// transcript — 774 segments
0:00 Welcome to the Top Voice Podcast
0:05 Hello and welcome to the Top Voice podcast. It is February 10th. I'm
0:12 excited to have Jason Saltzman on uh from CB Insights to talk about what
0:20 actually happened with AI in 2025 at a at an ecosystem level. We talk
0:25 about AI on this show a lot as you would expect. It's an important and frequently
0:32 discussed topic in the world today. But Jason's got a unique view about the
0:37 entire market ecosystem of players, investments, funding, and opportunities
0:43 that we're going to get into. Last week, we talked about AI for small and medium
0:48 businesses, and we're going up a level today to talk about the bigger picture,
0:53 Introduction to Jason Saltzman and CB Insights
0:53 and I think what that means for 2026. Jason, I'm going to give you a chance to
0:57 tell us a little bit more about you here in just a moment. For those of you that
1:03 are tuning in to the live conversation, please do let us know where you're
1:07 tuning in from and feel free to leave us a question or comment. We love to get
1:11 those. We love to address those in the conversation. Always great to take
1:15 angles from people who are seeing this in real time. And for those of you who
1:18 might be listening on your favorite podcast, please do be sure to subscribe.
1:25 That's the best way to support the show. So, Jason, tell us who you are. Tell us
1:30 what you do. Let's start with that. >> Yeah, thanks for having me, Michael. Uh,
1:35 I'm Jason. I'm the head of insights at CB Insights, which at this point, no
1:39 matter how many times I say it, still seems like a slightly funny job title uh
1:43 or maybe the best job title that I could possibly have uh at a company with
1:49 insights in the name. Uh we do you know maybe to give the lay of the land on
1:53 sort of what we will be covering across the rest of this conversation. We do
1:57 predictive intelligence on private companies and private markets with you
2:01 know trillions of data points across 11 million plus companies with everything
2:06 from sort of predictive signals on where a company is headed uh to sort of your
2:11 standard data like funding and financials business relationships hiring
2:18 uh insights and a whole bunch more. Uh really trying to understand the
2:21 companies and technologies that will define the future and sort of how that
2:26 ecosystem interacts from you know sort of the inception stage at you know preed
2:31 seed stage companies all the way up through how the large public companies
2:35 and the incumbent players are interacting with those companies through
2:39 investments acquisitions partnerships etc. Yeah, if you love data, this is a
2:46 conversation for you. We we the the I've looked at the dashboards that you all
2:49 produce on this. We might do some screen sharing here today for those of you that
2:54 are watching the video and just to jump into that. But if you've ever enjoyed
2:59 looking at data really I have a history with CV insights from my time in the big
3:03 four world. Amazing work that you all do and and I know more about what you do.
3:07 Really, really great stuff. Uh, in addition, you left a piece out. In
3:13 addition to being a big data guy, former professional cyclist, I think if that
3:16 we're not going to talk about that today, but if anyone has questions, find
3:21 Jason. He can speak to that as well. >> I'll also say that's probably at least
3:25 where some of the data nery first came from. Cycling is a sport with a lot of
3:29 data and a lot of telemetry and to use it for sports science uh and training
3:34 purposes. Certainly, you know, I I had many dashboards on my own performance
3:38 before I had dashboards on private company performance. >> Love it. Love it. So, Jason, let's let's
3:44 jump in and I think the best place to start is maybe just summarizing the
3:50 broad themes of what I'd call the AI market in 2025. Kind what what happened?
3:57 Walk us through the big themes. So, I mean, I think you can't talk about
4:01 the AI market in 2025 without talking about the fact that it was a record
4:04 setting year for private company equity funding, right? We almost uh more than
4:12 doubled the previous record from 2021 uh with 225 or 226 billion worth of
4:20 equity funding into AI startups. Uh when you start to go and look beyond that,
4:24 right, you know, there's sort of a deeper underlying current which is that you
4:31 know we had a lot of that funding that went to the largest LLM developers right
4:37 the open AIS anthropics XIS of the world right so those three companies alone
4:44 accounted for 41% of all AI funding in 2025 right so yes you could pull them
4:48 out it would have still been a record year uh but at that point it would have
4:53 been just barely a record year and I think that's generally the story in not
4:59 just AI but venture more broadly from 2025 which is that we saw this
5:04 incredible either record or rebound depending on how you want which of the
5:10 subsets you wanted to look at and most of that was driven by mega deals or uber
5:17 hyper mega deals right out into the billion dollar billions of dollar deals
5:23 that we hadn't really seen before and really dictated the sort of venture
5:29 landscape. So you saw this ecosystem where you got you know I guess towards up until the midpoint of
5:39 2025 in the AI uh sector right you know we were still seeing record deal volume
5:43 but then even that started to taper off uh in the back half of the year but
5:49 funding dollar amount stayed high and so we're getting to this sort of bifurcated
5:56 or dual track market that we've seen seen in the venture ecosystem more
6:00 broadly as well. And a lot of what that tells us is that you know people are
6:05 placing really big bets on a select few companies that have either justified
6:10 those bets or justified the promise of those bets and the rest are you know now
6:17 facing stricter financial pressure uh and sort of back to the fundamentals of
6:21 what a good company is. We're no longer in what I'm going to call the early AI
6:29 genai uh craze of late 2022, early 2023 where if you slappedai after your
6:34 company name and made a website, uh pitch decks were flying or term sheets
6:38 were flying in left, right, and center. >> Right. Right. You you mentioned the big
6:43 players and I think everyone's familiar with the three that you mentioned,
6:48 OpenAI, Anthropic, and XAI, but th this is a this is a multi-tiered market in a
6:53 lot of ways. And and so let's maybe just give people a landscape of when we talk
6:59 about other than those big three, what are the what are the layers of the Genai
7:04 marketplace right now that are driving maybe some of those other investments,
7:08 opportunities, deals? So I mean I think the places that we're
7:13 seeing the most uh activity now is sort of in what I'll call the infrastructure
7:17 layer right so the things that allow those models to operate uh the things
7:23 that will change the future of things like semiconductors chips uh energy
7:32 supply sort of the and then the next two layers become the agentic infrastructure and the
7:39 agentic workspace or work surface, right? That now has this massive promise
7:44 to disrupt the sort of global labor spend as opposed to just software and
7:48 technology spend, right? You know, if I can have an agent that goes and
7:52 I don't know, crawls through all of our data, which we have many of these, but
7:56 goes through and crawls through all of our data and tries to understand, hey,
8:00 if I want to pit these two companies against each other and compare them like
8:05 a financial analyst would or a venture associate would, right? I can now go
8:09 build an agent to do this, right? And we've seen one of the other major things
8:14 from 2025 that will continue through this year is this time or even just 11
8:23 months ago. Uh in 2025, we had just under 400 AI AI companies, right? That's
8:28 not to say that there weren't many more agents, but those were the companies
8:33 specifically working on AI agents both at sort of the infrastructure and
8:38 orchestration layer and then sort of at specific uh vertical tasks uh for those
8:44 agents. So right AI agents for finance, AI agents for healthcare, AI agents for
8:51 marketing and sales. Now we have uh approaching 2100 companies after right
8:56 11 months. So >> wow is right in terms of what's hot agents are still hot. The applications
9:04 for those agents across verticals are very hot and again it's largely based on
9:10 the promise that these agents can either change the way that work is done or
9:14 disintermediate a lot of the labor spend. Right? when we look at I'll take
9:21 the customer service uh space as an example right the promise
9:26 forever of AI for customer service is that you could reduce the number of
9:30 people required to get to the requisite number of resolutions for the number of
9:36 tickets that you had coming in for a long time uh that was sort of this pie
9:43 in the sky thing of you could give an agent a script and you could Right.
