Steven Tuck

We raised a $200M Series C, now we're putting it to work. New compute sleds with the latest @AMDServer EPYC CPUs are coming and we're working with @XsightLabs on next-gen switch silicon with a fully open ISA. @TheRegister has the details: buff.ly/vmW0P6A

Our $200M Series C oxide.computer/blog/our-200m-…

Thrilled to share that we have closed our $200M Series C. Thank you to Thomas Tull's @USITfund for leading, along with @EclipseVentures, Riot Ventures, @JaneStreetGroup, @intelcapital, @counterpartvc, @friendsfamcap and all of our additional investors. buff.ly/ozCrZUE

Thrilled to be running Finance at @oxidecomputer! Thank you @sdtuck & @bcantrill for the opportunity. Oxide has 15 open positions right now with more on the way. Come build the future with us!

@geerlingguy @jessfraz @asnip8 We have hosted environments for this very purpose! Two studio racks in winter maybe, not summer.

@jessfraz @asnip8 @sdtuck If there was any STL area business that got one I'd love to check it out. I don't have room for a 2nd rack in my studio otherwise I'd set up a little oxide GeerlingCloud 😀

I would love @geerlingguy to get hands on this, I could request one for Demo Purposes but maybe @jessfraz can hook you up.

100% Committed. Let’s get to work!

This AI experiment will take a long time to finish. The opportunity for increased productivity is simply too strong. A consistent theme I've followed is that the market for creating foundational models is a fraction of the market for inference. What does that mean for semis? I have no idea. That's more Dan's space. What it means for enterprise IT is that your accelerated compute isn't going to look anything like the hyperscalers training foundational models or providing burst capabilities for fine tuning and inferencing. You'll need to think through command and control similar to how you operate other edge deployments.

Thank you SUSECON! Did you miss your chance for a hands-on demo? Sign up for access to your own slice of an Oxide Cloud Computer here: oxide.computer/remote-access

Get hands-on with an Oxide Cloud Computer this week only at SUSECON!

How long does your cloud provider take to create a Kubernetes cluster? Thanks to @oxidecomputer we were able to test on-prem scaling with Omni buff.ly/41dOiGO

⚾ 10x All-Star ⚾ All-time stolen base leader (1,406) ⚾ Single-season stolen base record (130 in ‘82) ⚾ Most runs scored (2,295) ⚾ 81 leadoff HRs ⚾ 1990 AL MVP ⚾ 1989 WS Champ ⚾ First-ballot Hall of Famer

Oxide Computer definitely wins the best website award this year. Also, look at this responsiveness, it's insane 💀

New @oxidecomputer site (oxide.computer) detail that I spent way too much time on. Connecting lines between dynamic elements -> fixed point on an image that work across different window sizes.

.@Livermore_Lab looks to make #HPC a little cloudier with Oxide's rackscale compute platform No cabling, nodes interconnected by backplane that provides power and 12.8Tbps of switching capacity Quote by @tgamblin in the @TheRegister article theregister.com/2024/11/18/lln… #SC24

Remembering Charles Beeler oxide.computer/blog/rememberi…

12x - improvement in cooling efficiency 2x - improvement in total compute density 55% - reduction in total power consumption Learn more in our latest blog entry okt.to/wGCKxB

Last week, AMD launched their Zen 5-based server CPU, Turin. Tomorrow on Oxide and Friends, @ahl and I will be joined by George Cozma of @ChipsandCheese9 to talk about Turin -- as well as the system we're building around it. Join us, 5p Pacific! discord.gg/QrcKGTTPrF?eve…

Infrastructure as code is dead. Welcome to infrastructure as models! I wrote about my experience using @thesysteminit to manage @oxidecomputer resources. Link below. 👇🏽

Reflections on Founder Mode oxide.computer/blog/reflectio…

This is exactly what we are hearing from our customers @oxidecomputer We are in the early days of a massive shift back to private cloud. Fortunately, this time you don't have to hire 100+ engineers and piece together imperfect fragments of an integrated solution from multiple providers. There is a better way and the team at Oxide has the answer for you: oxide.computer @bcantrill @sdtuck