The SaaS Apocalypse Reality Check

How much are companies actually spending per employee on AI? Is AI winner take all? Is SaaS dead and is anthropic screwed after the Fable five dust up with the White House. Let's get to the bottom of these questions Separating fact and fiction with ramp lead economist Eic Kzian right after this. This summer, Fandou is the best place to bet on goals. Including equalizers. Uhuh Follies? Yep, headters. Every goal is worth more on fandou. So let there be goals. You customers get three hundred fiftyteen bonus bets guaranteed when you bet five dollars for seven days. twenty one plus in present and select dates. First online real money wager only minimum five dollar wager required for seven consecutive days, five dollars first deposit required. Bonus issued as n withdrawal bonus bets, which expiire seven days after receipt. restrictions applies see full terms at fanou dot com slash sportsbook Gambling problem call one eight hundred Gambler or one hundred My Ret. Depending on who you ask, between eighty and ninety five percent of enterprise AI projects fail. To get AI to work for you, you don't need more tokens. You need better people. A board pairs powerful proprietary tools with senior engineers who've seen it all. That combination means your project doesn't stall, doesn't drift, and doesn't fall. It shifts. Whether you're a startup that needs to get to market or an enterprise with complex legacy challenges, a board delivers exactly what your business needs fast is your partner for AI transformation. Visit abard. com and let's build something together. Welcome to Big Technology podcast, a show for cool headed and nuanced conversation of the tech world and beyond, there are a lot of rumors flying. In the AI industry, a lot of narratives flying in the AI industry in what better way to attack these than to look at the actual data and separate the fact from fiction. Well we're going to do it today. We're going to talk of course about How much companies are spending for employee on AI and whether token maxing is a thing, But we're also going to get into The news of the week, which is whether the White House is Fable five banned Eectively putting export controls on Anthropics model will have a long term damage looking at What happened the last time Anthropic had a dust up with the department Joined by Eric Arazian. He is the lead economist. At RAampM, he's publishing some great stuff at the RAampP Economics Lab. I've been a reader of his work for a long time, and I'm thrilled to welcome him to the show. Era, great to see you. Great to talk to you again. great to be on the show for the first time. Yes, first of many, I'm sure. So let's talk right off the bat about What we can anticipate the impact of the government's Fable five export controls on anthropics business to be, right? because we've seen a version of this before with the Department of Defense naming anthropic a supply chain risk. this is obviously on a bigger scale And you know, even if they go back on the band, there might be some impact I keep calling it a ban. G back on the export controls, which is effectively a ban There may be some impact here. So what can we expect Well, you're right to look at the Department of Defense decision from earlier this spring as the closest Uh recent example that we might use to inform how it's going to affect Anthropics business or business adoption going forward. I mean, let's go back to the spring when the Department of Defense labeled Anthropic supply chain risk, usually in most industries, for most software vendors When you are labeled as such, businesses are not likely to continue to use that vendor going forward So if you're just going to think about this from first principles We would have expected businesses to shut off their anthropic subscriptions for new businesses to not want to sign up for businesses to explicitly not want to use anthropic going forward because of a true security concern that the government is citing and because they might want to engage with the government on government contract That's not what happened earlier this spring, if anything spring was when we saw anthropics adoption really accelerate with businesses. It was coming off the heels of the successful launches of Cloud Code last year Finally starting to move into a more popular posture with non technical users and U Just this past month we showed that anthropic is now the most popular AI model used by U.S. businesses. accord to RAM data Several of those assumptions didn't come to be. I think there are two main reasons why. One is that Most businesses didn't really seem to take the Department of Defense's label very seriously You know, it was kind of like, okay, yeah, the Department of Defense is saying this, but This was one of the best models Available still is the best model available popular with businesses. The second reason is that the Department of Defense lost a lot of credibility over the continuing weeks. when it acknowledged that it would be issuing exceptions both internally and externally for businesses that wanted to use anthropic anyway So if you're going to go back to the example from earlier this spring If anything, it's probably the case that the Department of Defense's supply chain risk labeling accelerated anthropics adoption with businesses U And this label now Anything puts it on a very interesting competitive standpoint with open AI. Who has the better model? probably the one that the federal government has suggested is so powerful it must be controlled. and and brand strengthen that That's right. So we took on this show all the time about whether anthropics positioning around safety and the fact that it's you know, taking this steadfast approach when