[GUEST ACCESS MODE: Data is scrambled or limited to provide examples. Make requests using your API key to unlock full data. Check https://lunarcrush.ai/auth for authentication information.]  Kamahsutra [@TheKamaHsutra](/creator/twitter/TheKamaHsutra) on x 4133 followers Created: 2025-07-15 16:26:47 UTC They do have a small fleet of GPUs that is at or near full utilization. Their fleet is going to more than 2x, but not with more Hoppers that a lot of the current neoclouds have a high % of in their fleet inventory. $IREN will be 2400 Blackwells and 1900 Hoppers, that is XX% Blackwells. I would like someone to find me another neocloud that has that type of fleet ratio. Most of the other neoclouds won't even disclose their full fleet but we know they're full of Hoppers. Hoppers were easy to deploy because there was very little barrier to deployment into existing DC infrastructure. However, this will not be the case with Blackwells. $IREN is the only one right now that can do 80kw rack density with air-cooling for B200s in their existing DCs in BC, this is an enormous capex savings in running Blackwells. They will also retrofit liquid cooling for B300s and higher density B200s in X rack configurations when the time comes. It's ironic that when $IREN talks about rack density, and when the IREN bulls have been providing a lot of data and pounding on the importance of data center design from first principles it largely gets dismissed. But when @GoldmanSachs put outs a deck for $NBIS that highlights the added complexity of deploying Blackwells as barrier to entry, their stock starts to pump. Has GS Research been reading our posts and listening to our spaces anonymously? 🤔 If you're looking for a neocloud that can grow the fastest fleet of Blackwells, with the lowest blended in capex cost, I highly doubt you can find someone who can beat $IREN at this game. Yes, this will require a lot of capital, but this also applies to any other player. However, do consider that $IREN gets a huge amount of FCF to fund their capex in CSP growth from $BTC mining. This will be over $50mil (should be over $60mil this month) per month running at XX EH for a full month. The other neoclouds are still struggling to find any margin of profitability to use as FCF at this point. $IREN naysayers keep pointing to having no proprietary software stack. However, $NVDA DGX Cloud Lepton solves a lot of this problem, while IREN develops more of their proprietary software stack, if needed. We keep saying, IREN has the most important layer locked down: the base layer. To build anything, you need the strong foundation.  XXXXX engagements  **Related Topics** [iren](/topic/iren) [$iren](/topic/$iren) [Post Link](https://x.com/TheKamaHsutra/status/1945158111213330833)
[GUEST ACCESS MODE: Data is scrambled or limited to provide examples. Make requests using your API key to unlock full data. Check https://lunarcrush.ai/auth for authentication information.]
Kamahsutra @TheKamaHsutra on x 4133 followers
Created: 2025-07-15 16:26:47 UTC
They do have a small fleet of GPUs that is at or near full utilization. Their fleet is going to more than 2x, but not with more Hoppers that a lot of the current neoclouds have a high % of in their fleet inventory. $IREN will be 2400 Blackwells and 1900 Hoppers, that is XX% Blackwells. I would like someone to find me another neocloud that has that type of fleet ratio. Most of the other neoclouds won't even disclose their full fleet but we know they're full of Hoppers.
Hoppers were easy to deploy because there was very little barrier to deployment into existing DC infrastructure. However, this will not be the case with Blackwells. $IREN is the only one right now that can do 80kw rack density with air-cooling for B200s in their existing DCs in BC, this is an enormous capex savings in running Blackwells. They will also retrofit liquid cooling for B300s and higher density B200s in X rack configurations when the time comes.
It's ironic that when $IREN talks about rack density, and when the IREN bulls have been providing a lot of data and pounding on the importance of data center design from first principles it largely gets dismissed. But when @GoldmanSachs put outs a deck for $NBIS that highlights the added complexity of deploying Blackwells as barrier to entry, their stock starts to pump. Has GS Research been reading our posts and listening to our spaces anonymously? 🤔
If you're looking for a neocloud that can grow the fastest fleet of Blackwells, with the lowest blended in capex cost, I highly doubt you can find someone who can beat $IREN at this game. Yes, this will require a lot of capital, but this also applies to any other player. However, do consider that $IREN gets a huge amount of FCF to fund their capex in CSP growth from $BTC mining. This will be over $50mil (should be over $60mil this month) per month running at XX EH for a full month. The other neoclouds are still struggling to find any margin of profitability to use as FCF at this point.
$IREN naysayers keep pointing to having no proprietary software stack. However, $NVDA DGX Cloud Lepton solves a lot of this problem, while IREN develops more of their proprietary software stack, if needed. We keep saying, IREN has the most important layer locked down: the base layer. To build anything, you need the strong foundation.
XXXXX engagements
