# Modal CEO on the Future of Inference, the Chaos of the GPU Market & Origins of the Sandbox Bet

Markdown copy of https://realalpha.app/v/HJDekXH-h5c/ on Real Alpha, the best minutes from long interviews with the people building AI. Cite and link the page, not this copy.

Erik Bernhardsson argues that inference providers need a broad infrastructure platform—not just cheaper token serving—to build durable differentiation.

- Show: Unsupervised Learning
- Hosts: [Jacob Effron](https://realalpha.app/people/jacob-effron/) & Pat Chase
- Guest: [Erik Bernhardsson](https://realalpha.app/builders/erik-bernhardsson/) is co-founder and CEO of Modal.
- Published: Oct 28, 2026
- Length: 59 min

## The best minutes

- [5:19](https://realalpha.app/v/HJDekXH-h5c/?t=319) [Modal’s early sandbox bet became a major revenue line](https://realalpha.app/s/modals-early-sandbox-bet-became-a-major-revenue-line/) (5 min)
- [9:59](https://realalpha.app/v/HJDekXH-h5c/?t=599) [Modal built core infrastructure to make cold starts faster](https://realalpha.app/s/modal-built-core-infrastructure-to-make-cold-starts-faster/) (3 min)
- [42:58](https://realalpha.app/v/HJDekXH-h5c/?t=2578) [NeoCloud growth depends on debt and GPU resale value](https://realalpha.app/s/neocloud-growth-depends-on-debt-and-gpu-resale-value/) (5 min)

## Also worth watching

- [1:23](https://realalpha.app/v/HJDekXH-h5c/?t=83) [GPU procurement remains a fragmented, informal market](https://realalpha.app/s/gpu-procurement-remains-a-fragmented-informal-market/) (4 min)
- [21:21](https://realalpha.app/v/HJDekXH-h5c/?t=1281) [Inference providers need a platform beyond token serving](https://realalpha.app/s/inference-providers-need-a-platform-beyond-token-serving/) (4 min)
- [27:01](https://realalpha.app/v/HJDekXH-h5c/?t=1621) [Agents write most of Modal’s merged code](https://realalpha.app/s/agents-write-most-of-modals-merged-code/) (3 min)
- [30:17](https://realalpha.app/v/HJDekXH-h5c/?t=1817) [Agents make software deprecations harder](https://realalpha.app/s/agents-make-software-deprecations-harder/) (3 min)
- [16:04](https://realalpha.app/v/HJDekXH-h5c/?t=964) [Most workloads should use the cheapest GPU that fits](https://realalpha.app/s/most-workloads-should-use-the-cheapest-gpu-that-fits/) (4 min)

Skip inference research and open-source discussion, roadmap prioritization, and the closing quickfire on hiring and Sweden.

## Full analysis: Erik Bernhardsson says sandboxes reached a $300 million run rate

- Modal's sandboxes are about one-third of revenue and a $300 million standalone run-rate business.
- That supports Bernhardsson's platform thesis, but basic token serving still faces price pressure.
- Founders should test adjacent-product demand while modeling fragmented supply and possible rental-price increases.

Its 6 sections, what it means for you and what to watch are on [the page](https://realalpha.app/v/HJDekXH-h5c/#analysis), free with an account:

- [7:48](https://realalpha.app/v/HJDekXH-h5c/?t=468) Sandboxes turned a small test into a major business
- [11:22](https://realalpha.app/v/HJDekXH-h5c/?t=682) Custom infrastructure makes speed reusable across products
- [21:51](https://realalpha.app/v/HJDekXH-h5c/?t=1311) Platform breadth offers more durable defense than token serving
- [1:50](https://realalpha.app/v/HJDekXH-h5c/?t=110) GPU aggregation reduces sourcing work, not market friction
- [47:11](https://realalpha.app/v/HJDekXH-h5c/?t=2831) GPU commitments expose clouds to funding and resale risk
- [55:57](https://realalpha.app/v/HJDekXH-h5c/?t=3357) Bernhardsson expects GPU rental prices to rise

## Summary

Bernhardsson is the co-founder and CEO of Modal, which began building a serverless compute product around late 2021 and later added sandboxes and LLM inference. The YouTube description says Modal is past a $1B run rate.

Bernhardsson says GPU sourcing remains fragmented and operationally messy, and that Modal’s multi-cloud capacity aggregation helps customers manage shortages, reliability and pricing. He argues that inference serving alone risks commoditization, while a broader platform built on hard infrastructure and adjacent products can create more durable differentiation. He sees opportunity in low-latency multimodal workloads, training and agent infrastructure, though he says it is difficult to know which software patterns will last. On market structure, he views NeoClouds as viable but capital-intensive and predicts GPU rental prices will rise over the next 12 months as major buyers keep acquiring capacity.

### Minute by minute

- [1:23](https://realalpha.app/v/HJDekXH-h5c/?t=83) **GPU sourcing remains fragmented.** Bernhardsson says compute is available, but Modal often has to accept higher prices or longer terms and makes two or three compute deals a week. It works with 25 clouds and aggregates capacity, while sourcing can still depend on informal WhatsApp groups and personal introductions. He says reliability, location and hardware differences make compute difficult to trade as a uniform commodity.
- [5:19](https://realalpha.app/v/HJDekXH-h5c/?t=319) **Modal expands beyond serverless compute.** Bernhardsson says Modal began with serverless functions, then grew through Stable Diffusion and adjacent audio and video workloads. The company launched sandboxes in 2023; after coding models improved, the product took off in 2025 and reached about a third of revenue. LLM inference also reached about a third of revenue, leading Modal to launch dedicated inference products.
- [9:59](https://realalpha.app/v/HJDekXH-h5c/?t=599) **Custom infrastructure speeds launches.** Bernhardsson says Modal built its own file system, container runtime and scheduler rather than relying heavily on off-the-shelf software. He says that work supports fast cold starts, CPU and GPU snapshotting, and loading data to GPUs in seconds. Modal also built a system using peer-to-peer RDMA to move data between nodes quickly.
- [12:29](https://realalpha.app/v/HJDekXH-h5c/?t=749) **Inference optimization remains empirical.** Bernhardsson says inference still involves testing techniques such as disaggregated prefill, KV caching, faster kernels and quantization. Modal has a team of 20 to 30 people focused on model performance, while model architectures may change the value of current optimizations. He says infrastructure gives Modal room to benefit from open-source inference engines and contribute improvements back.
- [16:04](https://realalpha.app/v/HJDekXH-h5c/?t=964) **Different workloads fit different chips.** Bernhardsson says NVIDIA remains ahead, but older GPUs can serve smaller models and workloads such as audio, transcription and biotech more cost-effectively. He says the usual heuristic is to choose the least expensive GPU with enough memory for the model. Modal is exploring automatically switching chip types to improve utilization while keeping costs consistent for customers, and he expects model sizes to keep growing for now.
- [21:21](https://realalpha.app/v/HJDekXH-h5c/?t=1281) **Inference rewards platform breadth.** Bernhardsson says straightforward token serving is becoming commoditized and that performance improvements will require continual investment. He sees a broader platform—including sandboxes, reinforcement learning, fine-tuning and batch processing—as a more durable advantage. He cites customers using several Modal products, including Cognition for inference, sandboxes and reinforcement learning.
- [24:53](https://realalpha.app/v/HJDekXH-h5c/?t=1493) **Roadmaps require ruthless prioritization.** Bernhardsson says Modal has about 150 engineers and more possible projects than it can pursue, so he weighs work by return and sometimes pauses projects already underway. He says AI agents now write about 95% of the lines of code merged at Modal, but engineers still need to supervise them, especially on infrastructure where reliability and scale matter. He has seen senior engineers use agents to build more quickly.
- [30:17](https://realalpha.app/v/HJDekXH-h5c/?t=1817) **Agents change how tools get adopted.** Bernhardsson says agents need the same fast feedback, good error messages and documentation as developers, but deprecations are harder because models may generate code based on old examples. Modal tests whether agent-written code builds and runs, and tracks whether coding agents recommend Modal for relevant tasks. He says getting into training data can give software tools more staying power.
- [33:18](https://realalpha.app/v/HJDekXH-h5c/?t=1998) **New infrastructure bets remain uncertain.** Bernhardsson says sandboxes began as a low-downside, high-upside bet that initially took only days to prototype, before demand forced repeated rewrites. He is interested in low-latency inference for voice and real-time video, mid-sized model training and reinforcement learning, and support for full operating systems. He also sees potential in agent orchestration and related infrastructure, but says the architectures are still evolving.
- [38:55](https://realalpha.app/v/HJDekXH-h5c/?t=2335) **Multimodal workloads open new uses.** Bernhardsson says robotics will often run on devices, while Modal can support testing, development, training and simulation. He sees growing activity in audio, music and image workloads, and expects more speech-to-speech applications. He says current voice assistants remain poor and believes connecting speech systems end to end could enable better products.
- [41:00](https://realalpha.app/v/HJDekXH-h5c/?t=2460) **Growth brings recurring rewrites.** The hosts describe Modal growing from about $2 million in run rate to more than $1 billion. Bernhardsson says revenue grew about 500 times over three years and that rapid growth has repeatedly forced the company to rewrite systems and rethink processes. He says he is comfortable with the mess of scaling quickly.
- [42:49](https://realalpha.app/v/HJDekXH-h5c/?t=2569) **NeoClouds face capital constraints.** Bernhardsson calls NeoClouds viable but capital-intensive businesses, with growth dependent on debt and their cost of capital. He says GPU lifetimes and resale values have been better than some expected, but high leverage can make equity values volatile. He warns that NeoLabs' GPU commitments can create implicit leverage: some may owe further payments after raising money upfront, and could shift toward reselling capacity if fundraising slows.
- [48:12](https://realalpha.app/v/HJDekXH-h5c/?t=2892) **Post-training must become a product.** Bernhardsson says a limited set of large companies will train or post-train models, while most companies will not need to do so. He sees opportunities to provide better models at lower cost, but says providers need a product rather than a service business. Modal plans to focus on a cluster of adjacent products rather than expand into every part of AI infrastructure.
- [51:36](https://realalpha.app/v/HJDekXH-h5c/?t=3096) **GPU rental prices may rise.** Bernhardsson says coding agents have not changed his view that demand for engineers will grow, though he thinks new graduates may face a harder near-term market. He says one lesson from building Modal was to run dependent work in parallel when possible. He predicts GPU rental prices will be higher in 12 months because he expects Anthropic and OpenAI to keep buying capacity.