the federal government has asked it to do something. I that's led to some of these actions by the federal government whether that's marketing Um, P the attent aside though It could end up being a boost for its business, you know, whether it wanted this to happen or not. I keep thinking about At least we're recording Monday and I keep thinking about the fact that you know When you're in Claud, it says, hey, Fable is not available right now And you're seeing all these posts on about people who are like, oh, those five minutes with Fable really give me a glimpse into this could be And you'd imagine that when it gets turned back on, that forbidden fruit effect just kind of drives the interest in fable, you know, the the model too powerful for government to let you let you use, the interest will probably just go through the roof Well, let's also note that part of Anthropics Uh product experience of using anthropic models at least over the last six months has been getting really comfortable with models being down. all the time 'cause anthropic has so many compte constraints that it is very common to be a user of anthropics models and to be hit with a Um a uh crash notice or something that just says, hey, the models are down right now or try again later an APIer. And yet, we would normally expect that that would drive a lot of users over to open AI models. Open AI models are You know, Openye doesn't have nearly as much of the compute issues that Anthropic does. And yet We haven't seen that kind of switching behavior yet over the past couple of months. It maybe some marginal switching happening from people who are using Clawd code, now switching over to open eyes codex Uh but If anything know, it makes me wonder both as a researcher and user of these models How long is This forbidden fruit effect of Anthropics model is going to last before it starts to turn users off who really just need to get to work and use the models paying for and have them work as expected Yeah, I mean, you're the econist, I don't know. Is there an economic theory that explains why people would stick with the vendor U that has constant interruptions even if there's another one with just as good or on par capabilities that doesn't have the turbulence Well, what we found is that these models are a little bit stickier than we thought they would be. And we talk about them being these commodities that we can just switch between one model and the other. And maybe at the model level you really can think of things like that. But in terms of how many employees at firms are actually using AI, Better models don't necessarily crereate the switching for employees and users. It's about the product experience around those models Clad Code was so successful, not because it was powered by the models of Claud, but because it was Uh D. integrated into your workflows in a very effective and agentic way that it was the first model and first experience that allowed engineer to execute on multi step tasks without They being a traap bot the entire time Uh, And so such that inccremental improvements of the model are important and helpful, but they weren't the whole story for U the actual growth of Cloud code And so can you can imagine that to be ' driving some of the stickiness between quQad code and codex O ironanthropic is that People get really used to the tools, the software that they're in. They like the experience that one provides over the other I also do think there is going to be A a sort of branding effect. where Anthropics AI safety posturing. Love it or hate it. There are people who really do like that they're the company that at least postures itself as being thoughtful about the effects of AI, whether or not they are the right guardians of that Yeah, it's possible that the federal government is sort of, you know, by disrupting anthropic, maybe u giving it a helping hand by sort of to making its safety messagings seem more legitimate and again, giving it that forbidden fruit effect. All right One more economics style, you know thought question for you then we can get into your data Uh You know, there have been recent reports that opening eye is looking to drop prices drastically And we talked about it on the Friday show that my thought was basically They are looking at lifetime value of potential customers. And if they were to drop their prices, they could get people used to using let's say a Codex and then keep them with the stickiness that you're talking about And so therefore, it would be a good move on their end to say we're going to drop prices to win people over, and then hopefully they'll stick with us Your thoughts Part of it I think is the very natural. This is an extremely competitive market where It's very common to see In a matter of months, a newcomer U take the lead over a relatively popular Um provid. You know, we've seen this in software before, but it's especially true and acute in AI. We saw this with Cer versus G up Copilot. know when coding agents came out, at least when AI cod, autoomte came out, two, three, four years ago Gitubp co pilot was the enterprise tool. That's what everybody used. Unsurprisingly is also backed by Microsoft Um Cursor comes out and within about a year, it hasn the majority of the market Then by the way, quod code comes out And then now Codcode has majority of the market. And then so we saw a similar story with open ananthropic. you know, in RAMps data, we have been tracking using our flagship research RAPI index. The share of firms in the United States that are using AI always paying for it through subscriptions or tokens directly. And then we break it out by which model they're paying for And for most of AI's commercial existence, twenty twenty three onwards, Open eye was clearly