9:47 We've all interacted with a bad little chatbot on the bottom of a website.
9:49 >> Y >> they're very very easy to break and eventually you would either need more
9:57 humans uh to interact or you would end up with very frustrated customers.
10:02 >> Now we are at the point where you start to see that customer service hiring has
10:08 and customer service uh employment has dramatically fallen in the last 12 to 18
10:15 months. And you see that that sort of falls in line with the revenue growth of
10:19 the companies working on customer service uh AI agents and systems. So
10:25 >> that's where the promise of agents lies and it's probably where for the next you
10:30 know until we get to the next breakthrough of what's after agents. Uh
10:35 you know it's where we'll see a lot of the venture focus and a lot of the AI
10:39 development in terms of what's actually getting used and where the commercial
10:44 traction and impact are. Yeah, 2100. That's incredible. And it speaks
10:48 to the competitive nature and the opportunity that I think much like the
10:53 dot boom that we saw, you get a few big players that start to set the pace and
10:58 then a number of companies that try to build off the back of that. And I want
11:01 to I want to talk about that. We were talking about this before the show
11:06 started. the the AI race, the race to gen AI, which is part of this
11:10 conversation, which is being dominated really by a few of the big LLM
11:16 developers. How does that play into the big bet piece versus this ecosystem of
11:21 small companies that are just kind of nipping at the edges of this emerging
11:26 market? Is that is that something that is driving those big bets of whoever
11:30 wins this open AAI anthropic X LLM off is going to be the winner?
11:37 Big Tech and AI Investments
11:40 I mean I think it's an interesting space to look at really the capex and the
11:47 expenditure required to build these models and these systems, right?
11:50 Everything from the energy required to the data centers required to the chips
11:56 required to get to a sufficient amount of compute to right you know we've all
12:01 heard the sound bites and the snippets of you know the the limit on what an AI
12:09 model can do is largely based on what I'm going to shorthand as compute and
12:16 data and energy and those are the sort of the three things that are some of the
12:21 most costly, right? You look at even stuff from last year on the data side
12:28 and meta paid up uh you know [snorts] well into the mid- teens billions. I'm
12:32 not going to remember the exact I can't remember whether it's 14.8 or 15.8 or
12:37 16.8 off the top of my head for scale uh and the team behind scale to really
12:41 ensure that they had the cleanest data advantage uh when it came to training
12:48 data and data annotation. And we've seen multiple deals in that space. And so you
12:53 think about what the money raised by and the capital raised by these large LLM providers goes
13:04 to and it's scale at the you know sort of size and scope that you would see
13:11 from the likes of Google or Meta or Amazon or Microsoft or Apple and or
13:17 Nvidia, right? And I'll rattle those off as the clear other players in, you know,
13:24 sort of the self-funded, so to speak, right, to what we were talking about
13:28 before we got on here, self-funded space. U they, these are trillion dollar
13:34 market cap companies with enough cash on hand to go and spend on all of this
13:38 infrastructure that's needed to build these systems. And now we're
13:46 seeing that because this innovation has happened largely from private companies
13:51 uh or a lot of this innovation is happening from private companies. The
13:55 companies that are rapidly attempting to operate at the scale of you know
14:00 trillion dollar plus public companies need to have as much cash on hand to
14:04 spend on the necessary components of that. Yeah, there there's going to be
14:09 the number of Harvard business case studies that are going to come out of
14:12 this market dynamic I think are going to be staggering. And let's talk about
14:19 those few we talked about Microsoft Meta some of the self-funded uh co-pilot if
14:25 you use Office 365 is embedded now. I think there's been some ups and downs
14:29 with that roll out that they've experienced. I know Meta's had their own
14:34 experience. How much of that just maybe from your perspective of I find it
14:39 fascinating that I said this years ago that you know Facebook before they
14:43 became Meta just they were they just became a big company like there's a life
14:47 cycle that happens to firms that grow up and become big companies. It's a very
14:52 normal natural thing. How much is that an enabler because it gives them the
14:58 cash to self-fund but also maybe a limiter because we're seeing the players
15:02 that are moving faster aren't that. I guess maybe X is a little bit of the
15:06 exception just given the transaction that happened there and there's there's
15:10 a lot of perspective on that but >> but I mean even then you get to X and
15:15 you get to look at it in terms of the fact that SpaceX now bought XAI and XAI
15:23 had bought bought X uh and all of that still sits within private markets and so
15:28 there's more value still sat within private markets than there ever has been
15:34 before. um to get to the what does this mean for the large you know sort of me
15:43 mega cap uh players right I think we'll see that you know we we've seen a
15:49 life cycle here even in the last 12 months 18 months 24 months of right
15:55 forever or for a while as when open AI and anthropic were the only two really in the ring
16:03 Uh, and now you go and you look at Google Gemini and I'm not someone will
16:08 tell me that they're not all interchangeable. I use different models
16:13 for different tasks. And if I want to go do image creation, I use nano banana. If
16:18 I want to go write thing, you know, scripts to run through a bunch of data,
16:21 I'll use cloud code. And if I want to write, I'll use chat GPT. And so you see
16:28 specialization amongst these different players and these different use cases.