## Transcript

Unsupervised Learning, hosted by Jacob Effron and Pat Chase, with Erik Bernhardsson.

Transcribed by machine (Whisper) from the episode’s audio, so a word may be off here and there.

[0:00](https://realalpha.app/v/HJDekXH-h5c/?t=0) How much of like an organized marketplaces compute these days versus like the Wild West? I went to GTC last year. It was like an open air drug market, but like GPUs. Walking around, like talking to people, trying to get GPUs and meeting people in hotel lobbies, exchanging numbers. It's very far from an efficient market. I mean, some people are working on this. Compute comes in many different shapes or forms. I also feel like structurally, it will be a little bit like sort of unwieldy for a long time. Like it's just not something you can like sort of buy and sell in an exchange because it's not like a fungible contract.

[0:33](https://realalpha.app/v/HJDekXH-h5c/?t=33) Welcome to Unsupervised Learning. As you can probably tell from my voice, I am not Jacob. I am Patrick Chase, GP at Redpoint. So awesome to have Eric on Unsupervised Learning for the second time. He is the founder and CEO at Modal. I have learned so much from working with him over the last three years. He is one of my favorite people to jam on AI ideas with. He has an incredible ability to move seamlessly from the lowest altitude of developer shipping code to the highest altitude of where the market is going and building what is missing. We talk about the mobile journey over the last three years, the crazy technology they have built to make it all work, inference, sandboxes, and what's coming next. Hope you enjoy the episode.

[1:15](https://realalpha.app/v/HJDekXH-h5c/?t=75) For you guys, there's obviously a ton of AI compute demand and lots of people going out and trying to reserve, you know, GPUs and CPUs. What's it like for you guys finding compute in the market today? It's pretty rough. I mean, like, you know, not rough in the sense that, like, we can actually find compute. I think there's a little bit of a misconception that, like, we're sort of, we are certainly scarce, you know, in terms of compute in certain ways. But, like, there is capacity you can get. You just have to look a little bit further, accept higher prices, accept longer terms, et cetera. And we do that a lot.

[1:50](https://realalpha.app/v/HJDekXH-h5c/?t=110) **Erik Bernhardsson:** I mean, we probably do like two compute deals every week at this point, if not three.

[1:55](https://realalpha.app/v/HJDekXH-h5c/?t=115) And so we keep, you know, stacking up, you know, more and more computing. And I really think that's the beauty of modal is we can aggregate all this demand and manage the capacity in aggregate much more efficiently. And the valid problem of modal has always been like, don't think of my capacity. Like we handle all of that, like annoying stuff, the sort of the day-to-day knife fighting of getting capacity. We work with 25 different clouds at this point and adding more all the time. How much of like an organized marketplaces compute these days versus like the Wild West, like random people that somehow have capacity and, you know, late night text messages?

[2:27](https://realalpha.app/v/HJDekXH-h5c/?t=147) It's a lot of late night WhatsApp groups and like it's pretty unorganized. Like, I mean, I went to GTC last year or so earlier this year in San Jose, the NVIDIA big conference they were doing. And it was like an open air drug market, but like GPUs, like you're just like walking around, like talking to people, trying to get GPUs and, you know, meeting people in hotel lobbies and you know exchanging numbers and you know i mean at some point we i think we got connected with someone who had gpus because i think our vp of sales cousin was golfing with someone and we got connected like there's many such cases like like it's very far from an efficient market i mean some people are working on this and i think over time i get a little bit more efficient but it's very heterogeneous like compute comes in many sort of different chips or forms. So I also feel like structurally, like it will not like it will be a little bit like sort of unwieldy for a long time. Like it's just not something you can like sort of buy and sell in an exchange because it's not like a fungible contract.

[3:29](https://realalpha.app/v/HJDekXH-h5c/?t=209) Yeah. I mean, the key differences being like, I guess, the kind of actual powered shell or size of the cluster and like the chips themselves or to your point, there's a lot of people have been trying to build like, hey, you know, compute as this fungible single asset that you could build the marketplace around what are some of the kind of the key differences you observe i think a lot of it's actually boils down to reliability i mean first of all it's like you know differences in like disk and networking and of course cpu and memory and things like like all these things vary and region or like location like they're all over the world We have GPUs in the U.S., of course, all over the U.S., but we also have GPUs in Indonesia, all over Europe, et cetera.

[4:10](https://realalpha.app/v/HJDekXH-h5c/?t=250) So location is one of those big factors, but also reliability. And that's another thing we do exceptionally well. We aggregate all this capacity, and then when neoclouds die, or when they have failures, which they often have, because some of these are former crypto miners who haven't really thought about, you know, running a 24-7 operation, we can just take it out of rotation. Or like when GPUs fall off the bus or whatever, like in many cases, we just file a ticket automatically and someone goes and looks at it. Have you found the data providers, the companies that provide pricing for GPU and that sort of things, have you found them to be helpful?

[4:45](https://realalpha.app/v/HJDekXH-h5c/?t=285) I think moderately helpful. I think the pricing less so than just their ability to connect with a lot of providers. I don't fully trust the pricing today, but I think some of these providers are actually very good at like, you know, knowing all the neoclouds and we work with them sometimes to get connections with cloud providers. The pricing, I think over time will get more accurate. Like our internal pricing does not match what we see publicly when we buy capacity. But I think over time, hopefully, like we will have more efficient like external indices. I think that would actually be very useful.

[5:16](https://realalpha.app/v/HJDekXH-h5c/?t=316) You've been on this like rocket ship journey these past years. And so lots of things we'd love to dig into for our listeners on the business side. Maybe starting, you know, I think one of the interesting things about the history of modal is, for a while it felt like you publicly said, I think on your first unsupervised learning episode, that you were less focused on these LLM inference calls, specifically like a simple API call for a model. And you kind of recently released a product there. Maybe talk a little bit about the journey and kind of that initial reticence and what made you decide that like now is the right time to really double down on this.

[5:53](https://realalpha.app/v/HJDekXH-h5c/?t=353) Yeah, so we started building this in late 2021, basically, or early, somewhere around that time. And the initial product was this serverless function. And we thought it'd be really useful for a wide range of stuff around like data, AI, machine learning, broadly speaking, kind of like looked at a lot of different use cases. But then like ChatGPT came late 2022 and Stable Diffusion also came around the same time. It turned out like we were like super well positioned to host a number of Stable Diffusion applications. So that was like the first year of growth was like all Stable Diffusion. And then we started like expanding into like adjacent use cases like audio and video and stuff like that.

