Tuen, for people who don't know what AI inference is beyond its how you serve the models, I'm really curious to hear you explain it because it feels like the most exciting frenetic part of the AI stack right now and you're thick in the middle of it with B 10 and I'd love to know yeah what it means for people who yeah they they know the term but they don't know much else. >> Yeah, absolutely. So what is inference?
So inf inference is the way you serve models.
Quite often um customers will come to us they say to us hey we have this model we want to use this open weights model or we have postrained this model um we need to you know figure out how to actually use it in our applications.
And so when you work with a closed frontier model provider you kind of get this for free.
kind of get this for free. you know they they have the model but they vertically own everything from the model all the way to um the chips it runs on all the way up to the APIs um that you are using to call that model when it comes to post-train models or your own models or
openweight models you don't get that for free and so you'll come to an inference provider and and you'll say hey I need to run this in production um here is some scale approximate scale I need to run this at um and then you'll come to a company like base 10 now there is so much that goes into actually running a model. Um and you know for for us like
Um and you know for for us like that's at the infrastructure level problems.
So that is like hey where is the actual compute going to come from to run this model.
Um so we take care of that.
Um there is the orchestration piece which is hey like how do I um get this user request to the right chip with the right model on it. We solved that problem.
And then there's like the chip level details where it's like how can I make this model run very very fast.
And so really when it comes to inference, you're really like you're not just looking for an API.
You're not just looking for GPUs.
You're looking for like a performant reliable system that puts it all together and puts it behind behind an API.
And Ben kind of takes care of all that.
All right, that was kind of marketing.
So let's double click on like let's actually double click on what does that actually mean.
So you know BAS's like three three different layers of problems that we're solving for our customers.
So the first one is this idea of infrastructure and compute.
So you need you need chips to run these models on.
So for our customers, you know, if you're a customer running this model at um decent scale, you'll probably need thousands of chips to run this.
If you go to the market right now and say I need 1,000 B200s or GP300s or good luck, you know, good luck. It's not happening.
Not not happening as um they're gone.
Someone's already taken them.
So you'll come to like what we do is that we kind of take care of this compute procurement um from uh the standpoint of what you need that's necessary to run inference on for our customers.
Um we do this using like a pretty complex architecture where we're able to pull together compute from all sorts of different regions of cloud.
So we sit on top of 20 different clouds and 90 different regions which is completely abstracted away from our customers.
So that's the bottom heterogeneous compute layer.
I didn't even know there were 20 different clouds. That's >> Yeah, exactly.
Well, you know, you call them clouds, neoclouds, data centers, hyperscalers, you you call them to to us, they are, you know, sources of compute >> to some extent.
Um >> yeah, >> okay, let's go let's go up one layer.
That becomes a core inference layer.
And that's kind of what I describe, which is like, hey, I care about running these models really, really fast and really, really reliably.
Um, and so that's kind of like what Ban has done for a long time now.
That is what we are known for.
And our customers come to us not only saying, "Hey, we want the compute, but hey, we need the whole software stack on top of that, on top of this, so we can run this so it doesn't go down.
We can run this so it's very fast.
We can run this with the right numeric and um and so so on and so forth." That's the second layer.
Um we can talk as much as you want about that.
The the third layer is around all the primitives that come on top of that.
And that's kind of like, you know, these are inference adjacent primitives a lot of the time.
What I mean by that, these are things that either make your inference more powerful, they give you different paths to different types of inference.
Um, they make your models more powerful, but they're all related to inference in some way.
So that might be something like post- training models and RL on top of those models, which not only helps you train more models, but has a bunch of inference baked into it.
It might be something like sandboxes, um, which is how do you actually execute code?
>> We're going to get into that. Yeah. >> Yeah.
How do you actually execute code generated by these models?
Um it might be stuff like emails and routing.
When people talk about inference, they're kind of talking about all these problems together.
All these problems together and like but there are lots of different parts to it.
>> It's everything except training, right?
It's it's serving AI in production.
It's when you get a chat GPT response back or you use Muse or Grockbot or any of these things, they're all inference and and and inference is growing like crazy it seems like as a market.
But it also feels like it's hard to make sense of it.
There's a lot of players.
They're all approaching it in slightly different ways.
And I'd love to like better understand how base 10 differs in its approach from not only like your direct competitors and inference, but the traditional clouds because I think someone listening or watching this who uh is maybe not as in the weeds as we are would go, "Well, why would someone not just use AWS or Azure or cloud to Google Cloud to run this stuff because they have AI?
Why do we need these AI clouds doing inference? >> It's a good question.
Look, I I I think the the the most important thing with all this is that the market is just ginormous.
And this is important for a reason, but if you if you kind of project forward four or five years and there's some silly number a silly number associated with um how much AI spend is let's call it $5 trillion is what they say in 5 years from now.
It doesn't matter what that number is.
It's somewhere between 2 and 10 trillion.
Let's say AI runs about like a 50% gross margin.
