I think people are finally starting to understand why $IREN has operated the way it has. IREN took a lot of criticism for dilution. Maybe the timing was not perfect, but look at the sector now. WYFI, BTDR and NBIS are raising capital too, yet people are finding ways to call it bullish. When IREN does it, it is dilution. When others do it, it is funding growth. You cannot have it both ways. What matters is what shareholders get in return for that capital. The same applies to contracting. IREN spent years building the infrastructure first and has been willing to wait until capacity gets closer to delivery rather than rushing into deals. That gives them a stronger position to negotiate better economics. Then there is power. IREN has over 5 GW of secured power. This is very different from announcing a multi-gigawatt development pipeline that still depends on interconnection, transmission, permitting or other approvals. IREN got plenty of criticism for its strategy. Now we are watching much of the sector follow the same playbook.

"The Pentagon is now seeking $21 bn for drones/counter-drone capabilities, but the answer isn’t just spending more. We need layered defenses that can detect/track/jam + destroy cheap drone swarms without firing million-dollar missiles at thousand-dollar threats." = $ONDS

I think consolidation across the data center sector is inevitable over the next few years. There are simply too many companies sitting on valuable power portfolios for every one of them to independently become a major AI infrastructure company. Look across the former Bitcoin mining sector: $IREN $RIOT $MARA $CLSK $KEEL $WYFI and several others. These companies spent years securing land, power and electrical infrastructure for Bitcoin mining. AI has completely changed what those assets can potentially be used for and how valuable they are. But owning the power and developing it are two completely different things. Turning 500 MW or 1 GW into an AI campus requires billions of dollars. You need data centers, substations, cooling, networking, GPUs and everything else required to actually bring that capacity online. $WYFI is a good example. The company has a large development pipeline, but it has just announced a proposed $250 million convertible note offering. The stock has fallen heavily over the last two days following recent financing announcements. This sector requires an enormous amount of capital, and not every company sitting on valuable land + power has the balance sheet to develop it themselves. That is where I think consolidation comes in. A hyperscaler, neocloud or large infrastructure company has access to billions of dollars of capital. Instead of spending years finding suitable land, securing power, entering grid queues and going through the permitting process, acquiring an existing company or individual powered site could make far more sense. And the value of already secured power should only increase if bringing new capacity online becomes more difficult. Pennsylvania is a good example of what is starting to happen. More than 100 data center facilities have been proposed across the state, and Pennsylvania has now introduced stricter requirements for large developments. Developers need local approval, large projects must meet the state's GRID requirements before certain permits can move forward, and data centers have been removed from the Permit Fast Track program. This is also why investors need to pay attention to how companies describe their GW numbers. Secured power and a development pipeline are not the same thing. A company can have several gigawatts in its development pipeline without having secured the underlying power. Getting from a proposed site to an energized data center can take years, and there is no guarantee every project gets there. $IREN is a good example of the difference. IREN has historically been very clear about separating secured power from the wider opportunities it is pursuing. It spent years securing land and power before AI data centers became the opportunity they are today, and it has now shown with Horizon 1 that it can convert that infrastructure into actual AI capacity. That helps explain why IREN has separated itself so dramatically from the rest of the former Bitcoin miners. A few years ago these companies were largely valued on Bitcoin production, hash rate and mining economics. Now the market is starting to place a much higher value on something they accumulated along the way: land + secured power. I don't think every company on that list will independently make the transition. Some will build. Some will partner. Some will sell sites. Some will be acquired. There is a huge amount of valuable infrastructure sitting across the former Bitcoin mining sector. The companies with the capital and execution to develop it will become much larger. The ones that don't may still be sitting on exactly the assets somebody else wants to buy.

$WYFI is getting absolutely crushed. -10% yesterday -20% overnight The company announced a $250M convertible offering, with an option to increase it to $287.5M. That's roughly 25% of its pre-drop market cap. It's a huge raise for a company this size, so I understand the reaction, but these are convertible notes rather than a straight equity offering and much of the capital is going towards its data center expansion. This could turn into a great buying opportunity depending on the final terms.

$IREN & $NUAI - This is why grid-connected power is one of the biggest moats in AI infrastructure, and more people are starting to realize it. This is extremely bullish for companies like IREN that have already secured gigawatts of power. It is also extremely bullish for NUAI because its focus on behind-the-meter power means it does not have to rely entirely on the grid. As grid connection queues get longer, being able to develop power directly alongside data centers gives it a major advantage.

AI compute is not a winner-takes-all market. There will be several winners. You can pick the company you like most or buy $NCLD ETF. Top 10 holdings: $NBIS 33.27% $CRWV 27.20% $IREN 7.43% $HUT 4.94% $APLD 4.29% $WULF 4.06% $GLXY 3.98% $CIFR 3.44% $RIOT 3.24% $CORZ 3.06%

That is insane for a company founded in 2021. To put it into context, Apple generated $416 billion in revenue in FY2025, while Anthropic is now projecting $190–200 billion by 2028. That would be nearly half of Apple’s current annual revenue.

I think $CIFR, $BTDR and $KEEL are great acquisition targets. They control large power portfolios, valuable sites and grid connections that would take years to secure. For a HS, buying one would be far quicker than building the same pipeline from scratch.

$IREN - Jensen Huang is basically confirming what I have been saying for a while. The biggest bottleneck in AI infrastructure is land, power and shell. NVIDIA is now willing to help secure these resources because compute demand is growing faster than infrastructure can be built. This is extremely relevant for companies like IREN that have spent years securing large power positions. The GPUs will change every few years, but the land, power and grid connections can support generation after generation of compute. This is exactly why I believe IREN's >5 GW power portfolio is so valuable. Stock +5% today. It's only matter of time.

I thought this was an interesting slide from Piper Sandler showing the new AI infrastructure stack. $IREN sits in Layer 1: Power, Grid & Real Estate. $CRWV and $NBIS sit further up the stack in both Layer 6, Orchestration & Workloads, and Layer 7, GPUaaS & Cloud. What’s also worth noting is that IREN has now acquired Mirantis, which Piper places in Layer 6. So IREN now has exposure at both ends of the stack, from the physical infrastructure at the bottom to orchestration and workloads further up. Source: Piper Sandler & StockSavvyShay

$IREN & $NBIS are clearly the top picks in the compute sector, but the opportunity is far too large for one or two companies. I believe there will easily be 10+ winners because the amount of power needed to meet compute demand over the next decade will be astronomical.

Trump came to save the day. $IREN, $NBIS, $CIFR, $NUAI Politics aside, Trump is right here. Data centers are current economic engines. They bring billions of dollars of investment into a state, create thousands of construction jobs, hundreds of permanent high-paying jobs, generate long-term tax revenue, and attract other businesses to the area. AI companies will build where they can secure power, land, and permits the fastest. If one state makes it difficult to build, that investment simply moves somewhere else.