:25
  to lock in before they hit the weekend? For this
• 16:28
  scale of investment, Anthropic has raised a massive
• 16:31
  $30 billion war chest. They're valued at $380
• 16:35
  billion, and they're using that money to build
• 16:37
  the physical infrastructure of AI. Huge money
• 16:40
  for huge compute. This is heavy industry now.
• 16:43
  Second, the business model is real. They hit
• 16:46
  $14 billion in revenue by focusing strictly on
• 16:49
  enterprise needs -making businesses run better,
• 16:52
  not just making funny videos. This validates
• 16:54
  the entire industry economically. It's not a
• 16:57
  bubble if the cash register is actually ringing.
• 16:59
  And third, the Counter -Strike. OpenAI is firing
• 17:03
  back with GPT -5 .2 and, crucially, new hardware
• 17:07
  partnerships with Cerebras. They're trying to
• 17:09
  innovate their way out of a code red and break
• 17:11
  the hardware monopoly that has constrained the
• 17:13
  industry. It's a clash of the titans, and we
• 17:16
  are just watching the sparks fly. It's going
• 17:17
  to be a fascinating year to watch these two duke
• 17:20
  it out. And remember, the technology you use
• 17:22
  today might be building the technology you use
• 17:25
  tomorrow. Literally, in the case of Codex. That
• 17:28
  is a thought to chew on. Well, that wraps up
• 17:31
  another edition of Connecting the Dots. We're
• 17:32
  here to help at Snarful, so reach out on our
• 17:34
  website if you have any more questions. Have
• 17:36
  a great weekend, everyone. Stay safe out there.