So that means there's about you know$1 to5 trillion of inference spent in the market and and you know if you think about how that's going to be divvied up between hyperscalers between closed frontier model between inference price like base 10 um any which way you cut it um that's somewhere you know between 10 to$50 trillion of market cap that's going to be divided between the people who are capturing capturing inference.
inference. If you then start to break down how different folks are approaching inference um everyone has like a slightly different take I'd say so like you know there's a the hyperscalers who have like historically been really good um at infrastructure like things but they've also you know been really good partners so I think the hyperscalers will have inference solutions and you
know they do today but they're also really good partners um to the um inference companies themselves like we we have really strong relationships with all the folks you mentioned um there um The reason why you'd use us over hypers scale, you know, really a lot of it just comes down to like building specialized software and it's like, hey, we do one thing, they do many things. And so like
And so like when you're when you're in a market that is moving as fast as this market where you need to run stuff in production and and when your inference provider is down, um your product is down, you you generally tend to who will go who will move as fast as you in the most customer aligned way and give you the most attention.
And that is base 10 today cuz just cuz we give you um we do fewer things and allows us to be bestin-class at those few focused things that we that we really do. Yeah.
And and and I think it goes beyond that because I think you know there's just so many layers of this problem that I described and we're kind of solving it layer by layer there.
Whereas I think a lot of other folks and like you know whether hyperscalers whoever else are like kind of like hey I've got some basic inference solution that works off the shelf but isn't really dialed in to serving users in production at scale.
This episode is brought to you by Mercury AI native banking that's loved by more than 300,000 entrepreneurs including me. Visit mercury. com to learn more.
Mercury is a fintech not a bank.
Check the show notes for details.
Thanks also to Granola, the AI notepad for people in back-to-back meetings.
It works everywhere you do and lets you focus on what matters. Try it at granola.
ai/sources and use the code sources for 3 months off.
This episode is also brought to you by Jira by Adlassian where teams and agents get the context, coordination, and control to move work forward. Try it free at jira. com. That's jir. com.
The criticism maybe of of the market you're in is that these inference clouds, they're just reselling, you know, GPU capacity at a very thin margin and they really have to prove the software layer still.
>> And you just said like the software is what matters and you have an acquisition we I want to that you just announced that I want to talk about today on that.
But can you explain that a little more like why the software matters so much because as you said like the chips are hard to get and if you've got chips you know you see the base 10 ads in San Francisco and it's on the billboards and it's like very clear it's like you're saying I've got compute then that seems valuable but at the same time you're saying well the software is actually what matters.
Yeah, look, compute's necessary, but like the software is what differentiates, right?
So, it's like I I don't want to like downplay the the role of comput or comput's obviously like the market constraining thing right now.
Um, at the same time, at the same time, there is so much value to be driven to be built on top of that.
The way we think about this is like if if you go if you if you wind back a few months from now, there's been this really big push towards owned intelligence.
And this what is the idea of owned intelligence?
The idea of open intelligence is like hey because of cost control because of data sovereignty because of you know credible permanence which is the idea that you don't want models just to disappear on you one day is very very important um for every company in the world to some extent to own their own intelligence. Cool. What does that mean?
Well, what what that means is that they need to have the capability to not only run open source models but and open way models um in addition to closed frontier models but to figure out like when to use which how to use their data and the and the user feedback um from their customers to improve these models.
customers to improve these models. what customers are realizing like like inference is one part of that and I would argue it's probably like the backbone of that to some extent but there's all these things that you need to do to un unlock that continual learning loop that makes customers kind of internalize um their intelligence so
that might be inference that might be um how do you evalidate eval each of these models as they come out it might be how do you post- train models with RO environments um to make these models better um it might be like, hey, the execution environments for when these models are producing code and doing their own things for them to do this work. Um, and then eventually it's like
Um, and then eventually it's like how do you get a better model out of it all that you run it again and kick off this whole cycle again.
That whole like thing I just described to you is what we call like a continual learning loop to some extent.
some extent. And that is what the software layer that we are like that's a software stack like the software stack we are building um is all those things I just described to you and all of them in their entirety which in a lot of ways become their own hyperscaler or new
cloud when we build it and compute is necessary underneath that until you have that thing until you have that loop working the compute's kind of useless or it it is a it it is exactly what you said it is reselling of you know taking access to to it and putting it somewhere else. But the minute that um we build
But the minute that um we build the software layer for that continual learning loop, what we are doing is like providing every enterprise, every customer in the world some credible advantage they are getting from running their own intelligence.
And what that does for them is it protects their margins and as a result of them protecting their margins, they're willing to pay a lot of that margin to us >> because we because we we we are kind of making them independent um again to some extent. >> That's interesting.
I I haven't heard continual learning framed in that way.
I mean, when you spend time in the labs, you know, the frontier labs, they're talking about continual learning in the sense of the models themselves learning and getting better constantly uh and and the research loop closing and being aided by AI, but you're talking about it in like a a sovereign company AI sense of of controlling your own destiny, which I I hadn't heard before. >> Yeah.
Well, I mean, and it's exactly that, right?