the dominant player. someomewhere hovering between twenty to U thirty, forty percent of businesses in the US were actively paying for openen AI And it wasn't really budging much. it was just a very gradual increase, particularly twenty twenty three, twenty twenty four onwards No one was really thinking about anthropic, which was popular with technical users, but otherwise wasn't this sort broadly understood competitor in the market Second half of twenty twenty five, we see month over month percentage point increases. in the shareiff firms that are using anthropics models coming to the forefront last month when anthropic overtook O AI and actual business adoption So Anthropic sits at about forty one percent of US firms are using Anthropic thirty nine point five percent of firms are using open AI Anthropics is still growing, OpenIe is relatively flat And then even in the sectors that are early adopters of AI, we're seeing that growth contined to grow and accelerate Well open eye holds relatively flat. So consuming a dynamic market where you could expect All of the involved players to want to compete with each other But I actually think that You know, right now, the focus is on open end anthropic, but there are a lot of players that are underrated. Google, I think is extremely underrated And I think might end up being one of the big winers here that no one's talking about Okay, but then briefly the price the price war or the price Undercut. Do you think that's enough to dislodge the stickiness. I know you don't have the data on it, but There has to be some formula out there about, you know plays into usage So I do think it's going to be enough to I think look, we're going to come to a headad. at some point where We know businesses keep demanding some better control over AI spent. You know, we went through a sort of token maxing era Everyone was talking about, okay, we need to spend as much as possible on tokens And then now we're in the sort of new era where businesses are saying, hey, we need to actually rein in token spend or at least understand where it should be can continue to rise, but we at least need some control over what it is Neither open AR noranthropic have built products that allow firms to actually manage their token spend nor have they built products that incentivize firms to keep those cost under control. If anything, open phanthropic are incentivized to have firm spend as much as possible So far, You know, they can compete on price reductions But really what firms are asking for is some degree of control You know, maybe that means, hey, build us something that allows us to smart route tasks over to the most performant but also most efficient model for that task Other competitors are offering that It's not usually open and ananthropic though U So your data shows that like again, like you said, Anthropic has overtaken open eye with business speent Very briefly, Eara The criticism of the Ramp economics lab Wh they're well placed or not has been that Yeah, you're looking at, you know, companies with ramp. cards using Ramp which to lean towards uh, you know, sort of startups and forward companies and it's not representative of the full economy Your thoughts The way normally think about this is that So we have actually a pretty good distribution in our data set across sectors. However, no matter the sector The businesses on our platform are inherently more tech forward and that they're using something like Ramp to manage there in general Um I do see that as a strength for a couple of reasons. One, AI is a very new in natation technology and one that is not effectively tracked by other data setets. It's not effectively tracked by government dataets either, which have their own set of criticisms as far as how they are surveying firms about AI spend and also the firms they are surveying themselves being a little bit self selected Uh AI spend, if anything, is skewed over toward these tech forward businesses, such that if you do want to understand how businesses are spending and operating on AI, It actually behooves you to look at these very forward thinking businesses that have been leading this charge. There are more likely to be early adopters of this technology and whether or not they are representative of the average firm in the United States, we know they are not It's more likely now than not that the average firm will look more like these firms in a couple years than they look like the average from today So I think if you want to be forward looking, you want to look into the future a little bit. You probably want to use this kind of datas set. For what it's worse, I actually think that in many ways we underestimate adoption. Okay. All right, we'll get into more methodology later, but you know given that these are the forward looking companies, let's get into some more of your data because I think that there have been some narratives about token waste. that your data has a little bit of a different perspective on that I think should we should just discuss U So you've talked about what is the cost of being an AI pilled company it's seven thousand four hundred and forty nine dollars per employee. per month So you said the top percent of firms spend that much on employee per month ten percent spend six hundred and eleven dollars per employee per month And the median firm spends just eleven dollars, like the cost of an enterprise seat on enterprise ChatTPT or a cloud subscription Now seven thousand four hundred dollars a month is pretty high on a technology, like for a single person to spend that much on a technology seat or license is somewhat unheard of. Um, However, When you think about the headlines that we've been seeing, a company left clawed on and spent a half billion