Jacob Zachariasen

Vejle Boldklub - Happy amateur small time tech investor, tying to get a grasp of things

[Jacob Zachariasen @Nephdk]

I agree with the core idea here.

The $IREN thesis should not stop at “power and land.” That is the entry ticket. The deeper question is whether IREN can turn scarce power into efficient, high-utilization, revenue-generating AI compute.

That is where vertical integration may matter: data center design, rack density, liquid cooling, network topology, DSX alignment, Mirantis/k0rdent, orchestration and operational learning loops.

If IREN can compound those capabilities across sites, the moat becomes much deeper than pre-positioned power alone.

Where I would be more cautious is on the strongest claims.

“Only IREN”, “DSX flagship”, or “2GW becoming 4-8GW equivalent compute” still need proof. We need benchmarks, deployment data, utilization, uptime, margins and real customer outcomes.

So I agree with the direction:

Power is the foundation.
AI factory execution is the real test.

Bullish, but the numbers still have to validate the vision.

[Jacob Zachariasen @Nephdk]

@FransBakker9812 Im totally in sync with you here. Nice view by the way. Enjoy your sunday

[Jacob Zachariasen @Nephdk]

I don’t think this has to be a one-winner market. $CRWV, $NBIS and $IREN can all win in different parts of the AI cloud stack.

$CRWV has the revenue engine. $NBIS has a strong balance sheet and a clean AI cloud story.

But I lean $IREN because I think the real bottleneck is moving further down the stack: power, land, grid access, substations, cooling and time-to-compute.

Owning the physical layer matters.

If IREN executes, the combination of secured power, Microsoft, NVIDIA, Dell, DSX alignment, Mirantis and global expansion gives it a very asymmetric setup.

Still needs to prove ARR, utilization, margins and cash flow.

That is why I’m long.

[Jacob Zachariasen @Nephdk]

Fair question.

Execution risk is always worth discussing with a company scaling this fast. But so far, I agree that the evidence points more to progress than missed deadlines.

Microsoft financing closed. Dell/Blackwell procurement moving. DSX Air validation. Sweetwater progress. NVIDIA alignment. Mirantis added.

The hard numbers still need to follow: ARR, utilization, margins and cash flow.

But I am long $IREN because the execution signals so far keep moving in the right direction.

[Jacob Zachariasen @Nephdk]

@mikealfred

[Jacob Zachariasen @Nephdk]

$IREN today

[Jacob Zachariasen @Nephdk]

Goldman’s latest piece on US data center power demand is worth reading.

For me, the key investor takeaway is simple:

AI does not scale without power.

That is why $IREN remains such an interesting case. The market can debate AI cloud models, valuation, timing and execution, but power, land, grid access, substations, cooling and time-to-compute are becoming strategic assets.

This is exactly where IREN’s infrastructure-first model becomes relevant.

But the article also points to a broader opportunity.

The winners may not only be the AI cloud operators. The supplier layer behind them could matter just as much:

Power generation: $VST, $CEG, $GEV
Electrical infrastructure: $ETN, $SU.PA, $VRT
Grid and EPC: $PWR, $EME, $MYRG
Cooling and thermal: $VRT, $TT, $MOD
Cables and transmission: $PRY.MI, $NEX.PA

$IREN is one way to play the front line of AI infrastructure.

But the picks and shovels behind IREN, CoreWeave, Nebius, Microsoft, Dell/NVIDIA AI Factory and sovereign AI projects may become a major investment theme of their own.

Power is no longer a background input.

It is becoming part of the moat.

https://t.co/h3SAXJCMpy

[Jacob Zachariasen @Nephdk]

This is exactly why I am long $IREN.

There will always be religious debates about which AI cloud model is best. Asset-light, software-led, hyperscaler-style, infrastructure-first.

I get the debate.

But I believe in IREN’s model.