This idea of owned intelligence.
I think like Satia has talked a lot about this.
Jensen has talked a lot about this. >> Yeah.
>> You know, which is like, hey, your user data, your your signal from these models um the the improvements to these models themselves will become the core IP of every company in the world.
That like every enterprise like the ext the extent to which you thrive in a AI first world is the is the extent to which you own all those things.
In order to own those things, you need to own that loop that I just described to you.
And like that is what we are building for like that that is like you know you don't want that all that data to be the input to someone else's continual learning loop which is the loop you which is the loop you described about how how how the model provider companies get better at training training their models. >> Yeah.
>> Uh and like like that that to me is the software stock that is the differentiated value that you're providing.
That is when this gets a lot more interesting than just bare metal comput.
Does it feel to you like it does to me where in the last like 8 to 10 months it just really feels like AI has started becoming useful in a way that it was not before?
was not before? I mean I think that you know post chatbt obviously like that product took off and kind of kickstarted this whole boomer but it's was this kind of like better chatbot Google search thing for a while and I mean even in my
own work both you know as a as a podcaster but you know investor now too like its utility is incredible and I'm curious if you've been feeling that as well because you started base 10 in what like 2019 which was well ahead of all of this uh the models were not capable at all. No company would be running LLM in
No company would be running LLM in production even in like 2021 2022.
So yeah, I want to go to the founding story and how you saw this when you did, but also do you feel that what I'm describing about the shift in AI right now?
I >> I I think there was something in December that happened where these models just kind of had a capability shift.
I think it's happening again right now like in the last month where there's like another capability shift but but the last like six, seven months have been insane, right?
because so much has happened.
Like obviously it felt like coding models finally got really good.
The computer shortage happened.
Um open weight models started getting very very good and like you know it's like release after release after release.
It's just like oh there's another one there's another one there's another one.
Um and and I think what has happened as a result of this is that one people are paying attention to intelligence.
So intelligence um people using intelligence and and enterprising using intelligence has skyrocketed.
As a result earnings calls have become about using AI companies are being judged by how much they're spending on AI.
That in turn is like making them you know rethink hey what is our long-term cost strategy and how do we one you know get AI in more places and two how do we do it in in some way that doesn't blow up our companies.
Um, and I think that has just t taken over the entire narrative, but also just pushed adoption even further. It's kind of it's wild. >> It's wild.
And and I can't imagine being inside it the way you are.
I mean, we before we started this, you were saying I may need to step away cuz I'll get paged because one of our customers needs help.
Uh, cuz everyone's growing like crazy.
Um, before we get into more stuff, go back with me to to 2019.
You're starting this company with your co-founders.
You were in banking before. Is that right? How did you see this?
because it this this market didn't exist then. >> Yeah.
It's like my background is that um I grew up in Australia. I moved here for uni.
Um graduated uni in 2009.
I did what everyone does in 2009.
It's like oh, you know, you go work in banking. Yeah, that sounds right.
So, so I moved to New York. I worked in banking.
I worked in infrastructure finance.
Um basically doing like project financing for infrastructure projects like to roads and bridges and airports.
Um which is which is interest, you know, they're all data center people now, funnily enough.
Um after that I you know started working in technology.
I ended up working at a lab in um in Boston working on using machine learning um to predict and diagnose neuromuscular disease. That was in 2012.
From 2012 to 2019 machine learning was very very early.
Um you know it's kind of more classical models than more classical models than um large language models.
You know if you go to think about um what machine learning was being used for 15 years ago um it was stuff that like PayPal was doing.
stuff that like PayPal was doing. you know stuff that Facebook was doing and and then a lot of it was like fraud classification it was recommendations and like we were kind of working in that paradigm for a long time I started a bunch of copies didn't really go anywhere 2019 um started another company and we were really thinking at the time like two things one is like you know my
two co-founders of Phil and Amir um and then soon after that punkage um but Phil and Amir you know they're my best mates and so really the idea was like how do I how do I sell a company my friends that was number one number two number [laughter] Number two, which was um just as important was this idea that in 2019 we kind of knew that machine learning was going to be a big deal. Like OpenAI had
Like OpenAI had been around for a few years at that point.
Um they were doing stuff like you know smart people were going to work there.
Um but it was still very early like there was no there was no commercial proof whatsoever.
But like look machine learning looks like it's going to be pretty pretty big.
We don't know what that's going to look like.
know what that's going to look like. cuz we had no idea you know we're just like you know I like you know that's why I'm like so so in awe of all of our customers because I think application layer work is so challenging like actually you know actually understanding that end user problem in a domain requires so much product and domain expertise that frankly we just didn't
have in that way but what we knew was like machine we knew a budget about machine learning we thought it was going to be a big deal like oh well if we believe in this and we know a bit let's go just build a pick and shovels business alongside um to support machine learning And you know, as long as the market the market is big enough, we'll have a shot of building something big. I
I think that would have happened in 2022 where the market just accelerated in a way that none of us expected and then like quarter after quarter it just feels like you know we're like surely this is it, right?