16:33 And you'll see largely that as these models I I always struggle to dance around the
16:41 word commoditize uh but as they commoditize you know you'll see where the sort of advantages
16:52 of these incumbents being distribution embedding into existing processes and
16:58 the reality that a lot of the especially a lot of the business world is
17:01 relatively slow moving with long contracts and PE and change management
17:06 is hard that you know the people who want to use office the office suite of
17:13 products and use copilot within that office suite of products will find the
17:19 benefit as co-pilot gets better or there are more clawed plugins to excel or
17:27 I can tie in chat GPT directly to my PowerPoint right you We're in the right early days, early
17:37 innings to use the cliche phrase that >> uh reporters keep asking me if I can put
17:42 a number to what that means and I can't. Uh because we live in unprecedented
17:49 times right now. The reality of this is all of these tools and all of these
17:53 models will go somewhere and really where they'll go is where people are
17:57 doing their existing work for the next little bit. And then consumer or proumer
18:06 or habits within a work environment take longer to change than right like pe
18:12 people are the reason that AI adoption will be slower. Right. It's not that the
18:16 tools aren't aren't there and aren't useful. >> Yeah. Yeah. I I spend many of my hours a
18:22 day on this topic with clients going through that conversation itself.
18:27 Let I you said something and you you you you talked about the structure of the
18:31 market and I I want to maybe do a quick comparison. What what makes this moment
18:38 maybe different or the same from what we saw in the dot boom and then and then we
18:46 saw the bubble that came later and you can't predict the future but what does
18:50 the data tell us about you know is this a likely occurrence? Are we going to are
18:55 we following the same arc? Are we just repeating history with a different
18:59 technology? >> I think I think my answer to that has been
19:06 the right at least on the venture side and right you know on on the what is the
19:11 promise of value creation and returns generated from all of the money that is flooded
19:21 into the AI startup ecosystem. Venture has always been a power loss
19:25 game. It doesn't right you you you it takes right if we look at last year and
19:30 we look at the you know 226 million or billion dollars that went into AI funding last year.
19:39 That takes anthropic going public at 500 billion or open AI going public at a
19:49 trillion or right I'm making these numbers. >> Yeah. uh because by the time we get to
19:54 either of those companies going public, they'll have raised at yet another
20:00 asinine valuation. Uh the [snorts] right that returns all of the funding
20:07 that has gone into >> um these geni companies. >> Yeah.
20:13 >> AI startups more broadly. And so, right, power laws still hold even in a bubble.
20:17 Maybe power laws especially hold in a bubble, right? You can go back and you
20:21 can look at the cloud era companies and the companies that came out of that. You
20:24 can look at the dotcom era companies and say generally the same thing about all
20:29 of these technology cycles. I think the thing that is most unique about this cycle is the
20:36 speed at which it's happened and then sort of the promise right the speed at
20:40 which it's happened the distribution of the technology right you know so I
20:46 don't know off the top of my head the daily active users for chat GPT at this
20:50 point but you have reached far more people using this technology even at a
20:56 consumer level than you had for pretty much any other technology
21:02 in a decade, in the course of three years. And we're at the point now where
21:10 largely what you'll start to see is it's the speed of adoption of this technology
21:16 that's really going to dictate what's different and why people I mean it's
21:21 change is hard. I right if I have to know which model to use and that's a
21:25 different model every week, right? that creates a little bit more frenzy around
21:30 the technology and the technology adoption curve than you had for cloud or
21:36 the dot era. >> Yeah. If only somebody had written a book about change where you could figure
21:40 out what this >> Yeah, I might see it in the background.
21:45 >> Shameless that was a shameless plug, but it's >> this this whole episode's just been a
21:49 tea up to >> I appreciate that. I I want to ask about the other part of this macro and then
21:56 maybe get down to the micro which is if I look at the the.com boom and the rise
22:03 of of of Google the rise of Meta the rise of Microsoft you know Microsoft
22:09 strategy obviously they had the market really cornered on on
22:14 you know infrastructure processing for for the corporate world but when that
22:18 slowed down you know they reached out they bought all sorts of stuff they
22:21 bought link LinkedIn, they bought all of these different platforms. And you look
22:25 at Meta, they bought WhatsApp, they bought Instagram, and all of these
22:28 different things. And I I guess I can't help but wonder, is it going, you know,
22:34 is are the big bets maybe just as much about the core technology that each of
22:39 the big three are investing, but is but is it also a big bet that they're just
22:44 going to scoop up these individual use cases that start to fight it out in the
22:49 market over time? Are you asking from the large hyperscaler Yeah. perspective?
22:54 >> Yeah, >> it's unclear. I mean, I think for the
22:59 first time you see these companies acting scared in a sense of right the open AIS and the anthropics
23:10 of the world are currently leading on right sort of disintermediating a lot of
23:13 those workflows that they that these companies have previously owned. And
23:18 then you also get the promise of well, if we still own the distribution, if I
23:23 if I'm still Google and I own YouTube and now video creation goes through the and content
23:28 creation goes through the roof, I still win by selling ads. If I'm still
23:35 Microsoft and everyone's, you know, we're still going to get a generation of
23:38 people that use cloud inside Excel, well, they all still need Excel. uh if
23:44 right and the beauty for Amazon or AWS, Azure and GCP is that
23:50 they all still sell the cloud infrastructure that you know is making
23:55 money handover fist for them as everyone needs more cloud uh infrastructure to
24:03 use these AI tools and models. And so I mean right in the handwavy version of
24:09 I think they'll all be fine. Uh, I think they'll all be fine. It just may look
24:12 like a different set of Yeah. >> where they sit in the value stack.