[6:33](https://realalpha.app/v/HJDekXH-h5c/?t=393) And it was like early teams that, you know, they were deploying JNI inference. They didn't want to deal with down-aligned infrastructure. They wanted to move fast. Like we started aggregating a lot of different GPUs across a lot of different clouds, even back then. We always knew we wanted to be multi-product, but we started with the serverless, like the sort of compute primitives. We felt like it was so general purpose. And we focused a lot on like kind of what I think was like high code use cases, like focus on like machine learning engineers and researchers who want to write code and focus a lot on like being kind of lower in the stack, right?

[7:02](https://realalpha.app/v/HJDekXH-h5c/?t=422) And pretty early, we built also a storage layer, a file system to store data, in particular, container images. Then we launched a sandbox product in 2023. And the reason why we launched it back then, we started seeing people messing around with LMs to generate code. And we figured actually a lot of stuff we built on the infrastructure side is really well suited to run that code. And they're going to need a place to run that in an isolated way. didn't immediately it wasn't like an overnight success but like basically 2025 it kind of blew up because i think models got really good at writing code back then so about a year ago so that was sort of our venture into like kind of becoming like a multi-product vendor like starting about 12 months ago like sandboxes has been in a tier and grown like i don't know a thousand x

[7:48](https://realalpha.app/v/HJDekXH-h5c/?t=468) **Erik Bernhardsson:** almost or something like that it's about a third of our revenue right now standalone is a 300 million run rate business uh sandboxes uh so that's been incredible in terms of you know becoming a

[7:57](https://realalpha.app/v/HJDekXH-h5c/?t=477) multi-product vendor we just launched also a training So, sort of, you know, building, trying to lay out, like, all the different building blocks, and, you know, what does that suite of tool looks like, and what are the toolboxes in this, like, new world of, like, AI applications? Like, they're all going to need a bunch of new primitives, and, like, we want to offer all those primitives, and what does that look like? So LM inference was something We were a little bit on the sidelines for a while Until actually I would say the DeepSeq moment So that was about a year and a half ago Like early 2025 When DeepSeq came out Which was like the first Chinese open source model That was like actually really good And then I was like fuck it, we gotta get in here And like we started building up a team To do a lot of inference optimization Models has always been really good at LM inference But we didn't necessarily have a product for it In fact we didn't even have an inference product until very recently.

[8:48](https://realalpha.app/v/HJDekXH-h5c/?t=528) We were just like really good at hosting inference, including LLM inference, but it was like kind of custom. And a lot of the teams that want to run LLM inference a year or two ago, like they had to deploy it themselves or like we would work with them. But we started seeing a lot of demand with the DeepSeek moment over the last 18 months, started seeing a lot of growth. LLM inference is now about a third of our revenue too. So we've seen incredible growth with that. And a few months ago, we shipped a number of LLM specific features Because we think, you know, we kind of laid this like foundational layers, like the storage, the compute, the serverless functions, the file system, etc.

[9:25](https://realalpha.app/v/HJDekXH-h5c/?t=565) I think there's a lot of opportunity to move up in the stack. So we started moving up in the stack now and launched a dedicated inference product. So you can get like a single tenant models, like all the state of the art Chinese models or a bunch of US wants to like one single click or CLI command. You can just deploy on modal as well as a bunch of shared endpoints. You can just call models off the shelf without having to deploy them, fully token-based pricing. And so I think those are the first out of many more products to come slightly higher up in the layer. We think there's a lot of opportunity because we're sitting on this very solid foundation in the compute and the storage.

[10:02](https://realalpha.app/v/HJDekXH-h5c/?t=602) Maybe you guys will start at the bottom of the stack and work our way up. But going back in time, you touched on it, but you guys invested in some really hardcore technology in the early days. you built your own file system container runtime those types of things how does that show up in

[10:18](https://realalpha.app/v/HJDekXH-h5c/?t=618) **Erik Bernhardsson:** the product today um when the the experience developers can have yeah i think i think a massive source of alpha in this market is just like not being afraid to do hard things

[10:30](https://realalpha.app/v/HJDekXH-h5c/?t=630) uh so we went all in when we realized like in order to develop to to in order to to to deliver on the developer experience we wanted to have in the very early days we realized like kind of to build your own file system to do container images really fast so we we started building that in the very early days um and and that become like over the next few years you know it still is like a very massive source of alpha it's been it's been a longer journey for sure like we've had to go a much longer route like building our own file system building our own container runtime our own schedule like we don't use kubernetes and docker like we we use like very little off to shelf software. We used GVisor, which is a fantastic Google product for isolation.

[11:09](https://realalpha.app/v/HJDekXH-h5c/?t=669) Recently started using also KVM for certain tasks, but decided pretty early on that a lot of the stuff we're going to have to build ourselves. And now let's just do some really cool stuff. I think one of the best examples where that shows up is we can do cold starts extremely fast. Turn out like the initial developer experience that we wanted, like very fast feedback loop when you're launching stuff in the cloud. That is also very helpful when you're building auto-scaling applications that need to scale up and down very fast. So first of all, loading the containers very fast, but over time we also launched a couple of really cool features around CPU snapshotting and GPU snapshotting. So we can very fast load data from the disk or from network straight into the GPU and fire up even these large models in a few seconds, which is easy.

[12:00](https://realalpha.app/v/HJDekXH-h5c/?t=720) really cool. And over the last six months, we built this incredibly complex system that basically uses RDMA over peer-to-peer between the nodes to copy data very fast. So I think a lot of this stuff is like kind of, you know, accumulating and really starting to pay off. Like we built some crazy stuff. And now that's like, you know, we're really kind of, you know, reaping the dividends of that, those sort of investments that we made for many, many years. It was a longer route. But, you know, now I think we have some incredibly unique technology that's very hard to replicate. And then kind of moving up to inference, what do you view as, you guys have made, you know, tons of investments there as well and optimizations. What do you view as the unsolved problems in the inference world or in the inference stack? I think we're still pretty early with inference. I think, you know, a lot of stuff is like, it's like one of those like very empirical sciences. Like I think people are trying different things and seeing what works. I mean, And over time, like people have kind of figured out a lot of stuff, you know, take something like disaggregated pre-fill, right?

[13:01](https://realalpha.app/v/HJDekXH-h5c/?t=781) Like it's a very non-obvious idea. Like you sort of, you know, do the pre-fill on a set of nodes and then you copy the data over RDMA and do the decode and kind of swap. There's some like very non-obvious things, but like we've invested a lot in those types of things. The KV caching, there's a lot of complexity around like KV caching. Of course, like running more efficient kernels, that's something people do a lot. But there's a lot of these things that, you know, we have a team of, I think, 20, 30 people right now who are extremely focused on just like LM performance optimization. And, you know, it takes a lot of work.

[13:34](https://realalpha.app/v/HJDekXH-h5c/?t=814) You have to stay up with research. You have to read papers. You have to look at all the software coming out all the time and try and see what works, what doesn't work. It's still kind of an empirical science. Over time, I think, you know, patterns will emerge. I think also the models themselves will change very quickly. like now we have like mixture of experts and more complex things over time like who knows like maybe we'll swap out the transformers for something else and you know and that would change how like kv caching works like there's all these things like that could you know change overnight if someone figures out like a very different paradigm for all these things and maybe they will maybe they won't i don't know uh there's also obviously a lot of stuff around quantization and and other things um uh so so yeah it's it's it's a wild world out there but but yeah we're we're we're playing around with all these things. But it feels like there's so many folks that are working on, you know, different parts of improving inference. And obviously, what's I think been really interesting in the last two, three years is a lot of those optimizations and improvements have been in open source, right? And like, you know, for the broader community, maybe shared via research. As like a

[14:34](https://realalpha.app/v/HJDekXH-h5c/?t=874) CEO, how do you think about like, the level of what the kinds of problems you want your team to go solve versus, you know, it's like, hey, we're going to benefit from just the entire community, like pushing this space forward? Hi, I'm Anne Maris, and I'm the producer of Unsupervised learning. Thank you so much for listening to the show. And if you've enjoyed the conversation, please subscribe. Subscribing helps us grow, which helps us continue to lock down some amazing guests. Thank you so much for your listenership and see you next episode. As a CEO, how do you think about the level of the kinds of problems you want your team to go solve versus, you know, it's like, hey, we're going to benefit from just the entire community pushing this space forward? We actually love open source because I think a lot of our alpha is in the proprietary side, in the infrastructure layer, like all this cool stuff around autoscaling, for instance.

