Building @cursor_ai and @bot. Prev @tryramp @meta @fluidstack @microsoft
Joined September 2022
- Tweets84
- Following125
- Followers486
- Likes158
Just getting started. Trust me.
A nice moment for the team this week: @nasdaq featured Fluidstack’s $830M Series A on their Times Square tower.
We’re hiring across all departments: fluidstack.io/jobs
1
9
1,981
Openai agents at 10m users. the bottleneck was never demand... it's whether infra can keep pace
We're officially in the day and age where whether there's value in doing something with AI or not. We want to try. We want to spin off ai agents and ask them to try tasks for us. Ask them to solve problems for us. Ask them to do research on things we have questions about and give us answers.
It doesn't mean its always correct, but we're in a new dawn of the information age. The demand is greater than ever.
bloomberg.com/news/articles/…
6
663
We are well capitalized. We're trying to hire some of the best and brightest to help us on our mission.
If you think you could help. Please dm me or check out our jobs page!
In January, Fluidstack raised an $830M Series A at a $7.5B valuation, led by Situational Awareness, with participation from world-leading investors.
Fluidstack builds infrastructure for the leading AI labs - our goal is to be the fastest on the planet at deploying hundreds of gigawatts of compute.
As we stand at the event horizon of the singularity, humanity’s best chance is if democracies, with error correcting institutions, free speech, and checks on power, imbue those same principles into superintelligence. Whoever deploys frontier compute infrastructure fastest will decide whether Al expands human freedom or shrinks it.
We are hiring - come join us.
compute keeps being the bottleneck: kimi k3 paused signups in 48h because they ran out of gpus, not ideas.... nitter.gq/Kimi_Moonshot/status/2…
Introducing Kimi K3: Open Frontier Intelligence
🔹 2.8 Trillion Parameters, 1 Million Context, Native Multimodal
🔹 Kimi Delta Attention enables up to 6.3x faster decoding in million-token contexts
🔹 Attention Residuals deliver ~25% higher training efficiency at <2% additional cost
🔹 Built for long-horizon agentic coding and self-evolving workflows
Kimi K3 is now live on on Kimi.com, Kimi Work, Kimi Code, and the Kimi API.
Open Weights by July 27, 2026.
🔗 API: platform.kimi.ai
🔗 Tech blog: kimi.com/blog/kimi-k3
1
246
Sakthi Sankarraman retweeted
👀
A dinner on the future of AI infrastructure, Tuesday 7/21 in SF hosted by @fluidstack
We’re building datacenters the size of Central Park in months, not years, and hiring for basically everything.
Come eat and find out. Space is limited.
luma.com/osffvtq0
Hey come to dinner with us in SF! We would love to show you what we are working on!
A dinner on the future of AI infrastructure, Tuesday 7/21 in SF hosted by @fluidstack
We’re building datacenters the size of Central Park in months, not years, and hiring for basically everything.
Come eat and find out. Space is limited.
luma.com/osffvtq0
2
3
342
Sakthi Sankarraman retweeted
Our latest hackathon winner: RackScout, an AR copilot for the data hall. Put on the headset and the hall annotates itself - a glowing path to the rack you need, checklists pinned to the hardware, racks turning green as they go production-ready.
I’m looking to hire 5 fantastic software engineers to work on this project with me. If you want @AnthropicAI to be your biggest customer. Dm me. Must be US based.
anthropic.com/news/anthropic…
Compute capacity is heating up! More and more investment from big and small players.
TeraWulf Announces Anthropic Lease at Justified Data Campus and Sale of Majority Interest in Abernathy Joint Venture to Fluidstack 🚀
investors.terawulf.com/news-…
4
293
Jets have started to arrive for @allenandcompany annual conference in @sunvalley.
All the leaders in the AI space will be here from @sundarpichai of @Google and @finkd of @Meta. @Apple new ceo John Ternus, @DarioAmodei of @AnthropicAI, @JeffBezos of @amazon and @Prometheus and @sama of @OpenAI.
@netflix @Disney @warnerbros and ton of media moguls will be there as well.
We can expect a lot of deals to get done this week behind closed doors.
2
9
51,386
Please come work with @ArjChi and I. We’re building the future and we need help 🙏
we are hiring for many engineering roles
electrical, mechanical, SWE, MTS, TPM, and more
apply below or dm me.
if you don't see a role that fits you exactly, still dm me.
we make roles for exceptional talent.
fluidstack.io/jobs/c0457b1b-…
4
183
This is really well said. Getting the whole orchestra together to play music is difficult and if anyone backs out the whole symphony falls out of rhythm.
I also think this calls out there needs more competition for chips and different kinds of chips!
I've seen people on X talk about how this is a signal that demand is not as unconstrained as analysts make the market out to be - but I believe this is a backwards read
Building a NeoCloud, especially a GPUaaS one, is a lot like baking a cake where you need all of the ingredients to come together at once:
1. A viable site with power
2. A customer to offtake the capacity or a backstop agreement to provide contractual support
3. Financing for said equipment
4. Equipment delivery dates that meets customer and site timelines
Financing is the ingredient people underrate. Lenders want a collateral base, and they treat these offtake contracts as something they can securitize (think SaaS ARR-based lending). No contracted demand, no financing
Arrangements like this enable NeoClouds to meet insatiable demand by accessing more financing - this article speaks to that supply-side challenge, not demand weakness
theinformation.com/articles/…
4
6
183
Selling compute capacity has become an entirely new business line for many companies
We see that @Meta has declared they're going to sell their compute capacity, allow customers to use models on their platform, and charge per token.
We have @awscloud @Azure @googlecloud @OracleCloud selling a variety of GPU compute, with #awsbedrock offering a bring your own model approach.
We also have the #neocloud @CoreWeave @crusoe
@fluidstack @nebiusai @nscale who are building out and selling compute capacity.
The most important thing to notice is that, is that there is competition in the market place. Customers will have options. Innovation will push all the players to become more energy efficient, work with multiple chip manufacturers like @Broadcom, @nvidia, @AMD, and custom silicon like @amazon Trainium and @Google TPUs.