York morning trade.
We are looking at shares of Cumulus AI pulling back by nearly 8% following a 34% surge yesterday.
Now Cumulus AI is a cloud infrastructure as well as high performance computing company often described as a neo cloud provider that went public last week.
Now shares jumping yesterday as it signed a two year take or pay agreement to supply Nvidia Blackwell B 300.
To a GPU cloud marketplace with the contract valued at $18 million now instead of going after hyper scales, the company does focus specifically on lower costs, high efficiency compute for small to mid-market AI startups as well as research institutions and enterprise AI deployments and we will hear from the company's CEO right here on market movers next week.
Well joining us this morning to weigh in is even.
Dickens, CEO and principal analyst at Hyper Frame Research.
Steven, good morning.
Thank you so much for joining us.
So starting off a look at cumulus AI's neo cloud model, which does represent a complete shift away from traditional gem purpose cloud computing.
Give us an overview of this and do you think this fast track approach gives them a real lasting edge against power shortages as well as delays out there.
Hey Remy, thank you for having me on the show.
I think the way I look at Cumulus, I've been working with this team and spending time with Mike Manascolo, the CEO over the last few months since he joined.
Really where I see this company going is they've got a more modular model, um and also some very innovative ways to finance that modular model, and we're seeing inference moving out to the edge, so this is a company that's making a hard pivot from being a Bitcoin miner legacy into being this neo cloud provider with these innovative models for both finance and deployment.
And I think that's starting to resonate, they've signed deals, um, that you mentioned there in the in the introduction, they've also um made a commitment this week to purchase, I think it was 1600, over 1600 Nvidia B300s.
This is a company we're seeing some fluctuations.
They came to market through a direct listing last Thursday, so we're seeing some price discovery kind of in the stock, but I think long-term and structurally I think there's a lot to be sort of positive and bullish about with this company going forward.
And companies are shifting from building AI to actually running it 24/7, which we all know can get very expensive.
So do you think the company can run everyday tasks cheaper and more efficiently than big tech?
So I think the the approach here I touched on it briefly a few moments ago, is this modular approach.
Well, what we see with some of the other hyperscale is is putting big infrastructure into huge data centers.
Cumulus is taking a slightly different approach and putting that inference closer to their end customers in a more modular model.
I, I see this as a really positive way to drive out some of that token economics, and be able to bring that infrastructure closer to the customer so, To answer your question, I think this is an approach that's gonna find its base in the market.
I think there's a, there's an approach for all of the different, um, players that we're seeing.
We're seeing Korweave, we're seeing Volta, we're seeing Crusoe, we're seeing a bunch of others, they're all growing.
I think Humulus is gonna find its place amongst those names with its innovative model and just be able to service customers in a different way.
And the company is stacking Nvidia chips, so Cumulus does give companies direct access to Nvidia's latest chips about forcing them to use some of that software, but are tech teams actually choosing this over traditional cloud giants?
Well I think it's that optionality, people are looking for different, different angles, they're looking for being able to get access to these chips, being able to get access to them in a timely way.
They're also thinking about round triple agency on some of these transactions and some of these queries, being able to have options that are closer to where the workload needs to run, rather than going off to a, a large mega scale data center.
Um, hundreds and maybe even thousands of miles away being able to do that on a more modular level, so I think, do I see this as an either or?
No.
Do I see this as a way to bring optionality that's gonna resonate with some of those enterprise buyers?
Absolutely.
The Cumulus team's got that flexibility that I think some of those enterprises are gonna look for, and we just see explosive burgeoning demand and Cumulus can ride that wave.
And Steven, finally, before I let you go, we have less than 60 seconds here, so what is your take on IBM's earnings report?
So I think Arvin made a really fantastic, um, and unprecedented early announcement, so there was no real surprises.
Key things I took away was that those deals that slipped out of the quarter, over a third of those are already closed in the 1st 3 weeks of the next quarter.
I think we all should know that the mainframe cycle is cyclical.
They're 5 quarters into Z17, so I don't think there's any surprises if you actually track this stock and you actually know what's going on.
There's no surprises here.
I think long-term Arvin's strategy of mid single digits growth is above where expectations would have been 3 or 4 years ago, so I, I see this as long-term good, just a short term blip.
OK, Steven, well, thank you so much for joining us this morning.
I appreciate your time as well as all of your insights.
Thanks, Remy.