[15:22](https://realalpha.app/v/HJDekXH-h5c/?t=922) So I think on other inference providers, we've actually leaned very heavily into open source. We work very closely with both the VLM team and the SGLine team. We contribute very actively to both those repos and do a lot of optimizations with them. And so I think this may be a little bit of a different approach than maybe other inference providers. Of course, there's going to be certain things that you know we we keep internal but but almost everything we we push to to open source and i think it's you know it's because like you have so much alpha in the infrastructure anyway like we might as well just you know work with the amazing open source engines that are out there uh They weren't state-of-the-art a few years ago, but now they are, and that's an incredible thing that we'd love to leverage, frankly.

[16:07](https://realalpha.app/v/HJDekXH-h5c/?t=967) You were talking about heterogeneous compute and the different places where you're getting compute from and also how the inference stack is changing. How important are the long tail of chips to you guys, the non-NVIDIA chips and things like that, as you think about your supply or what your serving stack looks like over time? I mean, I think NVIDIA is still like the king here. Like, they have the best chips. There's way more demand for NVIDIA versus anything else. Like, you know, we're curious about some of the other chips. Like, you know, we made some experiments, like taking a look at it. But NVIDIA definitely has the best chips.

[16:44](https://realalpha.app/v/HJDekXH-h5c/?t=1004) But, you know, like looking at AMD, Terranium, TPUs, et cetera, like starting to be like pretty good. So we're always curious about that. And then, of course, there's a lot of like older chips too. And I think the challenge has been, like, for a lot of the LMs, like, you're kind of constantly pushing, like, you know, how many parameters can you fit? So typically, the latest and greatest LMs, like, they will always require, you know, a full node, maybe even in the future, even multiple nodes, who knows? So, like, eight B300s, like, they will require the latest chip and, like, eight of them, because that's how many you can fit on a node.

[17:17](https://realalpha.app/v/HJDekXH-h5c/?t=1037) but there's plenty of models outside of that like some older LMs smaller ones that have they're optimized for performance might be fewer parameters and then outside language there's a lot of stuff around like voice and audio transcription or speech synthesis or like pharma like bio stuff a lot of that stuff actually they don't require the latest and greatest GPUs they actually get more cost efficient serving on hoppers or even amperors or like you know different types of like l40x or rtx pro 6000 stuff like that so we actually i mean we are ramping up like very very aggressively on black walls uh primarily gb's now uh but we see actually like plenty of demand also in the older chips uh which i think is quite exciting there's just so much stuff going on people forget about the hoppers but like they're they are our most common type right now still at modal h100s we run some very like large production use cases based on hoppers how efficient like is model to chip matching today like you know in the industry at large do you feel like most people are using the like optimal you know chip or you know from a price perspective for what they're what they're working or is there still a lot of room for improvement there there's a little bit of room

[18:31](https://realalpha.app/v/HJDekXH-h5c/?t=1111) to improve i mean one of the things that we want to do over time is like kind of automate that for people i think in practice it's like less of a problem than it might seem like because often like

[18:41](https://realalpha.app/v/HJDekXH-h5c/?t=1121) **Erik Bernhardsson:** you're kind of memory constrained so like knowing how much memory you need for a model like you basically pick the cheapest gpu that fits the model that's usually like the heuristic that works in almost all cases so if you have you know if your model is 80 gigs like you can either fit

[18:56](https://realalpha.app/v/HJDekXH-h5c/?t=1136) it on a hopper or an older a100 80 gig i guess in that case it depends a little bit on the flops um but but in many cases it's like whatever whatever gpu like is the smallest that fits your model i think over time like there's a bunch of stuff we've been thinking about that where basically you just deploy something in modal and then we manage different chips and maybe automatically give you discounts because the more options we can have to run it on a lot of different chips, that lets us aggregate capacity in a better way and run things at higher utilization because then we can switch between different GPU types as there's underutilization on those.

[19:32](https://realalpha.app/v/HJDekXH-h5c/?t=1172) So that's something actually we're quite interested in doing over the next six months is giving people an option to deploy things in a way where we automatically switch types underneath in a way that's still consistent with your dollar per request or dollar per token or whatever. I think there's some incredible cool stuff you can do there. It's super interesting and one thing I imagine is so interesting in your role too is like you're obviously predicting for like what your compute fleet should be and obviously you see what the workloads are today and you probably have some sense of you know what models.

[20:00](https://realalpha.app/v/HJDekXH-h5c/?t=1200) and you're also trying to predict, like, you know, hey, a year or two from now, like, what, you know, what kind of workloads will we have? I think there's lots of, you know, there's lots of different debates about whether we're going to, you know, most of the, you know, most tokens running will be really large models or actually, you know, we've saturated most tasks and these models will get smaller and smaller. Like, you know, as you think about kind of the evolution of your fleet in the next year or two, like, what, you know, what's your perspective on, like, where you'll see relatively more different kinds of workloads and less?

[20:26](https://realalpha.app/v/HJDekXH-h5c/?t=1226) I think there's been kind of like basically over a trend of like, like larger and larger models over the last three, four years, I kind of, you know, I definitely feel like that trend is going to continue, like we're going to need, like, if B300s had more memory, I'm sure, you know, a Kimi K3 would be bigger. people are kind of like you know training the models in a way that they fit on these these chips so i i think on the at least on the lm side like on the you know lms are in a way like kind of a crude like type of world model like you're trying to like encode as much information as you can in these lms like you need like very big like almost like lookup tables of facts you know anyway it's kind of a conception i think about it and and and that sort of pushes like the memory requirements a lot i mean at some point like it might taper off like who knows but but like for now, like people will make use of every possible memory, uh, like the more memory, like they'll just train more parameters for sure. You obviously sit in like the, the, uh, in some ways I'm like the eye of the hurricane of like a very exciting, large and competitive market. And so I'm curious as you think about like, you know, what ultimately will kind of differentiate players

[21:34](https://realalpha.app/v/HJDekXH-h5c/?t=1294) in this space and like, you know, and, and, and win out. I think there, there's so many true believers in the size of the overall inference market. And I think lots of folks are asking what actually wins out in the end of the day. And I know you have a good thesis on this. So I would love for you to share with our listeners. I mean, I think a lot of the inference market will be quite rough, right? Like I think the sort of straight down the middle, just like, you know, serving tokens, you already see sort of open router basically commoditizing the underlying providers. I tend to think, you know, like the real alpha or like the real like point of differentiation like there will be a lot of differentiation just in like more efficient serving and but but it is a little bit of a race to the bottom and i think and i think there's going to be a lot of opportunities to sort of stay ahead of competition you're going to have to like put a lot of effort into that uh but i think more sort of durable differentiation will definitely be in like you know providing the entire platform of a lot of different things which is like i think something we are quite well positioned to do we have the sandboxes we have

[22:31](https://realalpha.app/v/HJDekXH-h5c/?t=1351) reinforcement learning a lot of customers doing also normal fine-tuning um a lot of customers obviously you know deploying other types of models we have batch processing i think i think my view

[22:43](https://realalpha.app/v/HJDekXH-h5c/?t=1363) **Erik Bernhardsson:** in general is like customers they don't want to buy like single trick ponies like they want to buy like a multi-product set of tools so that to me is like ultimately like probably the biggest

[22:52](https://realalpha.app/v/HJDekXH-h5c/?t=1372) source of differentiation is like you know how do you provide that that entire set of tools how do you provide this like very general purpose thing that lets engineers like iterate very fast deliver value and then like just a day-to-day like we're still gonna have to like go and do the knife fighting of like you know staying ahead of your competitors in terms of performance and use the latest kernels and the much more efficient kvcash and there's gonna be a lot of that stuff for a long time because there's so much money in it uh but i'm kind of pointing out i think that's sort of that's like less durable as an advantage i think that you know because over time like people these tricks will sort of spread and people figure out how to how to copy them so the more durable thing, I think, is like being that sort of portfolio of a lot of stuff around inference, not just the inference in itself. You know, having all the models and a lot of, you know, stuff like around reinforcement learning, maybe, or other things. You know, the biggest modal customers are on multiple products. Maybe can you touch on a few of the more interesting ones?

[23:49](https://realalpha.app/v/HJDekXH-h5c/?t=1429) Yeah, we have some incredibly cool customers like Cognition is using us at large scale for both for inference and sandboxes, to do a lot of reinforcement learning on modal, these sandboxes for the rollouts. We have customers like Ramp using us for background agents, for the internal coding agent that they actually launched as a standalone product today that they're selling. We have customers like Suno and Runway that use us for regenerative video and audio. Decagon is another customer that uses us for a lot of customer support stuff, voice and text. DoorDash uses us for a bunch of internal applications, mostly LMs, but also all kinds of other stuff.