Own the physical infrastructure.
Control the power, land, grid access and cooling.
Deploy the compute.
Then build the software and operating layer on top.
In AI, time-to-compute is becoming one of the biggest bottlenecks.
That is why I think IREN is different.
The numbers still need to follow. ARR, utilization, margins and cash flow will matter.

But I am not here for a quick trade.

I am long because I believe IREN is building something that can matter for years.

[Jacob Zachariasen @Nephdk]

#vejleb Jeg er positiv.Marcussen er en god signing. Dejligt at Hemmersam er tidligt ude af starthullerne

[Jacob Zachariasen @Nephdk]

$IREN still needs to prove the hard numbers: reported ARR, utilization, margins, cash flow and consistent execution.

That is the real test.

But it is also getting harder to ignore the soft datapoints stacking up: Microsoft deployment progressing, NVIDIA alignment, Dell AI Factory visibility, Sweetwater energization, Mirantis, improved funding structure and stronger institutional attention.

None of these individually prove the full thesis.

Together, they suggest something meaningful is building.

Bullish, but still watching the numbers.

[Jacob Zachariasen @Nephdk]

@Superligaen Ja tak @Vejle_B dejligt at få ro på netop den plads. En fuldgod substitut for Vekic.

[Jacob Zachariasen @Nephdk]

Fair point, and I agree that the disconnect between company guidance and analyst expectations is a big part of the setup.

So far, execution has clearly moved in the right direction: financing closed, GPUs arriving, Microsoft deployment progressing and ARR targets laid out.

Personally, I have strong confidence that the numbers will follow if IREN continues executing at this pace.

My only caution is that the market will eventually need to see those targets convert into reported revenue, utilization, margins and cash flow.

That is where the real rerating can happen.

[Jacob Zachariasen @Nephdk]

Congrats to the $IREN team on closing the $3.0B convertible notes offering.

A well-structured capital raise: long-dated, low coupon, full greenshoe exercise and capped call protection up to $110.30.

For shareholders, the key point is that the funding question is becoming less central. The next chapter is execution: turning capital into deployed compute, utilization, ARR, margins and cash flow.

Well done — now the real work begins.

[Jacob Zachariasen @Nephdk]

The neocloud wars are getting religious.

$IREN, $NBIS, $CRWV, $APLD, $CORZ, $WULF…

Every camp has its believers, skeptics, prophets and sinners.

But maybe we can all agree on one thing:

No matter who wins, they all need power, land, buildings, cooling, racks, fiber, transformers, switchgear and contractors.

So let’s call a temporary ceasefire.

Which listed “picks & shovels” companies are best positioned to supply the AI datacenter buildout?

This should be one of the few places where all tribes have a shared interest.

Drop your best ideas 👇

[Jacob Zachariasen @Nephdk]

Impressive quarter from $ONDS.

Revenue acceleration, backlog growth, raised guidance, strong gross margin and a much stronger balance sheet all point in the same direction: this is no longer just a promising autonomy story — $ONDS is starting to look like a real defense/autonomous systems platform.

The most encouraging part is not just the headline growth, but the broader execution across C-UAS, robotics, ISR, strategic acquisitions and the Palantir/SkyWeaver roadmap.

Still plenty to prove on integration, cash efficiency and converting backlog into profitable revenue — but this was a meaningful step forward.

[Jacob Zachariasen @Nephdk]

Strong framework, and I agree with the main point: $IREN financing story looks far more structured than the simple “capex funded by dilution” narrative.

Customer prepayments, GPU-backed financing, low-coupon converts, ATM capacity and strategic equity alignment all reduce the funding-risk overhang.

That said, I’d separate confirmed structure from inference. The Microsoft setup is tangible. Assuming NVIDIA or future deals follow the same playbook is plausible, but still optionality until confirmed.

For me, the key question now shifts from “can IREN fund the buildout?” to “can they turn this capital stack into deployed compute, utilization, ARR, margins and cash flow?”