And then you know and then and then it takes another turn.
I I was I was chatting with some folks that you know you work with with Guy and and Effie and Jill about this on Monday and and basically this idea that you know even 2022 when um you know they started thinking about machine learning and AI like it felt really early.
It didn't feel like it was obvious then either and it just feels like quarter after quarter it's like it just come online in a way that we like it's kind of gone beyond all our wildest dreams.
And so in 2019 when we started the company we were definitely building an infrastructure company.
I think that infrastructure that infrastructure company had many different facets.
One facet of it which was seemed to us at the time was the least important facet was being able to serve models.
>> Um and then that just became the entire company.
>> And you raised it 5 billion valuation in January of 2026 and then 13 billion valuation in June.
So >> in 6 months you almost tripled your valuation.
What changed in those months?
Is it what we've been talking about?
>> Yeah, it just it's all those things, right?
Which is like I think I think in the last like 6 months what has changed is that we've gone from thinking are open weight models going to be a thing is like to oh wow this might be a majority of token volume.
>> Do you think open will be that?
>> Do you think open will be that? I think open and custom and owned intelligence will become um like I think I think you know the the the future is mixed intelligence where you're using closed frontier using open weights models using post-trained models but I think custom
frontier models will probably like be some version some percentage of the models that are like uniquely served to needing the most powerful models in the world at all times and you know and the you know like you think about right now like think about what AI could solve 3 months ago versus what AI can solve now
and AI could solve a lot 3 months ago or two months ago even if you can run that two-month build intelligence for 90% of your tasks at a fraction of the cost and and own all that data sovereignty and ownership narrative that I talked about earlier like that's a no-brainer and it's very rational >> how are you feeling about the state of
open source though because the US is so behind China is still dominating I think a lot of people a lot of tech leaders are worried about what's happening and you know a bunch of cos are recently advocating for, you know, the government here in the US to protect open and and open weights and open source. But yeah,
But yeah, curious what you make of what's going on right now.
>> Yeah, look, I I I think I think it's a both a very exciting time and a and an interesting and challenging situation, right?
So, look, we we want more intelligence everywhere.
And I think you know openweight models are fundamentally a very important part of the ecosystem.
Like you know in in absence of open weights models you know all that ownership narrative that I just told you about is is gone like you need it.
They're they're a necessary precondition.
So we understand their importance to the independence and sovereignty of of American enterprise.
So that that's why I'll say one two.
So I feel really good that you know we there are options now.
So that's amazing on where they come from.
This seems temporary to me.
I think there will be American openweight models.
I hope in like from what we could tell there'll be really good American openweight models in weeks or not even months or quarters like soon enough there'll be really good American openweight models.
And I think you know once that flywheel kicks off of like hey we have stuff that is competitive with Chinese open source openweight models.
I think um a lot of the narrative that you're seeing right now which I think is a little bit more fear driven and like are we falling behind driven than it is capability driven um if that makes sense is um it I think that will kind of wash away as well.
I'll give you a good example of this where all the open weight stuff really kicked off 18 months ago when deepseek when deepseek v3 came out which is a fantastic model came out over Christmas of 24.
The first you know we were obviously serving that to a bunch of customers but the first thing I did was I was like hey okay we we've met so many enterprises over the last six or seven years.
I was like let me go and talk to 10 of them right now and see how they feel about them.
And the reality was that I'd say 90% of them if not more were like look we would never use this in production.
this is, you know, like like we we don't know where these are come from.
We don't really understand how this works, like you know, what have they been trained on. Um, yada yada yada.
And and I'd say they were like deep v3 was interesting because I think it was like the first time you'd had something which would approach the frontier but was still behind.
>> You know, it was still like three behind and like it wasn't clear how you'd run them and there was all this like uncertainty around was more of like >> and even though it was behind it was such a big moment, everyone was freaking out.
The air race has been reset yada yada. Yeah, 100%.
Like and it was like it was like the paper that they trained it on like a fraction of the cost and blah blah blah.
Fast forward a year now you have all sorts of models, you know, you have like you even have some really good American models like Inkling, but then you have like um you have GLM, you have Kimmy, you have Deepseek and you know hopefully Neimatron.
The Neatron family is getting really good from Nvidia as well.
Um all that narrative from 18 months ago is completely evaporated.
every enterprise we talk to now is very excited to run open source open weight models and I think even even like debating where they're coming from isn't really relevant.
I think they're just like this is just the new world and turns out that like fear uncertainty and doubt was more just capability driven as opposed to actual what does this mean driven.
So the minute they got good enough, everyone's like, "Can we use them?"
And and like, you know, the infrastructure needs to exist, open way models needs to exist to to run them.
But like we're in a really good spot right now in terms of um you know, entire owning that entire stack that's necessary to be able to run these models in production.
I think as that stack matures and obviously we are we are to some extent building that stack.
We are building that stack.
until that stack matures and as that stack matures I think the the barrier to adoption of open way models is going to go down is going to go down as well.
>> Does open have to keep growing for base 10 to keep growing?