[24:21](https://realalpha.app/v/HJDekXH-h5c/?t=1461) We actually have a lot of customers in computational biotech using us for various types of drug discovery, using, you know, molecular simulations and sequence alignment and all kinds of stuff. We have a bunch of customers in robotics, One X and a few other ones using us to train robot models, like doing a lot of training, but also batch processing of video and things like that. But a lot of customers in the data generation space, so customers who are building environments and running benchmarks in modal, often using sandboxes. One thing, you kind of mentioned this portfolio of products, and I think that's one of the most interesting parts of the business you're building.

[25:01](https://realalpha.app/v/HJDekXH-h5c/?t=1501) And I imagine, honestly, one of the hardest parts must be just prioritization of all the different things you could build. There's probably so many new things that come up of really exciting tools you can build for developers. And, you know, I guess I'm curious, obviously, you kind of, you know, with sandboxes, you have, you know, incredible products around, you know, coding agents around, you know, reinforcement learning. You know, how do you think about kind of building down that vector? There's also kind of like the fine tuning are all as a service stuff that some of your competitors seem like they prioritize.

[25:30](https://realalpha.app/v/HJDekXH-h5c/?t=1530) Like, how do you think about, I guess, you know, where to prioritize within that portfolio? I think it's an incredibly hard problem at, I mean, most startups, but especially multi-product companies, right? Like the sort of balance of like, you know, first of all, like, when do you go multi-product and then where do you put your bets? Like we're 150 engineers or something like that today. And I struggle with it every time, like every day. Like, you know, I wish we could, you know, do that and, you know, whatever. And like, but I kind of know that even if we had a thousand engineers or 10,000 engineers, like it's all, you know, there's a lot to build.

[26:03](https://realalpha.app/v/HJDekXH-h5c/?t=1563) It's like, you know, I wish we could build a rocket and go to the moon and whatever. Like, it just never goes away. so like you kind of have to be just like ruthless with prioritization and really think about you know what's the highest ROI and like where do you draw the line and you have to be sometimes like pretty aggressive also saying like actually like I know we started on this project I know you're really excited about it but like this other thing kind of came up so like we're gonna have to pause this for a second we'll get back to it later we've had to do that many times too it sucks when it happens but um I really think you know that sort of prioritization and like how do you how do you allocate resources I mean resource allocation It's like basically like, you know, the number one thing that I think a CEO has to really, you know, be ruthless about it, a company and like so much value is being created, like allocating the work to the right resources.

[26:50](https://realalpha.app/v/HJDekXH-h5c/?t=1610) The modal roadmap in each board meeting is extensive and, you know, very daunting. And I'm always impressed by the amount of software that the team can ship and hardcore software while keeping the system running. How much has, you know, you guys were building the company pre-ChatGPT and then, you know, post-ChatGPT moment. How much has that changed, you know, maybe on the product side, you know, what you guys were doing, but then also when you were thinking about building and the velocity that you can ship software at? Yeah, I mean, we've always benefited from, you know, being quite well positioned at a time where like AI kind of blew up.

[27:30](https://realalpha.app/v/HJDekXH-h5c/?t=1650) And that's been incredible to see. And we're powering so many different AI applications today that I had no idea would exist four years ago, five years ago when we started Modal. So that's been kind of incredible. And like you said, software in itself is also... How we build software has changed a lot. People write...

[27:45](https://realalpha.app/v/HJDekXH-h5c/?t=1665) **Erik Bernhardsson:** I mean, I think probably at this point, 95% of the lines of code that gets merged at Modal is probably written by agents to some extent.

[27:54](https://realalpha.app/v/HJDekXH-h5c/?t=1674) Infrastructure still, I feel like relies a lot on... Agents are kind of good at it. I've actually been on a team. Like the last few days, like building or last few months, actually, like writing like an insane amount of code using agents. And like, it's kind of incredible, like what they can build, but like you kind of still need humans to like supervise them quite a lot. And like, you know, infrastructure has a lot of tail risks around the reliability and taking and making sure that things scale and like stuff like that. So at least for now, I think software engineers are pretty safe. And I think we still benefit from having incredibly good engineers that can supervise these agents, I think probably for the next few years.

[28:30](https://realalpha.app/v/HJDekXH-h5c/?t=1710) Another thing I think has been kind of interesting to see is like, actually, like, I think the adoption pattern is actually the opposite of what I would have expected. Like, I thought it would be the junior people that would adopt agents the most. I've actually seen almost the other way around. It's like the very senior people, because for them, it's like giving them like 10 interns. Like, they've been kind of constrained by like typing speed for a long time, but they have all these ideas in their heads. And I think using agents, now they can get it like they can get it out like much faster. They're really good at building a test suite.

[28:56](https://realalpha.app/v/HJDekXH-h5c/?t=1736) Like, you know, give the agents like, here's an extensive test suite. It implements all of, you know, the POSIX standard for file systems. Now go write a new file system. And it just goes on, like, just writes a file system and takes, like, you know, three days to do it. So we've seen, you know, some of the very senior engineers, like, go, like, complete agent maxing. And I don't know. I mean, it's kind of incredible, like, what you can do with agents. You can move very fast with them. At the end of the day, it seems like humans are still safe as software engineers for the next few years. I don't know. We'll see.

[29:23](https://realalpha.app/v/HJDekXH-h5c/?t=1763) How have you found the velocity of shipping software has changed what you guys need to do with code review or CICD or deployment? Have those pieces of your development lifecycle changed at all? I think a little bit. I think we're coming up against some of those constraints. We've had to spend a lot more time on CICD. Even long before agents, I was a CTO for many years. I was always obsessed with CICD. It's very, very important to figure out. So in that sense, I don't know if it's changed my view as you really need to have automated deployments, automated tasks, all these things. Now, by the way, agents are pretty good at writing tests, too.

[30:01](https://realalpha.app/v/HJDekXH-h5c/?t=1801) They weren't like a year ago. So I don't know if we've necessarily yet run up against the foundational limits. But I mean, Anthropic, a couple of days ago, they announced they're moving off Git, I believe, or at least GitHub. We might run up against those limits at some point when we have 10x the amount of commits every day. I mean, it's interesting to think about kind of the, you know, how different the world is with these really powerful coding agents. And obviously, I think, you know, part of what I believe your co-founders talked about is kind of the shift even in who the consumer of modal is over time, right, from a developer to, you know, to agents themselves, right? And I'm curious, like, how you think about what that means for modal as a business and some of the decisions you might make differently as, you know, more and more of folks interacting with modal are either doing it via an agent or purely end-to-end an agent.

[30:48](https://realalpha.app/v/HJDekXH-h5c/?t=1848) it's actually changed like a lot less than you than i think i would have expected uh it turns out agents i mean like in the early days like we built modal in order to give engineers a better developer experience and the ability to move faster it turns out like agents kind of want the same thing like they want fast feedback loop they want like good error messages good documentation a lot of the stuff like that that engineers want to so so most of the stuff is actually quite similar i do think it's like certain things a little bit harder like one of the things like we found for instance like deprecations is much more annoying because uh agents are trained often

[31:23](https://realalpha.app/v/HJDekXH-h5c/?t=1883) **Erik Bernhardsson:** on like a lot of old code that exists online so removing features is much harder for us like we can't just rely on like you know remove it from the documentation and wait three months because like it's in the training data people will use you know lms to generate modal code and that will be

[31:39](https://realalpha.app/v/HJDekXH-h5c/?t=1899) like you know using deprecated features uh and that's like quite hard to change so we've had to change our deprecation window for quite a lot. Another thing we've started doing quite a lot is also setting up a lot of automated pipelines to basically have like LMs continuously write modal code and making sure it compiles and or not compiles. but it builds and runs, trying to just catch things that agents struggle with. And then maybe another thing we spend a lot of time on is more like agent optimization stuff, which is actually more similar to search engine optimization. If you can go and ask Cloud Code something like, hey, I want to do something with inference.

[32:20](https://realalpha.app/v/HJDekXH-h5c/?t=1940) I want to deploy this model and not say anything about modal and just see which tool does it recommend. And that obviously is something we wanted to recommend. model as much as possible so that's something we track internally make sure that you know over time like and actually i think that's like something that that um will change software engineering a little bit maybe almost like to do worse like i like i feel like in a way like because of the training data and and the training loops like there's more like you know like incumbency power in a way like there's more staying power in software today like if you get inside the training data and you get the models to recommend, you know, a certain framework. I've heard people say like, oh, I'm just going to use React for everything because agents are really good at it.