dollars in a month, Y data actually presents somewhat of a different picture that companies aren't sort of spending unrestrained rightight now, they seem to be uh, you know, sort of dipping their toe in the water as opposed to going all the way in and uh, you know, spending tokens like they're going out of business AI spin is the fastest growing nd category we've ever observed in our data setet Probably one of the fastest growing spend categories for businesses ever depending on how far back you go into business is defined as in prehistoric times I couldn't imagine any spend ramping faster than this Wh what could it be Exactly. And so since january twenty twenty five through may twentyeth twenty sixth, so last month per business spend on AI tokens fififteen s And that's amongst firms that were already spending on AI att the same time, AI spend itself isn't really that meaningfully large for most businesses. So it's grown a lot, but for the top quartile of firms They're spending on AI, top twenty five percent. It's only about two percent of business spend excluding payroll. You at one percent if you were to include payroll So it's grown a lot. That's why we got all these concerns from company executives about, you, how do I manage this growth But as far as its actual level It's relatively small So you'll notice, you know, when people talk about firms pulling back on a spend You know, they might be pulling back on A spend in some parts of the firm. They might be more mindful about which models they're using or making sure that teams don't have uncontrolled budgets But if you actually look at firms spend on AI in the last couple of months, just last month, it's still increased fourteen percent mononth over month So there are There's clear evidence on our platform too, that firms are making more cost to splint decisions. You know, we've seen an increase in AI spend being routed away from open anthropic and over to theseort open source platforms Last month, Deep Sek was one of the fastest growing vendors on Ramp And yet it's still a very small share of AI spend that is actually going through those rails It's a relatively small share of businesses that are using open source platforms in general and the vast majority of spend happening is still rising. So those cost displine measures are important, but they're really just occurring on the margin and they're not happening fast enough to hold back the rising slope of AI spent Do you think that there's going to be A moment where Some firms start to spend more on AI than they do spend, you know on they spend more on AI per employee than they spend on employee. For instance I don't know how accurate this was, but I think like we're actually trending in this direction where someone figured out how long how much it would cost to run the Fable or the Mythos API And they found out it was something like six hundred dollars an hour where if you like multiply that over a year, it's one point two million I So so what do you think when you think about the trajectory, do you think we're going to get to a place where people end up spending more on AI per employee than they spend on place themselves. I imagine some will. Yeah, for some firms, I imagine it makes a lot of sense Right, But for the typical firm, there's really it's really hard to benchmark where you should be And so the top one percent of firms spend seven thousand five hundred dollars per person per month on AI But that's the top one percent. So you can imagine that's a pretty tech heavy group that actually may include a lot of firms that are ultimately usings AI not just for the employeee's usage, but also for the underlying infrastructure of the firm. Maybe they' built a bunch of interimnal tools, right? So it all gets balanced out. And software engineers are double that typically, fifteen thousandllars or close to sixteen thousand doll a month Again, would depend on the firm becausecause you look at the typical firm on our platform and again, this is a ramp. So relatively tech forward platform, right And the median firm is only spending about eleven dollars per employee per month No You know, that's that's a chat subscription. That's like one of the low level open aye and anthropic subscriptions may be a little bit more on the margin Um So You know, it's another reminder, sort of how early we are, right? in that The vast majority of firms and this transformed our research approach because for a long time The last year and a half, my research has focused on trying to estimate the economic impact of AI But if there's no way to find that in productivity statistics, if people are still not sure what the ultimate gains of AI is going to be then Really the only way to start is, hey, how many firms are using AI And so our original version of Rampi index is just that It's, hey, what is a share of firms in the US that are even buying it and buying it month over month? to try to get some way of at least approaching the question, hey, is this valuable Now More than fifty percent of firms are using AI. fifty four percent of firms are using AI in some way are at least paying for it. So our question gets to transform a little bit, not just Wh's using AI is a valuable, but how are they using it How much are they spending on it What does it mean to be an effective user of these models and to deploy it effectively through your organization Be that's what's really interesting about AI too is that it is unevenly distributed and that certain sectors are more likely to adopt it than others. products themselves are also unevenly designed and distributed. and that's some of the best most advanced usages of AI are designed for certain job categories. coding agents It's like the most obvious commercially