24:17 >> Well, part of what I'm I'm also thinking about and we're we're coming up to the
24:21 we could keep going on this. This is like a this is like an MBA class in
24:26 market dynamics is, you know, you mentioned 2100 companies and if we
24:32 >> Well, and that's just for a right, you know, companies specifically working on
24:34 AI agents, >> right? And so, yeah. And so those all won't exist by the time this all shakes
24:42 out. And so I think the question is if I'm a if I'm, you know, if I'm a VC that
24:48 it's it's a different calculus than if I'm just a person in a company or I'm an
24:53 investor, what do I experiment with? What do I get attached to? How much is
24:57 going to change? How do I think about the world of work as a user of these
25:04 systems? and then also what it means for the work that I do. And you brought up
25:10 the productivity curves and the customer service experience and the layoffs that
25:14 we've already sort of seen. And you look at the 16,000 that Amazon cut just a
25:18 month and a half ago or whatever it is. You know, as the average person, all the
25:23 data is great, but as the average person, I'm I if I look up, there's
25:28 disruption. If I look left and right, there's disruption. If I look down,
25:30 Navigating the AI Landscape
25:32 there's disruption. How do I How do I navigate all of this given that we're
25:36 just regular people trying to kind of figure out the world here? [snorts]
25:41 >> I mean, I I'll I don't know that I have data that tells you how to navigate
25:44 this. Although, I'll I'll give you a pretty dash, right? We have dashboards
25:48 and we have agents and we have tools to help, right? The lay person, right? And
25:54 especially people in business contexts, right? We work directly with a lot of,
25:58 as you said, you know us from your time in professional services. We work
26:02 directly with a lot of those firms. We work directly with a lot of corp dev and
26:06 business development leaders to answer sort of exactly that question on the
26:11 business side. But as a normal day-to-day user of AI tools, right, and
26:18 someone who's curious about how this can impact your life in the work that you do or where
26:25 there's disruption potentially for the work that you do, right? I think the
26:28 answer is really just to go and use these tools and to be sort of an amateur and a student of what
26:37 can I do? How could I possibly reimagine the work I do on a daily basis? Right? I
26:42 think about right Excel has come up a lot. I think about all the work I used
26:48 to do in Excel and I still will do some stuff in Excel, but I'll do so much more
26:55 of it in cloud code running off of files that are accessed uh and let it run, you
27:02 know, more intelligent scripts and prompts over the top of that
27:08 data than I would in a few Excel shortcuts uh or formulas. And it really
27:13 allows you to rethink the way that you go through a lot of these processes and
27:18 sometimes it'll open up right for me it opens up new types of analyses that I
27:22 hadn't quite considered yet or it opens up new ways to look at this data or
27:25 relationships between the data right I would imagine that if you're a
27:31 saleserson uh right it gives you far more access to knowledge on any given
27:36 account or prospect than you've ever had before right you can think of all of
27:41 these ways in which it changes either speeds up the work you were doing or
27:45 changes the work that you can do. Yeah. Trying to think of it in those two
27:49 domains. And just >> I am I am doing I I am I'm dealing with
27:54 this literally as we speak >> the the process I use to manage this
27:59 podcast and the bookings and conversations and who I'm doing. Uh I'm
28:03 going to move into using Claude to navigate those two things. And there's a
28:07 whole sequence of events that happens when we book a date. And it's it's it
28:11 it's exactly that. And I think what I'm learning is maybe Claude's not a great
28:16 example because I know that Anthropic is going to survive in some way, but don't
28:20 also get too attached because it's probably likely that there'll be two or
28:25 three iterations of a new competitor, a new tool, a new use case. Uh you know,
28:29 even what Claude's capable of doing, right? I mean, you look at I think about
28:34 the things that I could do with pick your LLM of choice, right? pick [snorts]
28:39 your model of choice 12 months ago and I think about what I can do today and
28:45 right for on the company side right you almost have and on the startup side you
28:49 have to think about as a founder or as an as an investor what is more than one
28:56 or two or three model releases uh away from being obsoleted or cannibalized
29:03 >> and but as a user we get to as like a a daily driver we get to sort of be
29:07 dragged along with the pace of what these things can do and that's massively
29:13 exciting and a little bit terrifying. >> Y those two those those two experiences
29:18 often go hand in hand. I want to talk about and kind of give us a chance to
29:22 think about looking ahead and we've talked about a lot here. What what is
29:23 Looking Ahead to 2026
29:26 the you know the magic eightball if we shook it up and we look at 2026. Can we
29:32 glean anything that we expect to happen in the next 12 months based on what's
29:41 I think based on again I'll go towards the infrastructure sort of the spend on
29:47 infrastructure the massive potentially final private market raises from the
29:52 large uh model providers and then the sort of agents agents agents world that we're
30:01 now really firmly in. We're you know we're not just entering we're not just
30:04 dipping a toe into the agentic world. We're here and all of the things that
30:08 need to go alongside that. uh on the infrastructure side I think we'll get
30:12 you know more and more spend as people try to either figure out the cheaper you
30:18 know sort of the cheaper or more climate friendly ways to meet the energy and
30:26 compute demands uh on the large model side uh we probably see at least one
30:35 more round uh from both uh anthropic and open AAI before they go public but Both
30:39 of them are rumored uh and have made plenty of noise about going public uh
30:44 you know as early as later this year. And on the agentic side, right, I think
30:53 everyone now knows or almost everyone now knows what an agent is, what the
30:58 promise of having something that runs autonomously is. And now right a lot of what we saw
31:06 in the back half of last year and the you know the new companies and greater
31:12 funding to that spaces and what we've even seen in the first few months of
31:16 this year is that you know there's a massive demand for sort of the
31:21 orchestration and infrastructure and sort of the gu the guard rails around
31:26 how you run agents through a business or your own life. And so thinking about you
31:32 know I was looking I spend a little bit of time every day looking at the most
31:37 recent deals and there were a few deals last week uh where you saw actually
31:42 right we saw an investment from Anthropic in a company called Sapium
31:47 that is financial infrastructure for AI agents. So you think about all of the
31:51 different components of the more decisions that an agent can make, the
31:56 more places that it can operate and the more you need to guard rail it and
32:01 provided instructions and understand what it's doing and then you have multi-
32:05 aent systems and the orchestration between those agents and you know I
32:09 think we're still in the early days of what that means. We're going to see a
32:14 lot more funding to companies that operate in those domains. Sort of
32:18 enabling everyone from you and me to the world's largest enterprises to sort of build own
32:30 operate agents effectively and have some amount of measurement uh over the top of
32:35 them uh because you can't improve what you can't measure, >> right? And
32:41 then I mean I think you know largely we'll see that you'll get to the point where
32:48 there are now going to be some early winners in vertical specifics especially
32:53 in regulated industries right when you think about finance when you think about
32:56 healthcare when you think about legal and some of the others that either have
33:02 barriers to entry on uh regulation or barriers to entry on sort of domain knowledge. and
33:09 specificity. Uh you'll we'll start to see like the true leaders emerge in a
33:14 lot of these verticals. Uh or at least the true first leaders emerge in a lot
33:18 of these verticals. >> Wow. Uh that's that's that's a lot. You
33:22 that you know that used to be 10 years of business performance packed into 12
33:26 months. And I I think all of those things are very likely to to happen. And
33:31 I think you used a word in there a couple times about orchestration. And I
33:36 think we we think so much about the the job loss that's coming has come and will
33:41 continue to be impacted. But you also described a an emerging set of
33:47 industries and professions and careers and jobs that are about orchestration
33:52 and synchronizing and supporting people through these journeys. And I think it's
33:56 just another good lesson that when when one market dies, another one finds a way
34:01 to slowly emerge. and you see the investments starting to make bets on
34:05 where they think those winners will be and and to your point, it's a exciting
34:09 Conclusion and Final Thoughts
3:40 AI Market Overview for 2025
3:44 jump in and I think the best place to start is maybe just summarizing the
3:50 broad themes of what I'd call the AI market in 2025. Kind what what happened?