[33:04](https://realalpha.app/v/HJDekXH-h5c/?t=1984) I'm not going to try these like new fancy front end frameworks, right? Like, I think there's like something similar with like a lot of the frameworks today is like that staying power is actually very, very valuable now. So we obviously want to make sure that agents recommend modal as much as possible because that creates like a little bit of a flywheel. And one thing I'm struck by is you obviously had this like very prescient bet in 2023 on on sandboxes right it sounds like you guys had kind of this instinct of like hey this is going to be something that's required for for coding agents and then you know it took a few years for it to really take off and then you know it became this really important primitive i'm curious is there anything else that you guys are working on today that you feel like you're a little early on but like you know you're you're you're equally as excited and and how much of your product portfolio is like those kind of bets versus like okay we know this is a hair on fire like here and now problem jacob are you are you fishing for the next thing we should be investing in a hundred percent a hundred percent i'm nothing if not shameless i i feel like sandboxes they did create a little bit of like a high bar for us like

[33:58](https://realalpha.app/v/HJDekXH-h5c/?t=2038) everyone's like what's the next thing like you know what's the secret sauce you're working on and like what's cooking in the lab you know and and i i wish i had like an easy answer like yeah like 100 confidence you know this thing is it would you have answered sandboxes in 23 or was it just like part of a bunch of things you were doing no because what sandboxes represented back back then was like a little bit of like an asymmetric bat like sandboxes was not like a very hard thing for us to build back then like to getting out like a prototype of sandboxes took like actually like a few days and then i spent a little bit of time you know like improving the sdk and some other people like started building and then we started seeing some customers we rewrote it and then in 2025 when it started exploding then we had to rewrite the system like five times just to handle like more and more scale and make it faster and and and more reliable and things like that so but but the initial investment was actually pretty pretty basic so in that sense like it was a little bit like a sort of asymmetric bat like a kind of a low downside high upside bat i don't know if there's like anything similar today that we're working on

[34:54](https://realalpha.app/v/HJDekXH-h5c/?t=2094) i am i mean like just just kind of call out like a couple of like you know things that i'm quite bullish on and these probably represent like way larger investments uh like you know larger risks in terms of that size, low latency. I think there's going to be a lot of voice-to-voice, a lot of real-time video type stuff. That requires us to rethink a lot of how we manage to compute globally. How do we get down the number of hops? How do we scale up and down in different regions? Almost like building a CDN for inference. So I'm pretty excited about that. I think training is actually kind of an underrated thing.

[35:31](https://realalpha.app/v/HJDekXH-h5c/?t=2131) I think there's a lot of opportunity in training mid-sized models and reinforcement learning that's been a big bet for the next for the last couple years like getting all the rma infiniband type stuff working um but i'll say i think a lot of vm like stuff is quite interesting like i think you know supporting full-blown operating systems like mac and windows and android and iphone could be kind of cool um i think there's a lot of upside there so But I have a lot of those ideas in my head and we. we're like building a lot of those things. I don't know if there's anything that represents the same like sort of explosive potentials like sandboxes.

[36:06](https://realalpha.app/v/HJDekXH-h5c/?t=2166) I mean, maybe like if we have this podcast again in like three years, like we'll kind of look back and... We're going to make it a recurring at this point. So we'll look forward to revisiting that. You know, I guess like obviously, you know, sandboxes have become this fundamental primitive for, you know, coding agents. And it feels like that then brings you into this world where there's like, you know, I imagine you look at the current state of coding agents, you're like, God, there's so many more tools that would be helpful for these agents to make them more effective for people that use them.

[36:32](https://realalpha.app/v/HJDekXH-h5c/?t=2192) How do you think about whether you further build into that world or the sets of things that are interesting for you there? We do think about it, but I think we think about it more from an infrastructure perspective. I think it's very easy to kind of get lost in whatever is the latest today. When I look at the architecture of agent harnesses and how we're building agents and stuff like that, I still kind of feel like we're so early and these architectures are evolving so fast. and like what our job to some extent is like trying to figure out is like what are like the sort of core building blocks that we can put out today that like will last you know like people we need them for the next five years and and that to me is like less obvious like we are interested in like sort of agent orchestration as a category i think there's a lot of interesting stuff there um we're interested in like you know evals and model gateways and all kinds of other stuff around like the whole life cycle of learning from agents and innovating on agents.

[37:27](https://realalpha.app/v/HJDekXH-h5c/?t=2247) But I think it's less clear exactly how the architecture... My view is it will be more cloud-centric and it will be probably more durable objects, kind of workflow engine type driven. And we have some ideas of stuff to build in that space. But I find it harder to lay out exactly what agentic development will look like in five years partly because like i also think like to some extent like all software will like basically be agentic like i kind of hate and hate the word agentic in certain ways because i feel like it's like okay so it's using ai to make decisions like well shouldn't all software do that so like i almost like feel like if you look at it on a five-year horizon like i think like software like all software will become agentic and then what does software look like then like i i don't know but i you know i i hope we can build a lot of the building blocks to support all kinds of aspects of like, you know, in general, just like how software uses AI. Yeah. I mean, it still is like one of the hardest parts of building an infrastructure company, right? Of, you know, across the board for many people, you know, building AI for companies of, you know, what, what actually persists, right? Versus is a pattern, you know, is, is like

[38:32](https://realalpha.app/v/HJDekXH-h5c/?t=2312) rag from two years ago or like, you know, whatever it is. Rag is a good example. Exactly. Like if we had made a bet on like rag, like two years ago, like it would have been kind of a bad thing to do. So, yeah, but, but I would also, so it's hard, but I would also say like one of the benefits of building an infrastructure is like you can take a longer view and you don't have to like listen to all the noise. I mean, you have to listen to the noise, but like you kind of have to like ignore a lot of it and trying to figure out, OK, what are the patterns here? You kind of alluded to some of these other, you know, other workloads that like low latency workloads like voice to voice and real time video. I think there's a lot of chatter in the air about like multimodal inference, robotics inference, these other categories. I'd love to hear you just riff a little bit on like to what extent are those problems, you know, similar to the set of problems the inference markets had before? To what extent are they like different and new?

[39:18](https://realalpha.app/v/HJDekXH-h5c/?t=2358) I think they're all quite similar, but all, I mean, they're like snowflakes, unique in their own way. Robotics is going to be a lot of it on device, of course. So that's a big difference. And you know, I don't sort of envision like we're going to support, you know, most of what robotics looks like, because a lot of it is going to be on the devices themselves. But I think we're going to be able to support a lot of the testing and development and training of these of robotics and a lot of the simulation environments. So that's something I'm very excited about. On the multimodal stuff, I mean, there's audio and music and image and stuff like that.

[39:50](https://realalpha.app/v/HJDekXH-h5c/?t=2390) Like, people don't talk about it as much, but that keeps growing like crazy, too. I feel like LMS have kind of dominated the airspace in the last few years, but I'm just incredibly... you know excited about all the different applications like why are like voice assistants so bad in 2026 like i don't get it like they're just horrible like i have a google home and and like every question it just bungled it does not understand what i'm talking like you know so i just go to chat gpt like why can't you just connect all of it like end to end and train like a system so that that that can you know do like speech to speech i i i think there's a lot of we're going to see a lot of sort of incredibly cool applications in that space i always think about you know you look back like you know four years ago three years ago whatever you know could you have a guess that like things like instinct or suvino or like cognition or like none of those things like are very easy to guess like you know like three years ago so i kind of feel like also like if you look down the road like three years like most people don't have an idea of like what the world is going to look like in three years it's so hard to know but like what i do know is like kind of

[40:56](https://realalpha.app/v/HJDekXH-h5c/?t=2456) judging by the past three years it's going to be like there's going to be some really cool applications i just don't know what it's going to be but they're going to be very cool and hopefully they'll use modal i was looking back at an email you sent me uh with the modal series a pitch deck and this is what you said uh this was after we had signed a term sheet by the way we never actually reviewed a modal deck but here's something we made earlier that i just updated with the latest numbers could be good to review at some point when you sent this So you're at roughly 2 million in run rate. You now, you know, you've crossed a billion run rate, which is insane in three years.

[41:32](https://realalpha.app/v/HJDekXH-h5c/?t=2492) What's been the biggest kind of change to supporting that kind of scale in just, you know, a couple of years? So first of all, I actually realized like the other day, I said on another podcast that I've never made a pitch deck. And then I was actually thinking about it. I was like, maybe I did make a pitch deck after this series A because we had a couple of angels that came in and they were asking about it. So I was like, oh, fine. And I guess I'll put something together. So I realized I did lie. Like I did make a pitch deck, but it was actually really cool. You never showed it to us. So I don't know that it counts.