Are you really kind of tied at the hip with this?
>> That's a good question.
Um per like the intellectually honest answer is yes.
you know, you know, there's probably people who don't want me saying that, but um the the intellectually um honest answer is yes, which is like we we you know, we are we believe in the open weight ecosystem.
We and we want to give back to it and push it as well.
Like, you know, we we we launched something called base labs a few weeks ago, which is our research lab, which is like, hey, how can we make you know, we acquired a research team >> and then last year um they're doing really good work and we're like, how do we put all this stuff out as well?
But like we we are fundamentally you know behind the open weight ecosystem and and like we um the extent to which I don't think openweight models need to exist but I think customers need to want to own their own intelligence and that that needs to be a a a idea and concept that matures aggressively and I think open way models are just an accelerant to that.
that. So like does that make sense which is like you know like they're like a core ingredient of that >> and and does base lab signify that you're gonna start doing your own models base 10 >> if we need to like we will do whatever is necessary for customers to own their own intelligence at some like we post train models extensively with customers today um like so what that means is that
we take a base model and then we take their data and we set up their learning loops um to be able to you know have those models and run them in production today we are not doing large scale pre-training that's folks like Neatron on who are doing a lot of that stuff and we're building and we're part of the Neotron coalition and we're doing work with them there. Look at some point if
Look at some point if it's like you know we feel like we have a we have a differentiated advantage to training models ourselves and now you know the compute situation such that it allows it. Yeah, 100%.
Like I I think it'd be you know that is just our way of giving back to that ecosystem that we believe in. We try to power.
>> Mercury is a modern take on banking built for startups like mine.
When I decided to start my media business, Mercury was by far the most straightforward, fullfeatured banking solution for me to set up quickly.
The interface is intuitive and simple, saving me valuable time every day.
I use Mercury to track my spending, bills, and invoicing.
I love that I can delegate permissions to my team so they can keep things running for me in exactly the way I want them to.
My favorite part is how forward-looking Mercury is with AI.
Legacy banks are stuck in the past, but Mercury is built for how modern software works today.
I use its built-in command assistant to analyze cash flow and help me move money.
And Mercury also connects to other AI tools like ChatBT and Claude.
I use this feature all the time, and the folks at Mercury actually let me know that I'm one of the top users of it.
So, trust me, it's finally easy to get real time financial data about your business wherever you need it. Visit mercury.
com to learn more and apply online in minutes.
Mercury is a fintech company, not a FDIC insured bank.
Banking services provided through Choice Financial Group and Column NA members FDIC.
I spend a lot of time context switching between meetings, often with no time to process one before the next starts.
Thankfully, Granola runs in the background the whole time.
It's an easy to use AI notepad for meetings that works everywhere, even on phone calls.
I use Granola to recall what was said in meetings and create helpful summaries.
I use it every day to stay on top of what I need to get done with my team.
It connects to my email and suggests follow-ups for me to quickly review and send, saving me valuable time.
Granola isn't just a core part of my workflow.
It's basically my second brain. Try Granola at granola.
ai/sources and use the promo code sources for 3 months off.
AI is only as useful as the context it has, but when that context is scattered across tools, threads, and DMs, your team and your AI agents are flying blind.
That's the problem Jira by Atlassian solves.
What's the goal tied to your project?
What got decided last week in Slack DMs?
At Lassian's teamwork graph pulls all of the valuable pieces together from Jira, Confluence, GitHub, Slack, and more, so nothing falls through the cracks.
You get 44% more accurate results with 48% less token usage.
With Jira, you can easily share your work context with the AI agents you already love, like Claude, Cursor, and GitHub Copilot.
Assign them work directly or connect your tools through MCP.
All of this lets you spend less time digging through endless links and messages, chasing down what got decided and by who, and spend more time actually shipping. Learn more at jira. com. That's jir. com.
Framer is the AI website builder that brings agents into the same canvas where your website is designed, managed, and published so you can move faster without giving up your taste or control.
Framer powers the sources podcast website at podcast. sources.
news news, where you can find new episodes, transcripts, and a lot more.
Learn how you can get more out of your site from a Framer specialist or get started building for free today at framer.
com/sources for 30% off a Framer Pro annual plan. That's framer. com/sources for 30% off.
Rules and restrictions may apply.
It feels like everyone who has a stake in the compute buildout is going vertical.
You mean you've talked about Nvidia, which I know they're an investor as well, but like they've got Neatron now. They're doing models.
They just bought Hugging Face.
They're going more vertical in the compute stack up and down.
And it feels like everyone is realizing this is such a big market.
It's so important to control your own destiny that you've got base labs, you know, you're you're thinking about the model layer and how B 10 can play there, not just at the compute and the software stack for inference.
So I'm wondering do you agree with that take that kind of the incentives of of everything right now are driving companies in your position to try to own as much of the stack as possible?
>> Yeah, it's a good question.
Um I I can just speak for our ourselves like look like the reason why you own more and more of the stack is for two reasons, right?