advanced way that you could use AI to be productivity enhancing And yet that doesn't exist for most other firms. You know, maybe there's productivity enhancements that you can find in a lot of other jobs, but the products themselves are not well developed to Make that queer like the average user Um So A has unevenly distributed Uh So so it's ultimately going to be difficult as a firm to identify and benchmark against good usage of AI means, especially because the effects of AI, the productivity gains of AI are not likely to show up in your first couple months. know there's clearly a learning curve to implementing AI throughout your organization and even through your own personal workflow as employee And then beyond the learning curve, there's also this sort of minimum threshold of adoption I don't think anyone really expects you to get massive economic gains from everyone having a chat bot But butah you everyone having their own claud code for their job is much more compelling But you wouldn't know that if you're just using a chatbot for like a month or two and then you write off AI because it's like You know, what's the point of this But you look at the curves that you have in your research, right? And it's just a number just like three hockey sticks, right? If you look at the spend per employee per month of the top one percent of companies using AI, the top ten percent and the median company using AI It's like legitimately like a like a lightightly sloping line and then all of a sudden it shoots up in all three of those categories U So You're trying to you said you're trying to get to the answer of, you know, how is this technology valuable I And so I'd love to hear your perspective on what these numbers mean U Even though it's more, I guess quantitative or qualitative than quantitative, right do the fact that we're seeing these spend increases mean that companies are seeing an ROI on AI or is it still potentially in the sort of FOMO stage? I really am of the school of thought that businesses have no reason to be spending. this much money just out of a sense of obligation and FOMO You know, I get if we're talking about people buying stocks Right? But like companies making fairly large investments in software You know, that you could just talk about it externally without making Not only the large investments in software, but month over month increases in how much they are spending So I'm generally of the school of thought that if firms are doing this, They must be finding some value out of it. But there are some places that we look quantitatively for that evidence and also informs are thinking that this is different from most software markets So one is that the most advanced spenders on AI don't lock in with one vendor So this is fundamentally different from how we typically think of software where it's like If you're using a CRM, you're going to use one CRM Maybe you'll like experiment with a couple providers, but Ultimately you're going to sign with one. Um That's not the case with AI. The top one percent of spenders on AI use eight vendors on average whereereas the median maybe uses two And that's vendors fairly narrowly defined as LM providers and maybe some AI infrastructure companies Now, it's also not just experimentation You know, there's some amount of it that's always like the sort of continuing experimentation where it's like, okay, the AI models come out, but they also change so much So frequently that if you are an organization that is using and implementing AI effectively, you probably want to have paid access to all of the major model companies so that you can switch to the most effective model for whatever task makes sense or when a new model comes out, you can see if it makes sense for This this workflow or that workflow That's what being a good AI user often means to these firms in the top one percent But that is an unfamiliar idea for many businesses and Frank with business people at firms who are buying AI in charge of procurement I often get the question when we report open eye versus anthropic adoption rates Should I buy open A oranthropic And then when you whichich, you know, if you're someone who uses these day to day these models day to day, it's like a very surprising question because You would just say, well, you should try both. You should probably have access to both. That expensive to have access to both Um And so the question itself doesn't really register as a sensible question But if you are applying typical common practices of software procurement to AI You will find yourself asking that. So I think it's just off funding into a different market that people are not used to And that's what ends up driving a lot of the spend But I don't think firms would move along that advanced AI adoption curve if they weren't getting some benefit You know, you're not going to keep signing up for new vendors You're also probably not going to keep renewing vendors And if anything, in our datas set, we see that renewal rates increase year over year with firms more advanced So they're more likely to stick to the vendors that they've been using. as opposed to you know switch so frequently. One question about that I It's kind of remarkable, right? If you look at the graph of revenue that you're seeing from open AIN anthropic, it is it follows that I mean, you would imagine, right? it was going to follow that hockey stick as well h that that type of curve shape U Isn't it interesting that even as companies spread their spend across two to eight, vendors that those two have been able to grow the way that they have Well, it's because adoption You know, we measure at the firm level But then within the firm There's a lot more heterogeneity around who's using AI and how. And