00:00And we have a lot to talk about here but I do want to start with tech. The drawdown that
00:04we're seeing. I wonder what you think is behind that. Is it just that tech had the most room to
00:09fall after the run up that it had or is there something more foundational there.
00:14Well I think it's more foundational. It started in late October where people started to question the kind of stimulus
00:25for this AI trade which was OpenAI. Because remember they have 1.4 trillion in capital commitments over eight years.
00:34But their run rate revenues is 20 billion. How are you going to fund 1.4 trillion?
00:41And then recently you've seen them actually step back from some of the deals they had. Most notably last week
00:48was Disney. And if you remember they launched their video generation app Sora to great fanfare.
00:54And then they struck that deal with Disney where Disney was going to put in 1 billion dollars into OpenAI.
01:00And then OpenAI said okay we're not going to do that anymore.
01:04And that follows them not developing checkout for PayPal as well as they say you know we've got to focus
01:12on what we're doing. Which kind of speaks to the fact that they've got a lot more pressure on them.
01:18And that's just one example of this.
01:19Yeah. But I think it's the most notable because you know ChatGPT really started this whole AI trade in late
01:272022.
01:28Yeah. And you know it's wild to think here we are in 2026 and that has been the defining narrative
01:34of these markets for the better part of that time.
01:37But you think about the sector as a whole when it comes to tech. Do you think that there's further
01:42re-rating to be had at the industry level?
01:45Or are we you know pretty much approaching the bottom?
01:49Well I think it's sector by sector. I think in software we're nowhere near the bottom because it's really you
01:56could say since Mark Andreessen put that piece out in the Wall Street Journal back in 2011 of software eating
02:02the world.
02:03It's been you know 14 years of up straight up in software.
02:09Well AI is going to eat a lot of these software companies alive. And so you know you haven't solved
02:15figuring out who the winners are and losers are in five months.
02:19And so I think software as a whole has a lot further to go down.
02:23Now if you step back from that and you say well what else is going on?
02:27A big positive development is agentic AI and I think OpenClaw has captured everybody's attention.
02:34And that kind of got finalized at the end of January. The original version came out at the end of
02:40November.
02:40And if you look at the amount of tokens being used by agentic AI it's 10 to 100 times more
02:47than if you're just doing a chat based question.
02:50Right. So because then you have to go off and do a lot of things go to different sites create
02:57your spreadsheet go to different sources move things around.
03:00And so it takes a lot more computing power and you can see that in the data.
03:05And so it's going to take some time.
03:07But I think as you get towards the end of the year the companies that help power that are going
03:12to see their stock start to move again.
03:14And the analogy I would draw is what we're going through right now I think is akin to what Cisco
03:19went through in 1997 and 1998.
03:22Where if you go back and you look at that in 1997 Cisco stock at one point during 1997 was
03:29down 37%.
03:30It finished the year up 31.
03:33If you go to 1998 at one point it was down 38%.
03:36It finished the year up 150.
03:40So you're going to get these questions on have we reached the peak have we reached the top and Cisco
03:46was kind of the NVIDIA of its day the most valuable company in the world at one point the internet
03:51infrastructure provider.
03:53And so I think you're kind of going through a similar situation right now but agentic AI which we've been
03:59sort of waiting for for three years open claw being the best example of that is showing that there's some
04:05value to this and this is the next phase higher.
04:08Well Dan I want to look at semis right now Philly semiconductor index had been hanging in there right now
04:14it's down about 11% in the last three days is that an area that you would feel confident putting
04:19money to work and trying to buy some of these dips as opposed to as you mentioned steering clear of
04:23software.
04:23Well again I think you have to be selective because memory has been assumed to be in under supply through
04:342027 and you saw last week with Google release of their paper talking about how to save and get more
04:42efficient on memory by at least six to eight times.
04:46Google's already implemented that you know over a year ago but the rest of these vendors are going to look
04:52at that paper and go oh well these are things we can do to get more efficient and so it
04:57kind of depends on where you are.
04:59Now if you look at microprocessors going back to what we said earlier when you have an agent that's you
05:07know going to the SEC website pulling some data going to Bloomberg pulling some data then using it all creating
05:12a spreadsheet and then going ahead and querying you on what you need to do.
05:16That's not like one task over and over again like you had with training that requires actually a lot more
05:22CPU usage and CPU demand has been absolutely crushed with the move to AI early on.
05:31Now what you're seeing is CPU usage probably is going to go up two to four X from where it
05:36was prior to agents hitting the scene in AI and so I think that's potentially where your new bottleneck is.
05:41So for me the microprocessor vendors are actually pretty interesting such as an Intel because you know they've been sort
05:49of given up for dead to some degree as you move to GPUs are where you need to be and
05:56Nvidia.
05:56So I think you have to be careful about the stuff that people think is going to be in under
06:01supply forever like memory and then say okay where's the next area where there may be some bottlenecks and that
06:08looks like it might be CPUs.
06:10And Dan just a quick question here we're running out of time how are you thinking about the prospect of
06:15some of these companies going public whether it's an open AI or an Anthropik and what that can mean for
06:19the tech ecosystem from a public equities perspective.
06:22Well I think it depends on the company again I mean I'm not a believer in open AI and I've
06:27said that from the beginning because I didn't think you could spend 1.4 trillion if you're doing 20 billion
06:32in revenues and you're going to lose over 200 billion in cash flow until you get to profitability.
06:38Anthropik on the other hand is focused on enterprise where companies will pay you for products and they're going to
06:45get to profitability sooner they're more narrowly focused burning a lot less cash and so I think it depends on
06:51the company and the sequence of them coming public.
06:54Anthropik.
Comments