[42:01](https://realalpha.app/v/HJDekXH-h5c/?t=2521) Yeah. So I don't know if it counts. Like I never like sent a pitch deck to an investor. I sent it to a few angels actually. And that's the only pitch deck I've ever made actually. But in terms of the scale, yeah. I mean, we've grown, you know, 500X, I guess, in terms of revenue, something like that since then over the last three years. So a lot of growth. um i think i think in a way like i don't know maybe i'm like just you know my experience has mostly been at hyperscale startups like i was at spotify for seven years uh and some other places so so for me like there's like fast growth is kind of all i've seen like you know there's like you know every every six months you have to rewrite all the systems you have to kind of rethink everything you're doing you have to kind of change all the processes like you always feel like you're behind like i don't know i i kind of like the mess i guess we would love to you know since since we have your your mind on the podcast we'd love to kind of zoom out maybe talk a little bit about the broader inframarket and some things that have been in the news of late um you know maybe most top of mind these last few weeks i feel like there's been tons of discourse about

[43:03](https://realalpha.app/v/HJDekXH-h5c/?t=2583) neoclouds right they like you know how the debt's trading they've traded down they're kind of back up a little bit now and i'm curious like what do you think happens to this landscape over time I think the neoclots are like fine business models. I think people have been confusing it a little bit. They're like, wow, there's so much demand. Revenue is going to grow so fast. And then there's a little bit of almost like a VC mindset of thinking in revenue multiples. But I think all of these neoclots, growth is actually quite expensive for them because they have to go and raise a lot of debt. And so I don't know if necessarily just because they grow a lot, it's very capital inefficient.

[43:41](https://realalpha.app/v/HJDekXH-h5c/?t=2621) Like these companies are at their core, like, you know, financing vehicles, like they're very, they're like, they're doing leasing, like of hard assets. And that's, you know, that's like an inherently, you know, harder to scale the capital. Now that that market is kind of saying, well, we deployed a lot of capital into NeoCloud. It's like, maybe, maybe hold on for a second. Like, you know. We don't have that much more capital to lend. And so the interest is starting to go up. But also, by the way, like these companies have incredibly high leverage. So of course, they're going to have more volatility.

[44:10](https://realalpha.app/v/HJDekXH-h5c/?t=2650) Like the residual equity value is going to go up and down much faster on the stock market with GPU prices and stuff like that. But they're like fine business models. Like, you know, they kind of locked in customer contracts and raised debt against it. And like, you know, now they're paying off that debt. And in many cases, it turns out the residual value has been higher than people expected. Lifetime of GPUs has been longer than people expected. So they can resell once they paid it off like at a higher price than people thought. So like they're going to make a bunch of money from that. Like it takes a few years to get there.

[44:40](https://realalpha.app/v/HJDekXH-h5c/?t=2680) But I think in that sense, like Neoclass will do well. They're not particularly differentiated. So, you know, long-term, I think they're going to start with that. Yeah, I mean, you work with a lot of them. Like, is there any differentiation you see? I mean, it's like a little bit, but like at its core, it's like, you know, like relatively undifferentiated space. So, I mean, my view is like, you know, I mean, Elon was, like, tweeting the other day, like, all the new clouds are going to die. Like, I think that's kind of bullshit. I do think Elon had a point when, like, basically, like, when he implicitly was saying, like, SpaceX has a much lower cost of capital.

[45:12](https://realalpha.app/v/HJDekXH-h5c/?t=2712) I don't think there's a coincidence that if you look at the, like, the three successful hyperscalers, Amazon, Azure, or AWS, Azure, and GCP, that they're all attached to companies that have very solid cash flow. And they're all public companies that have very low cost of capital. because that at its core makes the life a lot easier if you want to build something that's very capital-intensive. So I think in general, that is something that Neoclasts will struggle with. They're going to have a high cost of capital and it's going to be hard for them to grow fast because of that. So we'll see about the market, but they're like decent business models.

[45:51](https://realalpha.app/v/HJDekXH-h5c/?t=2751) I think there might be an opportunity to buy the dip. I don't know. You had an interesting tweet that was a lot of the Neo Labs could become NeoClouds, you know, if their model doesn't work out, but they have locked in these contracts. I definitely could see the world going that way. If, you know, if you have valuable compute that continues to appreciate because there's so much demand. But yeah, I mean, Mistral kind of did that, right? And, you know, like it seems to be fine. I think a lot of people are making these like very large buying decisions and they kind of rationalize it. Like they're like, well, if you don't need the GPUs, we can just resell them, right?

[46:23](https://realalpha.app/v/HJDekXH-h5c/?t=2783) Like that's almost like a form of insurance. And you have to, you know, assume with like all these Neolab, you know, there's so many Neolabs out there that have each raised $1 billion. They always raise exactly $1 billion for some reason. I don't know why. I guess they like that number. And then buying GPUs. Very precise financial modeling that points to that. Yeah. So some of them, I think, will probably work out really well. And, you know, but I imagine in the grand, you know, like some of them will not work out. And if they don't work out, well, now they're sitting on a ton of GPUs. like you know let's resell those gpus right like especially if there's ever like a scenario where like vc funding pulls back because i think what's what's kind of a scary thing about some of these commitments is like you know effectively like what happens today when you make a gpu commitment is you only have to pay about 30 up front sometimes a little bit more sometimes a little bit less

[47:15](https://realalpha.app/v/HJDekXH-h5c/?t=2835) **Erik Bernhardsson:** but what that means is like you can raise a billion dollars you can go out and get gpus for three billion dollars and only pay a billion dollars up front i mean not exact math but the

[47:25](https://realalpha.app/v/HJDekXH-h5c/?t=2845) problem is like you know in a year now you're on the hook for another few hundred million and like every month like now you have to pay so so i i think that creates like a little bit of like implicit leverage in a lot of these companies and and and a dependence on subsequent fundraising and you know that works as long as they can go and raise vc dollars to fund the gpu bills but at some point it might not work and in that case i think you're going to see a little bit more of a secondary market where some of these Neolabs might, you know, pivot to doing a little bit more of the sort of NeoCloud style, like we have a bunch of GPUs or just resell it back to the cloud or I don't know.

[48:00](https://realalpha.app/v/HJDekXH-h5c/?t=2880) Bullish on AI as a whole, but, you know, I always, like, look for these, like, sort of structural, like, risks. So, overall, I think there's going to be, like, incredible application, incredible demand. But, like, you know, is 100% of the Neolabs going to succeed and, you know, be a trillion-dollar company? Like, probably not. You know, I feel like on the AI M4Sci, one big category that has gotten a lot of buzz is like this URL as a service market and kind of a bunch of folks, you know, almost forward deploying and like helping companies build their own models. And, you know, you've had folks like Satya saying, hey, every company is going to have their own model.

[48:30](https://realalpha.app/v/HJDekXH-h5c/?t=2910) Like, do you subscribe to that? And like, what do you kind of think of this, of that overall market? No, I think some companies will have their own models. I think that like weighted by dollar revenue, like I think most models will probably be post-strain, right? Right. Like, but but like weighted by company, like in terms of like numbers of companies, like most companies will not have the need to like post do their own post training. Right. But like the large ones, like Cognition is training their own code models and and Cursor did and all these companies are. And like those are the ones also serve a lot of tokens.

[49:06](https://realalpha.app/v/HJDekXH-h5c/?t=2946) So I think when you look at like reinforcement learning and post training, like certain types of parts of it is very attractive. right? Like you can work with companies and serve, you know, better models and do it at lower cost and higher throughput and all of this stuff. Some of it looks like it's kind of like just a service business. So I think a lot of the sort of tension over the next few years, when you look at post training, it's going to be, can you turn this into product? Like, can you sell that product to these companies? This is something we think a lot about is like, how do we create this like amazing product you can just kind of sell and not turn it into just like a service business. And then, you know, I guess last one before we move to the quickfire round would just be like, you know, I think there's this, this question about, you know, how many tools live under like a single infra provider in the AI world versus, you know, our separate categories. I've always struck when we talk to you about like just how much you want to go build and all these kinds of building blocks. I'm wondering if there's categories of AI infra, you know, that you think will be unique and distinct

[49:59](https://realalpha.app/v/HJDekXH-h5c/?t=2999) from the stuff you're building that you're optimistic on and that will kind of like live as still large standalone opportunities. Yeah. I mean, we're not going to go into model training i think like so i i think type safe with jav is going to be you know there's probably going to be a lot of these like custom models uh i'm sure in other verticals too you know other modalities like audio and video and music and like obviously on the application side is going to be a lot of incredible companies um uh and then there's probably going to be a lot of companies like focusing more on like software development and and stuff like that and and you You know, I don't know, like my general view on these things is like there's always like more boxes on the market map than you expect.