then within the person There's even more, I should stop using economics terms like heterogeneity when I'm on this podcast. Variation Yes know you have different teams that are still on different parts of the adoption curve And then you have individual people within those teams that for different tasks they're on different parts of the adoption curve becausecause people are still figuring out how to onboard This is what I mean about the learning curve. The firm is on a learning curve, The teams are on a learning curve, and then the individual itself is on a learning curve for their specific tasks trying to figure out Hey, can this task actually be done effectively? Have I even tried this task? or not Not to say that everything can be done by AI, I don't think everything can be. But I do think the more that you experiment with it, you will find what it is good at, what it's not good at And then if it's somewhat good at something, you kind of get better at understanding how do I modify my workflow so that Actually, maybe I take out this part of it, maybe I take out that part of it. This part's not really necessary anymore Oh now AI can actually do eighty percent of the job But you don't figure that out without experimentation. So I think that's why you see rising spend over time. We're clearly not at the point at which Um peopleeople have found like their benchmark level they I spend And if anything, if you want to look at our charts of AI spend per person , you don't see a token maxing era There's no point at which oh, it went up and then it went down It's like still going up Again Median firm is only spending eleven dollars a month. So there's a lot of R to grow, but even for the one percent still going up. All right, let's take a quick break. When we come back, I want to talk a little bit more about what you're seeing with Deep Sk because I think a lot of people expected that that deep Sk moment You know, sort of happened in February, January, February, twenty twenty five and then dissipated. But It is u growing once again, so we'll talk about that We'll talk about model orchestration, and then we'll talk about SS, right? whether the SS apocalypse is actually being born out in the data. So we'll do that right after this Switch up the way you play slots with new flex spins on Golden Nugget Online casino. Now you can earn and redeem spins on the slots you actually want to play. New Golden Nugget players get five hundred flex spins on your choice of over one hundred slots. Just play five dollars to unlock fifty spins a day for ten days. Use your flex spins across a variety of popular slots like Huff and Morepuff, Cash Eruption, Cleopatra, and more. The choice is yours. Download the app today and sign up with promo code Nugget onene That's promo code Nugget one to make any moment golden. In partnership with Golden Nugget Online Casino, Gambling probleblem call one eight hundred gambler. Please play responsibly. twenty one and over. physically present in Michigan, New Jery, Pennylvania and West Virginia only, Void in Ontario. Non withdrawable spins issued as fifty spins per day for ten days valid for select games only and expire each day after twenty four hours. See terms at goldennuggetcasino dot com slash promos. ends june twenty eighth at eleven fifty nine PM Eastern time So you're running out of closet space. The good news, you don't need to stop shopping. You just need to start selling with the real real The Real Real is the world's largest and most trusted resource for authenticated luxury resale. whether up's that mini bag that can't fit your phone or those boots you never fully broken The real Re handles everything, from photography and copywriting to shipping and pricing. so you can just sit back, get paid and make room for things that actually feel like you And with ten thousand plus new arrivals every single day from top designers like Prada, Celine, Louis Vuitton, and Luueve, offer up to ninety percent off retail you're bound to find something perfectly on brand to fill that extra closet space with Plus, this may only, you get an extra two hundred dollars to shop when you sell for the first time Make room for what feels like you Go to the realreal. com to start selling and get your extra two hundred dollars to keep shopping at the realreal. com That's the realreal d. com terms apply And we're back here on Big Technology podcast with Ramp lead economist Era Karazian. Erica, great to see you. Thank you again for being here. Let's just talk a little bit about the deep seek growth that you're seeing You're saying basically what you found in the data is that There is an increased reliance on these cheap open source models in Deep Sk When it comes to your list of trending models is number one I What do you think that says about the way that AI is adopting? Is it being adopted? Is it that U as, you know, some people have said the future is going to be that you have this like maybe super spart foundational model orchestrator that makes your big decisions for you like an Opus four point eight or a GPT five point five or six. And then you just sort of deliver the Um The more straightforward work to these smaller agents with the open source models I'm curious if you think that that is already being borne out in the data When I talk to businesses The single most important factor they they list for why they have not adopted AI comprehensively throughout their organization is a concern around the cost. Not just the cost, but not really knowing what the cost is. and knowing And hearing all these stories that, hey, once you adopt AI, it's really hard to control the costs And I think that is an indictment of open eye anthropic. who have not developed predictable pricing structures for their