It's either to deepen customer value and you know ideally you know add more value to them at the um at the end and hopefully skim skim off some of the top for yourself as you add that value.
Like that's that's one place and the other one is like to the extent that you feel blocked by the rest of the ecosystem.
The reason why like you know we we think about going going like building up and down the stack is like you know we build up the stack and that's all the primitives on top of it I talked about that third layer um because we think um that is how we acrew more value and add more value for the customer and like we we own more of that learning loop that's like a very valuedriven thesis.
valuedriven thesis. I think we go down the stack because you know we need to be in charge of our own destiny and we need to know that um you know we we think that in 12 like 12 to 24 months we will be you know one of the largest largest individual users of compute on the
planet like you know like like that that is that is the way that that is you know there'll be the labs and then there'll be base 10 you know like like that is the way that this is trending in order to make sure that we can continue to grow at the pace that we want to grow at
we just need to at least believe that we have control over some of those things that said downstream of us or upstream of us however you want to think about it and like that is why we ver like vertically integrated in that way >> can you give me a sense of the scale of base 10 today you mentioned you work
across dozens of clouds can you talk about revenue growth custom number of customers yeah >> anything that kind of puts a a weight to what you do >> what's the scale what's the scale that we do you know we do like you know I think 40 to 50 trillion tokens a day. That's a lot. That's, you know, That's a lot.
That's, you know, bigger than a lot.
>> And how much were you doing six months ago?
>> I I think I think I think token volume has grown 40x year on year. >> Wild. >> Yeah. It's kind of insane. And then so that's one.
Um two is like you know we have thousands of customers.
have thousands of customers. I'd say you know the the way to think about base 10 is that you know if you have if you have used any of the application layer companies and um we've probably touched part of your workflow today probably in the last few hours and like your tokens that you have consumed have probably run through base 10 um through all the apps
you use through all the um through all the products you love um and then in terms of in terms of revenue like I don't know what we have published said but it's like I think we've 10x in the last 12 months like it's like you know It it is everything about our revenue, about our business is just kind of multiplying as the market is accelerating and like that's it's a weird thing. It's a weird moment that
It's a weird moment that we're very very blessed to build in this market and then obviously there's a ton of like macro tailwinds that are of continuing to push.
>> And speaking of that, a trend I'm obsessed with right now is the rise of these agents using virtual machines and browsers for people to get things done.
I think you know very uh at the consumer level there's Muse which Meta just shipped.
I had Zach on the last episode talking about that.
>> Um Instinct is taking off in Silicon Valley world.
Uh there's startups like Town and then there's Grockbot and Chacht Work obviously but the idea is you take a a model and you give it a harness that includes a computer and a browser and the ability to log in and do things.
And most people in the world who use AI have not experienced this yet, right?
They're still in that chatbot uh paradigm of before, but this is really starting to take off and it also implies that token volume continues to go through the roof, I think. >> Yeah.
>> And and you just made an acquisition in this space. >> Yeah.
>> So, I want to understand that and how you think that is going to start intersecting with what you do at B 10. >> Yeah.
Look, I I think this just becomes, you know, these runtimes and execution environments for these the place where agents do work.
That's the way to think about it.
It's like you know and like um for us like this is like our first version of like starting to build stuff specifically for this idea of like agents you know like like instinct like town like Grockpot like where they're going to need to do their work and we've been thinking about this space for a long time now.
long time now. um we met um Paul and Black and the entire team and we were like just blown away by the technology that they built and like really the idea was was like hey like I don't think inference and sandboxes are different problems I think they are extensions and and related like you know where where
these models run um is where the execution environment should be and that makes a lot that makes a lot of sense to me and then we start to think about hey like it's so interesting when you start to think about all these problems together because you know a lot of these sandboxes or these execution
environments run on CPUs not GPUs >> right >> then you start to think about that owned infrastructure thing again it's like oh should these GPUs and CPUs be collocated should you know how how do you and like really we start to think you know it is both again the extension of the value that I talked about earlier for our
customers um where it's like you know where the agent if inference is primarily going to be driven by agents in the future um we're also going to need to understand where the agents are doing their and communicating themselves and that comes into those execution environments. It comes to the browser
It comes to the browser used and that's where we start to think about black and how that fits into base 10.
It's very very natural but then we start to go down the layer again and think about the layer of heterogeneous comput.
comput. like how does that change when you these two things are colloccated and like that is why this is such an interesting company I think like it's so complex and I think you know I just want to like bring this back to the first question you said about the competitive environment like inference is a very
very very special workload and it needs like new hyperscalers and new clouds to be built around it and that's like that's really how we think about not only what we do but how we are acquisitive and how we join forces with folks like black hole >> do you think that billions of people are eventually using agents with virtual machines. Like do you think this gets
Like do you think this gets really big? >> Yeah.
I mean I mean you kind of seeing it with like instinct and you know like >> and and love to instinct I I love it.
I use it but it's I don't think it's huge yet.
I don't think it has millions of users.
>> Oh no no no to totally but you could see it right which is like you know we we've been trying to solve this do work for me problem for so many years.