3:57 Funding Trends and Major Players in AI
3:57 Walk us through the big themes. So, I mean, I think you can't talk about
4:01 the AI market in 2025 without talking about the fact that it was a record
4:04 setting year for private company equity funding, right? We almost uh more than
4:12 doubled the previous record from 2021 uh with 225 or 226 billion worth of
4:20 equity funding into AI startups. Uh when you start to go and look beyond that,
4:24 right, you know, there's sort of a deeper underlying current which is that you
4:31 know we had a lot of that funding that went to the largest LLM developers right
4:37 the open AIS anthropics XIS of the world right so those three companies alone
4:44 accounted for 41% of all AI funding in 2025 right so yes you could pull them
4:48 out it would have still been a record year uh but at that point it would have
4:53 been just barely a record year and I think that's generally the story in not
4:59 just AI but venture more broadly from 2025 which is that we saw this
5:04 incredible either record or rebound depending on how you want which of the
5:10 subsets you wanted to look at and most of that was driven by mega deals or uber
5:17 hyper mega deals right out into the billion dollar billions of dollar deals
5:23 that we hadn't really seen before and really dictated the sort of venture
5:29 landscape. So you saw this ecosystem where you got you know I guess towards up until the midpoint of
5:39 2025 in the AI uh sector right you know we were still seeing record deal volume
5:43 but then even that started to taper off uh in the back half of the year but
5:49 funding dollar amount stayed high and so we're getting to this sort of bifurcated
5:56 or dual track market that we've seen seen in the venture ecosystem more
6:00 broadly as well. And a lot of what that tells us is that you know people are
6:05 placing really big bets on a select few companies that have either justified
6:10 those bets or justified the promise of those bets and the rest are you know now
6:17 facing stricter financial pressure uh and sort of back to the fundamentals of
6:21 what a good company is. We're no longer in what I'm going to call the early AI
6:29 genai uh craze of late 2022, early 2023 where if you slappedai after your
6:34 company name and made a website, uh pitch decks were flying or term sheets
6:38 were flying in left, right, and center. >> Right. Right. You you mentioned the big
6:43 players and I think everyone's familiar with the three that you mentioned,
6:47 The Rise of AI Agents
6:48 OpenAI, Anthropic, and XAI, but th this is a this is a multi-tiered market in a
6:53 lot of ways. And and so let's maybe just give people a landscape of when we talk
6:59 about other than those big three, what are the what are the layers of the Genai
7:04 marketplace right now that are driving maybe some of those other investments,
7:08 opportunities, deals? So I mean I think the places that we're
7:13 seeing the most uh activity now is sort of in what I'll call the infrastructure
7:17 layer right so the things that allow those models to operate uh the things
7:23 that will change the future of things like semiconductors chips uh energy
7:32 supply sort of the and then the next two layers become the agentic infrastructure and the
7:39 agentic workspace or work surface, right? That now has this massive promise
7:44 to disrupt the sort of global labor spend as opposed to just software and
7:48 technology spend, right? You know, if I can have an agent that goes and
7:52 I don't know, crawls through all of our data, which we have many of these, but
7:56 goes through and crawls through all of our data and tries to understand, hey,
8:00 if I want to pit these two companies against each other and compare them like
8:05 a financial analyst would or a venture associate would, right? I can now go
8:09 build an agent to do this, right? And we've seen one of the other major things
8:14 from 2025 that will continue through this year is this time or even just 11
8:23 months ago. Uh in 2025, we had just under 400 AI AI companies, right? That's
8:28 not to say that there weren't many more agents, but those were the companies
8:33 specifically working on AI agents both at sort of the infrastructure and
8:38 orchestration layer and then sort of at specific uh vertical tasks uh for those
8:44 agents. So right AI agents for finance, AI agents for healthcare, AI agents for
8:51 marketing and sales. Now we have uh approaching 2100 companies after right
8:56 11 months. So >> wow is right in terms of what's hot agents are still hot. The applications
9:04 for those agents across verticals are very hot and again it's largely based on
9:10 the promise that these agents can either change the way that work is done or
9:14 disintermediate a lot of the labor spend. Right? when we look at I'll take
9:15 Impact of AI on Customer Service
9:21 the customer service uh space as an example right the promise
9:26 forever of AI for customer service is that you could reduce the number of
9:30 people required to get to the requisite number of resolutions for the number of
9:36 tickets that you had coming in for a long time uh that was sort of this pie
9:43 in the sky thing of you could give an agent a script and you could Right.
9:47 We've all interacted with a bad little chatbot on the bottom of a website.
9:49 >> Y >> they're very very easy to break and eventually you would either need more
9:57 humans uh to interact or you would end up with very frustrated customers.
10:02 >> Now we are at the point where you start to see that customer service hiring has
10:08 and customer service uh employment has dramatically fallen in the last 12 to 18
10:15 months. And you see that that sort of falls in line with the revenue growth of
10:19 the companies working on customer service uh AI agents and systems. So
10:25 >> that's where the promise of agents lies and it's probably where for the next you
10:30 know until we get to the next breakthrough of what's after agents. Uh
10:35 you know it's where we'll see a lot of the venture focus and a lot of the AI
10:39 development in terms of what's actually getting used and where the commercial
10:44 traction and impact are. Yeah, 2100. That's incredible. And it speaks
10:48 to the competitive nature and the opportunity that I think much like the
10:53 dot boom that we saw, you get a few big players that start to set the pace and
10:58 then a number of companies that try to build off the back of that. And I want
11:01 to I want to talk about that. We were talking about this before the show
11:06 started. the the AI race, the race to gen AI, which is part of this
11:10 conversation, which is being dominated really by a few of the big LLM
11:16 developers. How does that play into the big bet piece versus this ecosystem of
11:21 small companies that are just kind of nipping at the edges of this emerging
11:26 market? Is that is that something that is driving those big bets of whoever
11:30 wins this open AAI anthropic X LLM off is going to be the winner?
11:40 I mean I think it's an interesting space to look at really the capex and the
11:47 expenditure required to build these models and these systems, right?