products If you adopt Oen as an enterprise, or anthropy as an enterprise, you are more or less at the behest of your employees as far as how much they spend in tokens. There are very few controls available to admins And so that is ultimately, I think been driving this growth and demand for open source platforms or open source models, or what you're describing these sort of like routing models where you know, you instead of sending your queries direct through openanthropic, you send it through is this like middle layer which then decides, hey, actually this very simple task I can send it to pretty cheap model or a cheaper model even through open iron andthropic and I'll go through that So that's one of the really popular ways to reduce spend What I want people to know? is that Yes, there's evidence that this is happening Uh On the margin, it's a little bit overrated So fiveive percent of firms on our platform. are even using these kinds of open source platforms Last year is one percent So it's five X growth and it's actually going faster than the growth that's happening for Okine Anthropic, but it's a relatively small percentage of companies and typically the most advanced companies as is You know, new newew starters, companies that are just starting to onboard to using AI are not starting with a Chinese model They're starting with open an anthropics So that's the first thing I'll say The second thing I'll say is that You know, we've seen the rise of DepSk in our data set before in early twenty twenty five when D Steps got this really buzzy W. U It's spiked in our dataet then too. it rose to about like I think around half a percent of businesses on our platform for about a month use D deep Sek And then very quickly fell back to Earth through one percent of businesses. And the reason why back then was There was a competit there was a stressful but also competitive response from the American model companies from Op Ananthropic. to offer cheaper but still very performant models that could compete with Deepseek. And so they centrally instituteed price cuts And businesses therefore had no incentive to be using Deepsk anyway To be clear, there are actual security and reputational concerns for businesses that are transacting directly with DeepSek And so you don't want to use Deepsk if you don't have to. So Last month in our data set, Deep Seak also had this breakout growth. one of the fastest growing vendors on RampM's platform But it's growing from a very small base. Only about ero four percent of businesses are using it and that's up from point one So four x increase, but It's extremely small And I think it's not going to be very durable Given that openine Anthropic are well positioned to response to that with some price cuts So I think Deep S is a little overrated I think the open source models in general are a little bit overrated. However, I do think models, companies like Google are very underrated You know, the maincern here dynamic here is that open and anthropic are not being responsive, effectively responsive to firms that want some Cost control and cost discipline Op Anthropic developed models that incentivize you to spend as much as possible on tokens And that makes sense for them because eighty percent of their revenue from businesses is token based. It's not subscriptions It's tokens That's not the case for Google. Google doesn't need firms to spend a lot of money on the tokens in the models. so it actually can offer that are routing It actually can offer these product experiences that give firms a little bit more cost control because they have way more revenue sources available to them They're also competitively well positioned because Google Workspace is already used by So virtually most businesses have some access to it And so Gemini is already a fairly popular model. It's just not Um thought of in this kind of discourse often And so I think Google's really the best positioned and relatively underrated in these kinds of conversations to take market share away from open eye and anthropic Google doesn't need you to spend tokens, then what does Google need you to do? just to use its model so you don't use the others It's not that Google doesn't need you to spend on tokens. they They of course produce revenue from tokens they are They are supported by so many more revenue streams. as well as the subscription revenue stream that makes them less dependent and less laser focused on exclusively having you spend more on tokens R Like so that is where they're a little bit better competitively positioned. I mean, if they wanted to, they could also Have a I be a little bit of a loss leader, it wouldn't be that big a deal. Right I mean, as long as you do somethingbodyI cloud As long as you're using Google's Cloud, right to store your data, for instance, then it's a win for them by making cheap models, they can even have Yeah, I'd be somewhat of a loss leader if it grows cloud and it has. They've been like sixty percent a quarter That's very interesting And their cheap models are extremely popular. Fly you know what they' Yeah, and they're also like a more mature company. U so you would imagine they won't have these problems with the federal government like an anthropic is happenving And, you know, the sort of, one of the responses has been go open source But a different response might actually be go Google because Yeah can trust that those services are going to stay up I would have bet on that continontuity just yet at least with the I kinds of announcements coming from the federal government, but we I get the point that you're making. Okay, All right, I'll take it. All right, I want to end with this apocalypse everyveryone's been talking about how SS is dead and certainly it makes