Like think about how many chat bots there are where like you know can you make this reservation for me? Can you pay my taxes?
Can you can you can you um book a fight for me?
I think this fundamentally changes how we do work.
changes how we do work. This is this I think this is what's so interesting which is you know I I think about our c our customers a lot and I think about customers like open evidence um which is like you know basically brings intelligence to the fingerprints of of of frontline physicians or a bridge
which is like making uh you know healthcare providers be more present by like you know being the ambient scribe rather than making them take notes um being time or I think about folks like Clay who are like kind of changing how go to market teams go to market teams or even what the texture of go to market
looks like or you obviously think of stuff like cursor um which is like coding or I don't know if you're familiar with whisper flow whisper flow is awesome um >> when I work walk around our office in Jackson Square I see a bunch of engineers whispering into their microphones to their agents >> um and what I what I see with all these
things is like we we in a lot of ways like tech and like the the communities that we are in are really really early adopters but you see the impact it's having on healthcare you see the impact it's having on the legal field or with support with companies like um Sierra and Decagon and Bland and so on and so forth. There
There is just no way that it would make sense for all those productivity gains and the changes in how they work to not flow to the consumer experience.
I think like instinct in town and these are just like the first example.
These are just the first examples of that.
And I think I think like you know could I see my my um my mom and dad being you know massive instinct users or town users or you know um 100%.
And like I think that that's where things get really exciting for like how how much more we have to like again this goes back to the first thing you said.
It's like inference is about to get like 100 or thousandx bigger.
Like I think that that's the crazy thing.
It's like the the amount of penetration we actually have in the market right now.
So it feels like you know 1% or 2% um not not not you know not some saturation point.
>> Is it scalable though if hundreds of millions eventually billions of people have agents or multiple agents running computers in the cloud?
Don't we just need a lot more data centers?
Like is this actually scalable?
>> Uh yes I I answer that. Yes.
We need a lot more data sets.
We need a lot more compute.
But you also got to remember that we we're we're like we're aggressively like lowering the cost of of inference as well.
And so like I think both things are true.
One like we need a lot more inference optimization and two we need a lot more power shell and land and um to put to do data centers and we'll get more efficient at building data centers as well.
But the good news is that you know there's a lot of land like you know I I like the country that you know I grew up in is like you know like the majority of it's unoccupied and you know the there's so much space and maybe in the US we're obviously dealing with stuff but um you know we have we we are thinking about and there are really interesting people thinking about how we create energy in in a sustainable way.
How do we figure out land?
How do we make these data centers more compact?
Um, and I think that will just become a very big big topic of conversation um, in the coming weeks and months and years.
>> A big theme on my show so far both with Sam Alman and Zach was the data center backlash and I'm curious how that touches your world.
Is it something that impacts you base 10 and your customers?
Is it something you're t thinking about a lot?
People are really reacting viscerally to to what's happening and I'm curious how you think about that.
>> Look, does it affect our business? Yeah.
like yes like you know like we we we need a lot of compute and we need to scale and we need to figure out how to how to get intelligence in the hands of more people and that's going to run a lot of comput and I think you know all these things to some extent just slow us down.
I think at the same time like you know there there are there are both valid concerns about like you know what this means for these talents and and and the countries where we're putting these things up but also like a lot of education to do.
I think we've done a pretty terrible job as the industry so far kind of bringing everyone else along and describing hey what does this actually mean?
How does this change things?
>> To me, it's more just it just shows to some extent the responsibility that we all have to be able to um make sure that this thing that we have seen the value acrew from grow to the potential that we see it has.
I don't know if it helps for us to be going and like you spread spreading FUD around these things.
Um but but I think it's to to me it's it's a lot of it just like education driven as opposed to anything else.
Well, I think it's very tough because yeah, I I think a lot of researchers at Anthropic especially, but also Open AI, they really do believe that what they're building has potentially negative effects.
I mean, everyone has seen recently the ex anthropic researcher who said, you know, 10% chance it may kill humanity, like most viral tweet of all time in terms of views within like 24 hours. >> Is that true? That's wild. >> It's true. Yes.
I think Elon confirmed that.
And so you see that and then that researchers on Anderson Cooper, you know, later that evening and like it's hard to balance all these things because you just listed a bunch of your customers who are doing genuinely impressive things with AI, helping doctors, um, you know, helping lawyers be better at their jobs, etc.
But at the same time, there's this existential just like fear and dread about what AI will do.
And I'm wondering I'm wondering if a company like Bayen that is that sits in such a critical part of the stack, you know, what role do you have to play in this conversation?
Have you thought about that?
>> Yeah, look, I I I I think a lot of it for us is just, you know, there's two things you could do, right?
There's like one, make sure we are listening and we understand like the like the actual environment.