11:50 Everything from the energy required to the data centers required to the chips
11:56 required to get to a sufficient amount of compute to right you know we've all
12:01 heard the sound bites and the snippets of you know the the limit on what an AI
12:09 model can do is largely based on what I'm going to shorthand as compute and
12:16 data and energy and those are the sort of the three things that are some of the
12:21 most costly, right? You look at even stuff from last year on the data side
12:28 and meta paid up uh you know [snorts] well into the mid- teens billions. I'm
12:32 not going to remember the exact I can't remember whether it's 14.8 or 15.8 or
12:37 16.8 off the top of my head for scale uh and the team behind scale to really
12:41 ensure that they had the cleanest data advantage uh when it came to training
12:48 data and data annotation. And we've seen multiple deals in that space. And so you
12:53 think about what the money raised by and the capital raised by these large LLM providers goes
13:04 to and it's scale at the you know sort of size and scope that you would see
13:11 from the likes of Google or Meta or Amazon or Microsoft or Apple and or
13:17 Nvidia, right? And I'll rattle those off as the clear other players in, you know,
13:24 sort of the self-funded, so to speak, right, to what we were talking about
13:28 before we got on here, self-funded space. U they, these are trillion dollar
13:34 market cap companies with enough cash on hand to go and spend on all of this
13:38 infrastructure that's needed to build these systems. And now we're
13:46 seeing that because this innovation has happened largely from private companies
13:51 uh or a lot of this innovation is happening from private companies. The
13:55 companies that are rapidly attempting to operate at the scale of you know
14:00 trillion dollar plus public companies need to have as much cash on hand to
14:04 spend on the necessary components of that. Yeah, there there's going to be
14:09 the number of Harvard business case studies that are going to come out of
14:12 this market dynamic I think are going to be staggering. And let's talk about
14:19 those few we talked about Microsoft Meta some of the self-funded uh co-pilot if
14:25 you use Office 365 is embedded now. I think there's been some ups and downs
14:29 with that roll out that they've experienced. I know Meta's had their own
14:34 experience. How much of that just maybe from your perspective of I find it
14:39 fascinating that I said this years ago that you know Facebook before they
14:43 became Meta just they were they just became a big company like there's a life
14:47 cycle that happens to firms that grow up and become big companies. It's a very
14:52 normal natural thing. How much is that an enabler because it gives them the
14:58 cash to self-fund but also maybe a limiter because we're seeing the players
15:02 that are moving faster aren't that. I guess maybe X is a little bit of the
15:06 exception just given the transaction that happened there and there's there's
15:10 a lot of perspective on that but >> but I mean even then you get to X and
15:15 you get to look at it in terms of the fact that SpaceX now bought XAI and XAI
15:23 had bought bought X uh and all of that still sits within private markets and so
15:28 there's more value still sat within private markets than there ever has been
15:34 before. um to get to the what does this mean for the large you know sort of me
15:43 mega cap uh players right I think we'll see that you know we we've seen a
15:49 life cycle here even in the last 12 months 18 months 24 months of right
15:55 forever or for a while as when open AI and anthropic were the only two really in the ring
16:03 Uh, and now you go and you look at Google Gemini and I'm not someone will
16:08 tell me that they're not all interchangeable. I use different models
16:13 for different tasks. And if I want to go do image creation, I use nano banana. If
16:18 I want to go write thing, you know, scripts to run through a bunch of data,
16:21 I'll use cloud code. And if I want to write, I'll use chat GPT. And so you see
16:28 specialization amongst these different players and these different use cases.
16:33 And you'll see largely that as these models I I always struggle to dance around the
16:41 word commoditize uh but as they commoditize you know you'll see where the sort of advantages
16:52 of these incumbents being distribution embedding into existing processes and
16:58 the reality that a lot of the especially a lot of the business world is
17:01 relatively slow moving with long contracts and PE and change management
17:06 is hard that you know the people who want to use office the office suite of
17:13 products and use copilot within that office suite of products will find the
17:19 benefit as co-pilot gets better or there are more clawed plugins to excel or
17:27 I can tie in chat GPT directly to my PowerPoint right you We're in the right early days, early
17:37 innings to use the cliche phrase that >> uh reporters keep asking me if I can put
17:42 a number to what that means and I can't. Uh because we live in unprecedented
17:49 times right now. The reality of this is all of these tools and all of these
17:53 models will go somewhere and really where they'll go is where people are
17:57 doing their existing work for the next little bit. And then consumer or proumer
18:06 or habits within a work environment take longer to change than right like pe
18:12 people are the reason that AI adoption will be slower. Right. It's not that the
18:16 tools aren't aren't there and aren't useful. >> Yeah. Yeah. I I spend many of my hours a
18:22 day on this topic with clients going through that conversation itself.
18:27 Let I you said something and you you you you talked about the structure of the
18:31 market and I I want to maybe do a quick comparison. What what makes this moment
18:38 maybe different or the same from what we saw in the dot boom and then and then we
18:46 saw the bubble that came later and you can't predict the future but what does
18:50 the data tell us about you know is this a likely occurrence? Are we going to are
18:55 we following the same arc? Are we just repeating history with a different
18:59 technology? >> I think I think my answer to that has been
19:06 the right at least on the venture side and right you know on on the what is the
19:11 promise of value creation and returns generated from all of the money that is flooded
19:21 into the AI startup ecosystem. Venture has always been a power loss
19:25 game. It doesn't right you you you it takes right if we look at last year and
19:30 we look at the you know 226 million or billion dollars that went into AI funding last year.
19:39 That takes anthropic going public at 500 billion or open AI going public at a
19:49 trillion or right I'm making these numbers. >> Yeah. uh because by the time we get to
19:54 either of those companies going public, they'll have raised at yet another
20:00 asinine valuation. Uh the [snorts] right that returns all of the funding
20:07 that has gone into >> um these geni companies. >> Yeah.
20:13 >> AI startups more broadly. And so, right, power laws still hold even in a bubble.
20:17 Maybe power laws especially hold in a bubble, right? You can go back and you
20:21 can look at the cloud era companies and the companies that came out of that. You
20:24 can look at the dotcom era companies and say generally the same thing about all
20:29 of these technology cycles. I think the thing that is most unique about this cycle is the
20:36 speed at which it's happened and then sort of the promise right the speed at
20:40 which it's happened the distribution of the technology right you know so I
20:46 don't know off the top of my head the daily active users for chat GPT at this
20:50 point but you have reached far more people using this technology even at a
20:56 consumer level than you had for pretty much any other technology
21:02 in a decade, in the course of three years. And we're at the point now where
21:10 largely what you'll start to see is it's the speed of adoption of this technology
21:16 that's really going to dictate what's different and why people I mean it's
21:21 change is hard. I right if I have to know which model to use and that's a
21:25 different model every week, right? that creates a little bit more frenzy around
21:30 the technology and the technology adoption curve than you had for cloud or
21:36 the dot era. >> Yeah. If only somebody had written a book about change where you could figure
21:40 out what this >> Yeah, I might see it in the background.