[50:39](https://realalpha.app/v/HJDekXH-h5c/?t=3039) But I also do think that overall, every company that is successful in the end, to some extent, has to become a multi-product company, too. So I don't know if that's a tension or not. Like, I think, you know, we obviously want to find, like I said, a cluster of like very adjacent things and do all of it in a way that, like, you know, maximizes the revenue and the shareholder value. um but i you know by no means you know are we going to try to expand into everything i think people kind of make this like false assumption also like you know you look at like anthropic or open ai and they're like why don't they move up to stock and launch their own law firm and launch their own whatever like they're very good at like serving these models and doing that at scale that's how they achieve scale right like and they're very good at doing that like i i don't think it makes sense for every component to do everything and so similar for us like we're not and do everything, but we're going to do a very large cluster of adjacent things and do exceptionally well. And I think there's going to be a lot of room in the market to create a bunch of very large $100 billion or trillion dollar opportunities in the market. We'll end on our quickfire round where we stuff

[51:43](https://realalpha.app/v/HJDekXH-h5c/?t=3103) probably a similar set of questions, what we asked you three years ago, maybe a little bit of an evolution. To start, maybe I'd love just one thing you've changed your mind on in AI in the last year. Well, the opportunity size of sandboxes has obviously been kind of crazy right like i i didn't think it would be anywhere close to you know i mean i thought I knew that in 23, and that's why you made the big bet. Yeah, so that's been kind of incredible. Coding agents have changed a lot of what it means to be an engineer. How, if at all, has it changed how you think about hiring? Not much. I think we're still going to need humans for a long time.

[52:18](https://realalpha.app/v/HJDekXH-h5c/?t=3138) I think the demand for humans is just going up the more productive they get. I think we've seen this for the last 30, 40, 50 years of software engineering. you know, whether it's compilers or high level programming languages or CICD, like every time you make engineers more productive, like the demand just goes up. And so I think we're going to need a lot of different, a lot more engineers. I do think maybe on the short term, it's going to be like a little bit harder for like new grads, like, you know, except for that, the very high side of the spectrum. But in the long run, I'm very bullish on still, you know, computer science.

[52:55](https://realalpha.app/v/HJDekXH-h5c/?t=3175) Actually, ironically, I didn't study computer science, but I'm very bullish on studying computer science in general. Like if my kids wanted to do it, like I would definitely encourage them. What's the biggest mistake you've made in building mobile so far? I don't know. I think one mistake I made in the past is that I think sometimes like you kind of want to sequentially do things. You're like, well, first we need to do this. This is the blocker. We'll get to this other thing later because this other thing is blocked on this. But like in a world where like everything is moving fast like one of the things i've learned is like sometimes you kind of need to do both at the same time like you need to assume that like oh we're doing this thing to unblock it you might as well do this other thing at the same time just in the interest of moving fast that's something i i'm naturally like not mega aggressive which is kind of ironic because like we've run 15x in the last 12 months i sort of had to retrain my brain to like be a little bit more like aggressive and saying like this is a massive market opportunity we need to be even more, you know, like hiring, deploying capital, like whatever it is, like massively just like

[53:54](https://realalpha.app/v/HJDekXH-h5c/?t=3234) paralyzing everything we can. Sweden has a ton of interesting startups that have emerged. What are your thoughts on why that is? And has the Swedish house mafia been replaced by the Swedish tech mafia? Did you know that like one of the Swedish house mafia members was my classmate? It's a fun trivia. Really? I did not know that. Yeah. Back in the early 90s, growing up in Sweden. uh when i was eight he told me my bike was ugly i was very upset uh yeah i mean we've had a lot of tech in in sweden spotify klarna you know a bunch of gaming studios uh i mean more recently obviously lovable and agora stuff like that i think my theory about like a lot of this stuff is like success kind of just creates success like you know a lot of like what makes people going to tech is like they know about some cousin's friend who made a lot of money in tech or something like that so a lot of it is i think it's just like a little bit of a flywheel effect i do think that sweden as a european country has always been pretty entrepreneurial and commercial so i do think there's a little bit of that like we've always been a little sort of maybe going back to the vikings like kind of let's go conquer the world in the past maybe it's been through music

[55:05](https://realalpha.app/v/HJDekXH-h5c/?t=3305) and furniture and stuff like that but now increasingly with tech so you know there's that like I think there's sort of a mentality in Sweden like you can actually Stockholm is like funny because like everyone in Stockholm thinks they're like this like mega global city that's like part of the world in a way where like I actually think they aren't but like I think the fact that they think they're so global is like kind of good like they have this like vision that like everyone cares about Stockholm but I actually kind of like that because it means like they're out there like building you know things and they're saying from day one like we're going to conquer the world like we're going to do this startup it's going to be awesome and I like that like the sort of megalomania I know this vision of like your elementary school reunion and just what a wild group that would be.

[55:44](https://realalpha.app/v/HJDekXH-h5c/?t=3344) That's quite the cast in what I imagine was not a particularly large class. Yeah, yeah, it's kind of a random fact. I guess a shameless question. Do you think GP rental prices 12 months from now will be higher or lower? I think higher. You know, it's funny, like actually. changed my mind on this like in the last 48 hours uh and i'll tell you why i i read the anthropic s1 and i actually had a you know agent summarize it for me uh but it was kind of incredible how profitable that company is now it took to the degree that i hadn't anticipated i thought that like in my mind like 48 hours ago my argument was like oh anthropic like they have really good gross margins but they're spending so much money on training that like as soon as they go public they're gonna have to like slow down and how they grow training and post training and they're gonna you know maybe grow inference as much but but not the other stuff as much but then like i realized like they're basically a profitable company so like they have like zero incentive to slow down on like capacity um capacity acquisition and and i imagine there's a similar dynamic going on in open ai and so i in open ai and anthropic has obviously been like the biggest

[56:51](https://realalpha.app/v/HJDekXH-h5c/?t=3411) two buyers of capacity out there and so the fact that like i don't think they're gonna slow down buying you know uh because of you know clearly like financially they're doing very well i think it will mean that like there's going to be a lot more demand and and i think for that reason uh prices will probably go up on the next 12 months i've heard you say that you've always thought modal you've thought of modal as your 20-year project you're five years in um what do the what do years 10 through 20 look like wow uh i mean hopefully i'll hire a bunch of people that I can do a lot of the sort of, you know, I'm too in the weeds of like everything right now.

[57:28](https://realalpha.app/v/HJDekXH-h5c/?t=3448) So I, but I don't know, like, I just like love this stuff. Like, I love to keep working on this. Like, I think, you know, year 10 to 15, we're going to look more like a conglomerate. We're going to have a lot of different products. We're going to have probably some like crazy bets going to moon and Mars or whatever. But I still hope to be quite involved in this sort of the high level, like resource allocation and what are the things we're betting on. And I'd probably be less involved in sort of day to day, like, you know, let's get a contract from this GPU vendor, like whatever. But like, I think that like sort of like high level, like resource allocation and stuff like that, to me, that's the most fun.

[58:04](https://realalpha.app/v/HJDekXH-h5c/?t=3484) And I hope to stay involved in that for another 20 years, hopefully, maybe longer. Yeah, I mean, if we keep current pacing, at least five more unsupervised learning episodes in that time frame, right? Yeah, exactly. Well, Eric, thank you so much for joining us. We really appreciate it. will give you kind of the last call to action for the audience. If there's, where can they go to learn more about Modal or anything that you would direct them to? Well, you should obviously go to modal.com. You should also run pip install Modal in Terminal and play around with it. You know, there's a lot of people maybe listening to the show who are building startups or thinking about building startups and, you know, would love to see all the cool stuff they're going to build.

[58:44](https://realalpha.app/v/HJDekXH-h5c/?t=3524) And, you know, hopefully at Modal, we can help you with those startups. So let me know if you reach out, by the way. I always love talking to customers, but feel free to play around with modal and looking forward to see what everyone's going to build. I'm Jacob Effron, and this has been Unsupervised Learning, a podcast where I get to talk to the smartest people in AI and ask them tons of questions about what's happening with models and what it means for businesses in the world. As I hope is clear, I have a ton of fun doing this. It's a nights and weekends project, in addition to my day job as an investor at Redpoint.

[59:12](https://realalpha.app/v/HJDekXH-h5c/?t=3552) But our ability to get these incredible guests on really comes from folks like you subscribing to the podcast, sharing it with friends. It's really what ultimately makes this whole thing work. And so please consider doing that. And thank you so much for your support and listening. We'll see you next episode.