We just crossed $100M in run rate.
Four quarters ago, we were at $1M. Then this year alone, we grew 37.5x.
Not long before that, we did a hard pivot- handed our old customers to competitors, rebuilt from scratch, and relaunched with a few months of runway left.
The bet was: models think on GPUs but work happens on CPUs.
So every agent that writes code, opens a browser, or runs an RL job needs a composable computer that's isolated, stateful, and ready in milliseconds.
$1M to $100M in about a year puts us in the same cohort as the likes of Anthropic and Cursor. It’s humbling to share that lane with companies we’ve learned so much from.
Show more
Ill never get a top tier 😅
Which startups are most likely to get an investment from top tier VCs?
I looked at ~60 attributes across 5000 startups and these were the 15 strongest.
#1# is being from a top tier fund. MBA is a negative signal
Full list of signals below:
Show more
Working on another secret project for @daytonaio 🧪
Any guesses what these are for?
Get it right and we might just send you one of whatever we’re making 👀
Credit is going to be the biggest constraint, not the GPUs themselves.
This will be the biggest trigger for consolidation in this ai wave.
Will be a fun 2027
I’m saying that in the medium term (next year or two), opencode will struggle to get their next clusters to scale, and their largest bottleneck will be credit. If i had to guess what @thdxr’s biggest problems in getting new clusters are, i suspect they would be the following:
- everyone is sold out
- all the new clusters have long lead times and no matter what dax wants to pay he can’t get it earlier
- everyone asks him for a large prepayment, say 30% to 100% on a long contract. If he asks for a shorter contract, the prepayment % goes up, not down.
- any cluster he wants is likely to be delayed
- he’s asked to bet his demand over the next several years, and if he’s wrong in one direction, the business either doesnt grow or blows up.
all of those problems are the same credit problem i am describing.
OpenCode’s incentives, regardless if it’s what they want to do, is to sign larger, long term contracts with credit worthy counterparties.
This is the same problem that @ivanburazin just hit with Daytona. Daytona isn’t IG, and the cloud they signed up with couldn’t get the debt, so that cloud came back and asked them for a larger prepayment, because the terms of the clouds debt wasn’t expected.
Show more
I migrated from e2b to @daytonaio a month ago.
I cannot explain how good their service is.
1. They gave @workflyai a grant 10K worth of credits + 40K coming after I run out on the first 10.
2. Instant slack-channel with the team. Super approachable and fast in replying.
3. I requested for an increase in Linux VMs capacity and they gave it in less than 10 minutes.
4. Support on Linux, Windows, and my understanding is that also MacOS.
5. LLMs (cursor, cc agents) are far more receptive on daytona's documentation. I guess because it has been created for agents.
If you are thinking about your sandbox provider, I really recommend daytona. I tried both, and I can say is far superior.
Show more
Dont dunk on people.
Everyone is going to get hacked...
Fuck around and find out.
When Railway was breached on 21 Oct, Northflank dunked on them and urged people to migrate over to them because of that breach.
9 days they got hacked, the same way.
Zero sympathy: instead of securing you were busy dunking
Show more
Thank you @daytonaio for the lovely gift and congrats on the series B :)
PSA: if the job description says on-site SF and not on-site/remote or hybrid, and you still don't want to move, please don’t send the message asking if it's possible to make an exception for you. Especially when I'm hearing from you for the first time.
I moved my family from Croatia to SF to build this company and there are people asking if they can fly in from Paris twice a year and make it work.
I'm sure remote/hybrid works very well for some of the most well-established companies out there, but most of us do well by coming to work together in one place. Totally fine if that's not for you but don't make absurd requests and waste either of our time.
Show more
People think of agents as " This will help me do X." That is correct but incomplete.
What they don't internalize is that the agent actually has to go out and do the thing.
It hits the same walls a human would. It needs access. So it needs credentials and tools to navigate systems that weren't built for it.
The agent can only do what you've equipped it to do.
Show more
We randomly selected 3 winners from eligible entries in the Mac Mini giveaway.
Congratulations to:
• @corylegendre
• @nitantbhartia
• @KayaHickin
Each of you will receive a Mac mini. We’ll reach out via DM shortly with next steps for shipping and verification.
Thank you to everyone who entered and shared what they’d build, the responses were great to see.
Show more
We have a few Mac minis left over from our Daytona event in San Francisco.
Thought we'd give 3 away to people building something cool.
To enter:
1. Follow @daytonaio
2. Reply with what you'd build with a Mac mini and tag 2 people who'd build something cool with one.
We'll randomly pick 3 winners and announce them on October 30th.
Winners will be contacted directly via DM.
Good luck!
No purchase necessary. US residents 18+. Rules below.
Show more
An agent won’t infer a pivot table
It won’t sit there and think through an entire spreadsheet in its head. It’ll just open Excel and execute the command. The same way you don’t navigate by memory anymore. You open maps and point.
The ratio of thinking to doing is not 1:1. It never was for humans either. So the CPU demand coming from agents at scale is going to dwarf what anyone is currently planning for.
Show more
Every time an agent executes, it offloads from the GPU to the CPU.
For one agent, that's fine. But what if you have to factor in a trillion agents doing that simultaneously?
The number of CPUs we will need to offload that work is going to be staggering. We're far from having that kind of supply.
Show more
My inbox right now is completely nuts. My EA and Grok Bot try to clear it and it just keeps filling back up.
Texts, whatsapps, emails, DMs, everything is flooded with congratulatory messages, sales/hiring pitches, partnership offers and whatnot.
I'm sure there are a bunch of genuine ones from my mutuals / our well wishers too
I'm really sorry if I haven’t been able to reply back to any of you. It’s a mess so I've just let it be for now.
Show more
state of AI startups in 2026
Daytona sandboxes are now built into AgentEnv, @scale_AI's open-source framework for RL environments.
Every environment and agent gets its own sandbox with Docker inside, so compose stacks, agent containers and image builds run unchanged.
Setup below 👇
Show more
Announcing the 3 winners tomorrow 👀
Follow @daytonaio and me so you don't miss it.
Stay tuned!
Gone.
4 x B200 nodes live today, if someone needs them. Hit me up.
Boom
Inspired by Daytona giving away an Nvidia DGX Spark supercomputer and $20,000 in cash if you're coming to Salt Lake City in November.
Join us:
We announced our Series B last week, and even my high school friends reached out to congratulate us :D
Founders should take some money off the table.
I read someone saying take 10% every round. That’s too much imo.
It should be enough that you don’t need the company to pay for your life but not so much that you’ve mentally cashed out.
If you're bullish on your company, even a small % seems like an opportunity cost.
Show more