21:45 >> Shameless that was a shameless plug, but it's >> this this whole episode's just been a
21:49 tea up to >> I appreciate that. I I want to ask about the other part of this macro and then
21:56 maybe get down to the micro which is if I look at the the.com boom and the rise
22:03 of of of Google the rise of Meta the rise of Microsoft you know Microsoft
22:09 strategy obviously they had the market really cornered on on
22:14 you know infrastructure processing for for the corporate world but when that
22:18 slowed down you know they reached out they bought all sorts of stuff they
22:21 bought link LinkedIn, they bought all of these different platforms. And you look
22:25 at Meta, they bought WhatsApp, they bought Instagram, and all of these
22:28 different things. And I I guess I can't help but wonder, is it going, you know,
22:34 is are the big bets maybe just as much about the core technology that each of
22:39 the big three are investing, but is but is it also a big bet that they're just
22:44 going to scoop up these individual use cases that start to fight it out in the
22:49 market over time? Are you asking from the large hyperscaler Yeah. perspective?
22:54 >> Yeah, >> it's unclear. I mean, I think for the
22:59 first time you see these companies acting scared in a sense of right the open AIS and the anthropics
23:10 of the world are currently leading on right sort of disintermediating a lot of
23:13 those workflows that they that these companies have previously owned. And
23:18 then you also get the promise of well, if we still own the distribution, if I
23:23 if I'm still Google and I own YouTube and now video creation goes through the and content
23:28 creation goes through the roof, I still win by selling ads. If I'm still
23:35 Microsoft and everyone's, you know, we're still going to get a generation of
23:38 people that use cloud inside Excel, well, they all still need Excel. uh if
23:44 right and the beauty for Amazon or AWS, Azure and GCP is that
23:50 they all still sell the cloud infrastructure that you know is making
23:55 money handover fist for them as everyone needs more cloud uh infrastructure to
24:03 use these AI tools and models. And so I mean right in the handwavy version of
24:09 I think they'll all be fine. Uh, I think they'll all be fine. It just may look
24:12 like a different set of Yeah. >> where they sit in the value stack.
24:17 >> Well, part of what I'm I'm also thinking about and we're we're coming up to the
24:21 we could keep going on this. This is like a this is like an MBA class in
24:26 market dynamics is, you know, you mentioned 2100 companies and if we
24:32 >> Well, and that's just for a right, you know, companies specifically working on
24:34 AI agents, >> right? And so, yeah. And so those all won't exist by the time this all shakes
24:42 out. And so I think the question is if I'm a if I'm, you know, if I'm a VC that
24:48 it's it's a different calculus than if I'm just a person in a company or I'm an
24:53 investor, what do I experiment with? What do I get attached to? How much is
24:57 going to change? How do I think about the world of work as a user of these
25:04 systems? and then also what it means for the work that I do. And you brought up
25:10 the productivity curves and the customer service experience and the layoffs that
25:14 we've already sort of seen. And you look at the 16,000 that Amazon cut just a
25:18 month and a half ago or whatever it is. You know, as the average person, all the
25:23 data is great, but as the average person, I'm I if I look up, there's
25:28 disruption. If I look left and right, there's disruption. If I look down,
25:32 there's disruption. How do I How do I navigate all of this given that we're
25:36 just regular people trying to kind of figure out the world here? [snorts]
25:41 >> I mean, I I'll I don't know that I have data that tells you how to navigate
25:44 this. Although, I'll I'll give you a pretty dash, right? We have dashboards
25:48 and we have agents and we have tools to help, right? The lay person, right? And
25:54 especially people in business contexts, right? We work directly with a lot of,
25:58 as you said, you know us from your time in professional services. We work
26:02 directly with a lot of those firms. We work directly with a lot of corp dev and
26:06 business development leaders to answer sort of exactly that question on the
26:11 business side. But as a normal day-to-day user of AI tools, right, and
26:18 someone who's curious about how this can impact your life in the work that you do or where
26:25 there's disruption potentially for the work that you do, right? I think the
26:28 answer is really just to go and use these tools and to be sort of an amateur and a student of what
26:37 can I do? How could I possibly reimagine the work I do on a daily basis? Right? I
26:42 think about right Excel has come up a lot. I think about all the work I used
26:48 to do in Excel and I still will do some stuff in Excel, but I'll do so much more
26:55 of it in cloud code running off of files that are accessed uh and let it run, you
27:02 know, more intelligent scripts and prompts over the top of that
27:08 data than I would in a few Excel shortcuts uh or formulas. And it really
27:13 allows you to rethink the way that you go through a lot of these processes and
27:18 sometimes it'll open up right for me it opens up new types of analyses that I
27:22 hadn't quite considered yet or it opens up new ways to look at this data or
27:25 relationships between the data right I would imagine that if you're a
27:31 saleserson uh right it gives you far more access to knowledge on any given
27:36 account or prospect than you've ever had before right you can think of all of
27:41 these ways in which it changes either speeds up the work you were doing or
27:45 changes the work that you can do. Yeah. Trying to think of it in those two
27:49 domains. And just >> I am I am doing I I am I'm dealing with
27:54 this literally as we speak >> the the process I use to manage this
27:59 podcast and the bookings and conversations and who I'm doing. Uh I'm
28:03 going to move into using Claude to navigate those two things. And there's a
28:07 whole sequence of events that happens when we book a date. And it's it's it
28:11 it's exactly that. And I think what I'm learning is maybe Claude's not a great
28:16 example because I know that Anthropic is going to survive in some way, but don't
28:20 also get too attached because it's probably likely that there'll be two or
28:25 three iterations of a new competitor, a new tool, a new use case. Uh you know,
28:29 even what Claude's capable of doing, right? I mean, you look at I think about
28:34 the things that I could do with pick your LLM of choice, right? pick [snorts]
28:39 your model of choice 12 months ago and I think about what I can do today and
28:45 right for on the company side right you almost have and on the startup side you
28:49 have to think about as a founder or as an as an investor what is more than one
28:56 or two or three model releases uh away from being obsoleted or cannibalized
29:03 >> and but as a user we get to as like a a daily driver we get to sort of be
29:07 dragged along with the pace of what these things can do and that's massively
29:13 exciting and a little bit terrifying. >> Y those two those those two experiences
29:18 often go hand in hand. I want to talk about and kind of give us a chance to
29:22 think about looking ahead and we've talked about a lot here. What what is
29:26 the you know the magic eightball if we shook it up and we look at 2026. Can we
29:32 glean anything that we expect to happen in the next 12 months based on what's
29:41 I think based on again I'll go towards the infrastructure sort of the spend on
29:47 infrastructure the massive potentially final private market raises from the
29:52 large uh model providers and then the sort of agents agents agents world that we're
30:01 now really firmly in. We're you know we're not just entering we're not just
30:04 dipping a toe into the agentic world. We're here and all of the things that
30:08 need to go alongside that. uh on the infrastructure side I think we'll get
30:12 you know more and more spend as people try to either figure out the cheaper you