sense as a headline Um, you know, if you're trying to be provocative and you're thinking about what AI can do, But you actually have a post saying the death of SaS has been Greatly exaggerated. So talk through what you're seeing there and why this apocalypse hasn't fully materialized in the way people expected But there's two ways that I think about SaSpocypps. One is that traditional SaaS companies are going to lose a significant amount of market share to open ananthropic. The second way that I think about SAS apocalypse is that Every existing SaaS company just needs to rethink its pricing model in that things are increasingly moving to token based spend or usage based spend. SaaS companies are going to become their own little AI companies, perhaps And so the typical way that we think about SaS pricing being seat based is going to go out the window and every SaS company needs to rethink its whole product and model So I found both of those to be a little bit overrated. as far as actual business behavior in our data set So the first part of Spocalypse, whether or not Oanthropical will Eat every other company We're just not seeing that Like I'll use CRM as an example, right? This is one of those things that's just purchased by so many businesses Look like eighty percent of the market share for CRMs is just directly going to Salesforce. Firms are trying to buy CRM, buy salesforce and then it' some buy hubspot, whatever And that's just always, that's just been the case and it's held that way. And yet in our data set, we can actually see month over month growth in small, but Mighty, if you will, AI native competitors to CRM to Salesforce and Hubspot Audio, that's a London based company has a very low market share today, but is one of the fastest growing vendors on our platform as well. and has apparently durable growth to There's some evidence already to say that Hey first of all, open ananthropic haven't offered there on CRM Theoretically someone could vibecode their own CRM But also The companies are signing up for an audio. They that's a tech forward company They know that they could vibecode their own CRM And they and however, they're still buying as opposed to building themselves And we see that across different kinds of software categories Of course, another one is figma, right? So a claud design comes out. everyone thinks the figmook is going to go under Figma over the last couple of months has continued being one of the fastest growing vendors on our platform and extremely durable software vendors whether or not that's all going to change going forward you know, who's to say, but what I will say is that there is no indication, at least in our data, that there are even early signs of a slowdown amongst these kinds of SaS fundnders. The legacy vendors definitely have some competitive threats But the competitor threats aren't just open anthropic. They are actually AI native software vendors that are taking market share today So then on pricing, that's the second part. where You know, everyone's talking about, oh, we're just going to be paying based on token for everything That's also not quite happening. Um based contracts are still the vast majority of spend for most software. It's like sixty to seventy five percent. U The rest of it is really just flat platform subscriptions metered usage is Extremely small, like five percent less than five percent. And at many traditional SaaS companies that have offered their own sort of metered usage like an Adobe You can now pay for Adobe by credits It's still only like half a percent of their revenue. And then notably even for the AI companies, they're actually growing on subscription spend faster than they are growing on token spend So even for them You know, there is still this demand for subscription based spend So I generally land on Spocalypse as like, hey, maybe these things will happen generally being made by pronouncements from product readers as far as where the data is on actual business behavior overrated. So where do you land on this idea that, you know won't be necessarily that AI can just vibe code everyvery application Um But that You're AI becomes effectively a operating system So you type into codex like what you need and then it opens up figma and works through figma for you Could you see that being the future interface? And if that's the case, if effectively You know, the chat pods are a front end of all software How do you think that might change pricing I think that one makes a lot of sense I mean, I've seen that just as my in my own user experience that I'm increasingly if a product has some integration with SaS that I'm already using I'm more like to interact with it through the models than I am through the GUI Now I'll still maybe go into the website and make my own changes for certain things U As a product experience, it's actually pretty good How that's going to affect spend I mean, look, I do think we'll probably see a steady increase in the kind of spend that is token based in the kind of spend that is aentic, but I think it is Orated, you know, I still think this work tends to be directed by an individual person who is likely going to have an individual seat. I mean, if the AI companies themselves like are still seeing this kind of subscription based growth, and I think that's the best evidence. Any other trends or sort of narrative busts that you've been looking at recently that you think we can share before we go We're thinking a lot about the jobs impact of AI paper coming out about that mostost likely in a few weeks So I'd tell people to keep an eye out for that And otherwise we write about all of our data at rM d. com slash data on substack. Yeah Fall me on substack as well. Yeah, econlab. sububstack. com U