Like we can't live in a bubble.
bubble. Like this is kind of the thing I said to you earlier as well which is like living in San Francisco we just have like a very specific view of the world or we the way we see it and like seeing how it's affecting doctors how it's affecting lawyers like that's that helps ground you to like hey this is the thing and then two is like telling those
stories is very important and and and highlighting those use cases and three you know I think the biggest thing is like look we have had lots of powerful technology for many years you know like there's been um and the way we the way we work around them is to figure out like what are guardrails and what are the balances that we put um to make sure that they don't cause undue harm. But
But also, you know, like that's the amazing thing about academic research and and policy is that, you know, there's a counterbalance to these things.
It's like, you know, when we we encourage we encourage um security hackers to find day zero day zero vulnerabilities and that is so we can build an ecosystem around that to patch them and fix them and roll them out.
I think it's the same thing will happen with AI.
I think the biggest challenge right now is just like things are moving very very very fast. >> Yeah.
>> And we're not having that time to react.
I think like that that is like probably where the responsibility comes in which is like take stock of what's happening and then and and just be thoughtful about the future we don't live in and then make sure that we bring people around and uh along and um you know maybe don't drink too much of our own Kool-Aid. >> Yeah.
Uh speaking of things coming fast, I've heard you say that um inference may be the last market post AGI.
uh which is an interesting take.
Um the the idea that maybe the labs just you know reach their training goals and uh AI training stops and the billions of dollars spent on training uh shifts to inference.
Do you really believe that?
And like how do you see that practically playing out?
Because you know I just spent a lot of time in open AI and they were talking about AGI with Astra and um you know Sam and Greg told me they've basically got it.
They feel like they have AGI but the world is continuing right and like new models are still shipping and so I'm curious.
I'm curious uh if even Astra and what a OpenAI said recently has changed your view on that or not.
>> All that claim is is like is is that we just have a lot of AI and like and whether it's the final market or not is definitely going to be the largest market. That's for sure.
Like running these models in production and and serving end user value is going and agents doing work this way is going to be massive.
be massive. Um the final market claim is really just that all you know even in a world which you have AGI and like again like I'm not um I'm not educated enough to to be able to you know um to speak to that to speak >> the researchers who are building it can't speak to it either don't worry they're all they all have different
opinions >> I can't speak that but what I can tell you is like look without without AGI infrance is going to be the biggest market ever with AGI um well the only thing left to do is for these models to run it'll be the only market left >> cuz these models are going to run on in loops and do everything for us. And if
And if that's the case, it's like, well, we're going to need a lot of comput and a lot of software to be able to run these models really really fast.
And that's that that's what that claim is.
Yeah, I think I believe that.
>> I want to end here because I think it's a really interesting look at where things are going.
You all shipped I believe it was an MCP server and a skill so that coding agents can use base 10 directly. >> Yeah.
So does this imply that agents are becoming your customers?
>> Yeah, a agents are deploying models in base 10 for sure and running them which is kind of wild.
A agents agents are deciding how which model to use when >> and and how recent of a phenomenon is this?
>> Oh like I think this has been happening in some form for a while.
We're just getting you know the agents are just getting good and the models are getting good.
Are you imagining a future where like your customer support instead of you getting paged mid interview to go deal with a customer that had needs something fixed like base 10 agents are talking to customer agents?
Well, if you go if you if if you go into our like instant channels when things happen, the agent is debugging for like like while we are working on figuring stuff out.
The agent's debugging by itself.
It has access to all our code.
It has access to everything about the model.
It knows everything about the architecture.
everything about the architecture. it knows about the customer issues come up and it's like hey there might be something unique about this model that's using this data in this way you should go check this out and then the extent to which we give it power to do that thing itself and patch it like today we don't like we we run stuff in production for
you know in a and these are really important things and like we don't want we don't want to delegate all that like such like large impact actions um but I think for sure I think these just like all these problems just start getting solved by agents and what what happens the human skill has to change to figure out how to interact with the data to give them the most leverage and vice versa. >> What is the last thing you and your team
>> What is the last thing you and your team are going to be delegating to agents?
>> The last thing hang hanging out with each other.
I don't know like you know the human contact like I went to I went to dinner with a few of my colleagues yesterday that that was a highlight you know there's no time I'm delegating you know my social interactions.
Um >> but is that maybe all we've got left? >> That's all we got.
I mean that's all we ever had.
I guess what I would argue with you, all we had is people. All we had was people.
That's a great place to end.
I think >> this is getting way too philosophical. We got to end this.
No, Tan, I really appreciate it.
I appreciate you uh chatting through all these things with me. Thanks for joining. >> Oh, thanks, Alex. Thanks for having me.
>> Banking should feel like modern software.
Get everything you need in one place. Visit mercury.
com to learn more and apply online in minutes.
Mercury is a fintech, not a bank.
Check the show notes for details.
Granola is the best AI notepad I've tried.
It works everywhere on a video or phone call, in person, or an Apple Watch. Try it now at granola.
ai/sources and use the promo code sources at checkout for 3 months off.
Jira Lasting is where your team and your agents work from the same context. Try it free at jira. com. That's ji. com.
Framer is the AI native website builder that lets you build faster without giving up control. Visit framer. com/sources for 30% off.
Rules and restrictions may apply.