30:18 know sort of the cheaper or more climate friendly ways to meet the energy and
30:26 compute demands uh on the large model side uh we probably see at least one
30:35 more round uh from both uh anthropic and open AAI before they go public but Both
30:39 of them are rumored uh and have made plenty of noise about going public uh
30:44 you know as early as later this year. And on the agentic side, right, I think
30:53 everyone now knows or almost everyone now knows what an agent is, what the
30:58 promise of having something that runs autonomously is. And now right a lot of what we saw
31:06 in the back half of last year and the you know the new companies and greater
31:12 funding to that spaces and what we've even seen in the first few months of
31:16 this year is that you know there's a massive demand for sort of the
31:21 orchestration and infrastructure and sort of the gu the guard rails around
31:26 how you run agents through a business or your own life. And so thinking about you
31:32 know I was looking I spend a little bit of time every day looking at the most
31:37 recent deals and there were a few deals last week uh where you saw actually
31:42 right we saw an investment from Anthropic in a company called Sapium
31:47 that is financial infrastructure for AI agents. So you think about all of the
31:51 different components of the more decisions that an agent can make, the
31:56 more places that it can operate and the more you need to guard rail it and
32:01 provided instructions and understand what it's doing and then you have multi-
32:05 aent systems and the orchestration between those agents and you know I
32:09 think we're still in the early days of what that means. We're going to see a
32:14 lot more funding to companies that operate in those domains. Sort of
32:18 enabling everyone from you and me to the world's largest enterprises to sort of build own
32:30 operate agents effectively and have some amount of measurement uh over the top of
32:35 them uh because you can't improve what you can't measure, >> right? And
32:41 then I mean I think you know largely we'll see that you'll get to the point where
32:48 there are now going to be some early winners in vertical specifics especially
32:53 in regulated industries right when you think about finance when you think about
32:56 healthcare when you think about legal and some of the others that either have
33:02 barriers to entry on uh regulation or barriers to entry on sort of domain knowledge. and
33:09 specificity. Uh you'll we'll start to see like the true leaders emerge in a
33:14 lot of these verticals. Uh or at least the true first leaders emerge in a lot
33:18 of these verticals. >> Wow. Uh that's that's that's a lot. You
33:22 that you know that used to be 10 years of business performance packed into 12
33:26 months. And I I think all of those things are very likely to to happen. And
33:31 I think you used a word in there a couple times about orchestration. And I
33:36 think we we think so much about the the job loss that's coming has come and will
33:41 continue to be impacted. But you also described a an emerging set of
33:47 industries and professions and careers and jobs that are about orchestration
33:52 and synchronizing and supporting people through these journeys. And I think it's
33:56 just another good lesson that when when one market dies, another one finds a way
34:01 to slowly emerge. and you see the investments starting to make bets on
34:05 where they think those winners will be and and to your point, it's a exciting
34:10 and terrifying time. That's just such a great summary to kind of end the
34:13 conversation on. I usually do a big wrap-up question, but you just did it so
34:18 well there on it. Uh I think it's a good spot to end the conversation about
34:22 what's coming and a good setup for your bonus question as I mentioned before.
34:26 And we do we we always have a bonus question that is left for you by the
34:30 last guest and they always work out really nicely and have a have a way to
34:35 tie in especially for your world. Uh so here's the question that's been left for
34:39 you is what is the irreplaceable experience your customers get from you
34:52 The irreplaceable experience. I think I think a lot of it for us right
34:58 now is remains the fact that we have you know very rare validated data across
35:05 both publicly available and proprietary information and then a decade plus of
35:09 smart people thinking about what that data means for the future. uh right you know it's not
35:18 just reporting on the fact that company XYZ raised $20 million or company ABC
35:27 acquired uh Acme Widgets for a billion dollars right it's really what that
35:31 means for the market and you know the ability to then train systems and models
35:39 uh AI tools and agents on top of both the data and sort of what we'll call the
35:45 analyst uh or expert brain, right? It's the combination of the two of those and
35:49 the ability to then tailor that into sort of predictive intelligence on what
35:54 it means for you and what your or your company needs to do next uh to stay
35:58 competitive is >> yeah, that's very well said. I think
36:02 your company should give you a little a little bonus for that. That was a well
36:07 off the top of your head summary of the value of CB Insights and what you all do
36:11 and what you do. Jason, I your level of depth, your level of insight. Um, we'll
36:15 put some I'll give you a chance here to talk about where people can find you,
36:19 but it really is an exciting and incredible time to look at the data and
36:23 where we're traveling and where we're headed. So, so with that, where can we
36:28 find you, follow you, learn more about the topics you're discussing? We talked
36:31 about some dashboards. I didn't get a chance to share them today, but where
36:35 can we track all of that? >> Yeah, absolutely. I mean, so maybe a
36:39 little bit with the uh title of this podcast, I am very active on LinkedIn. I
36:44 share the latest insights on what we're seeing, what I'm seeing across the data
36:49 uh that we have and the predictive intelligence that comes from that almost
36:56 daily. Uh if not daily. Uh so Jason Saltzman on LinkedIn. And then really
37:02 all all of this is enabled by the fact that I get to be a someone who has
37:07 access to too much data. Uh so right can't uh not plug CB insights and what
37:13 you've mentioned as our right if you go and you Google the CB Insights Genai
37:19 signal tracker that's a live updated uh dashboard not just on funding data but
37:23 really on you know sort of what's coming next uh in AI based on where we're
37:29 seeing market traction exit probability uh etc. And then we always put out
37:34 research. Uh there is more and more research that will keep coming out and
37:38 more and more interactive forms of that research that look start to look more
37:44 like the dashboard and for anyone who's sort of excited and wants to talk about
37:48 you know what's happening in private companies and private markets. Uh always
37:53 excited to chat and give you give you a tour demo of the CB Insights platform
37:57 and let you go poke around and get smart yourself. >> Love it. I love it. We'll we'll put all
38:02 those links into the show notes as well. Jason, this has been really really
38:05 great. I would love to keep going. I think you've got so much insight, so
38:08 much that you shared about where we've been and where we're headed and a lot of
38:12 good lessons learned from the past that I think we can all take with us to make
38:17 informed choices and I think you summarized it well. Get going, get
38:21 involved, start using, start learning, start experimenting. Uh the time is now
38:25 and that's a theme that's come up many times on this show. So, thank you once
38:30 again, Jason. Thanks for those of you that tuned in live. I appreciate the
38:33 comment in the chat about the smooth plug. Occasionally, we get those in. I
38:37 think I know who that was. It says user, but I think I know who that was. And,
38:41 uh, we're excited to keep this conversation going. Thank you so much
38:44 for being a part of the Top Voice podcast community, and we'll see you
