AI Infrastructure Buildout
Hey traders,
If you’ve been following the AI boom, you may know that the big hyperscalers (AWS, Azure, Google Cloud) can’t keep up. Demand for raw GPU compute and the infrastructure required to run high-level AI models is expanding exponentially as more companies and individuals adopt AI. Multiple large companies, including Meta, Amazon, Dell, and Block, have already come out this year and stated that they are using advanced generative AI rather than hiring more workers, and in some cases, they are also cutting thousands of workers as well.
This is only the beginning of larger AI adoption. Usage will continue to explode across a variety of use cases, from complex agentic AI large language models capable of high-level reasoning and coding to robots like Optimus and autonomous cars already available in cities like Austin and Las Vegas.
The Neocloud
That’s exactly where the Neocloud comes into play.
Neoclouds are companies that build and run data centers packed with thousands of the world’s most powerful AI chips (mostly Nvidia GPUs). Their only business is renting time on those chips to AI companies so they can train and run massive models. They don’t offer websites, storage, or email like the old clouds did. They exist because the hyperscalers simply don’t have enough capacity or the specialized infrastructure required — and the gap is measured in billions of dollars in signed contracts.
Danielle Shay Joined Making Money on Fox Business with Charles Payne
Yesterday, I joined Charles Payne on Fox Business to discuss it with him.
Watch my Fox Business segment here:
I told Charles the same thing I’m telling you today: this infrastructure build-out is still in the very early innings. Agentic AI + smarter LLMs + massive incoming demand = exponentially more compute needed. The companies that are first in line to handle this workload are where I am focusing.
Funded by the Magnificent 7 Names
- Hyperscalers are writing $2B–$17B checks to these names because they can’t build fast enough.
- We’re watching the largest infrastructure spend in human history (Nvidia CEO Jensen Huang highlighted this on Monday, March 16th, at Nvidia’s GTC event).
- Many of these stocks still trade at fractions of their projected 2027 revenue — I think largely because it’s so unproven at this point, and there is a lot of speculation around it
My Current Neocloud Focus
Here are a few focus points from the varying names I have selected:
Pure Neocloud Tier
- NBIS (Nebius) – The standout. Never touched Bitcoin. Spun out of Yandex with elite engineers and $2.5B cash. Already sold out for Q1, $17B+ Microsoft deal locked. Wall Street’s favorite “CoreWeave 2.0 with a tech edge.”
Previous Bitcoin Miners (companies that already have the setup, but are now switching their models)
- IREN (Iren Ltd) – Green energy king turning mining sites into AI factories. Microsoft deal + Nvidia Blackwell access.
- APLD (Applied Digital Corp) – Owns cheap power, just landed massive hyperscaler leases. Revenue was up 250% last quarter.
- CIFR (Cifr Digital Inc.) – Aggressive pivot from mining to $9B+ AI/HPC leases. Rebranding and converting facilities right now.
Bottom Line for Investors & Traders
This isn’t 2023 crypto hype. These are real multi-year contracts, real revenue ramps, and real scarcity in a growing space. Yes, earnings volatility is real (we saw it with CRWV and the pivots), and the high short-float percentages demonstrate that investors continue to give this space the stink eye. However, that’s why this is a longer-term plan for me.
I will continue trading these names on a short-term basis, especially when they break out above previous highs and short squeeze, and I’ll hold them on a long-term basis unless they break critical support and something fundamental changes. For example, I use the 50 SMA on the weekly chart as a key technical stop loss zone.
Happy trading,
Danielle Shay
@traderDanielle
VP of Options @ Simpler Trading
Fox Business | CNBC | StockCharts
