The Post-Earnings Tantrum, The $6B "Dilution" Myth, and The Unstoppable Neocloud

G'day fellow investors. The post-earnings reaction to IREN is textbook volatility: shares tanked on no new hyperscaler announcements and an EPS miss from the mining-heavy Q4 P&L.

But if you’re panicking over a backward-looking, non-cash accounting loss driven by a transition from ASICs to GPUs, you are missing the forest for the trees. Algorithms and surface-level analysts are still pricing IREN as a legacy Bitcoin miner.

Let's dismantle the panic, look at the reality of the recent $6B ATM announcement, and review why this period of choppiness is nothing to be worried about in the long term.

The Elephant in the Room: Misunderstanding the $6B ATM

The biggest boogeyman spooking retail recently has been the new $6 billion ATM (At-The-Market) offering.

Don’t get me wrong; a $6b ATM looks ridiculous relative to IREN’s current market cap of ~$13.5b. If they tapped the entire amount at today’s levels, the share count could increase by >40%. But that’s a massive “IF”, and most people are treating it like a certainty.

Let’s apply some logic. Why would IREN need to tap this ATM today? They don’t. They are fully funded for the Microsoft deal, having raised over $3b through converts, ~$1.93b of prepayments, and secured $3.6b of GPU financing at an incredible <6% interest rate. They don’t need a single additional dollar to execute the MSFT buildout.

Okay, what about the new ~$3.5b GPU order for 50k B300 units? First, those costs come due gradually through H2, on terms that are 30 days post-shipment. Second, management explicitly stated these will be financed primarily through the same sources they’ve used over the past quarters: prepayments, converts, and GPU financing.

So what’s the actual purpose of this $6b ATM? It is primarily a backstop. It establishes ultimate confidence for prospective hyperscalers that IREN can deliver on massive, multi-gigawatt deployments, strengthening their negotiating position by eliminating any doubts around capital availability.

So will IREN tap it eventually? Almost certainly at some point, but likely when the share price is materially higher. Net dilution from this ATM will most likely come in well below 20%, and could even be closer to ~8–12% (or less) if management stays conservative. Relative to the growth opportunity in sight, that’s a remarkably small price to pay.

Competition Heating Up: The Structural Moat (IREN vs. NBIS)

Competition is undeniably heating up across the sector (as seen with recent deals like NBIS / Meta), but it's crucial to distinguish between real structural profits and empty revenue hype.

https://www.cnbc.com/2026/03/16/meta-nebius-ai-infrastructure.html

Take Nebius (NBIS), for example. NBIS is bleeding cash structurally: gross expenses and depreciation consumed 110% of their quarterly revenues ($249.2m costs vs $227.7m revenue). Even adjusting their aggressive 4-year depreciation to an industry-standard 5 years, costs still eat 93.5% of revenue. These aren't one-off hits, they're structural costs that scale directly with revenue growth. You can't "grow out" of GPU depreciation or gross expenses; they stick like glue. Worse, NBIS allocates its pipeline to low-margin "bare metal" hyperscaler deals, commoditised bulk compute with zero pricing power. They have no moat and no path to profits.

Contrast this with IREN. The full hyperscaler contract transparency unlocks precise, bankable modeling. For the Microsoft deal, management discloses everything: $9.7B total value, $1.94B ARR, 85% EBITDA margins, $1.93B prepays, and sub-6% GPU debt terms. There is no black box. This visibility confirms IREN prints structural profits as a neo-hyperscaler (scaling to billions in net income), while peers like NBIS chase ARR hype with costs eating their revenue. IREN is the profitable one today.

Execution: The April Microsoft Deadline & Sweetwater

While Wall Street cries over EPS and potential dilution, the real story is operational execution.

IREN's Horizon 1-4 construction remains entirely on track to meet Microsoft's GPU deployment timelines. We are fast approaching the critical April/Q2 window where these clusters will begin coming online. At the same time, Sweetwater 1 (1,400 MW) remains on track to connect to the grid in Q2. Civil works for liquid-cooled data centers are already underway.

IREN is actively transforming its 810MW of operating data centers into a revenue engine that targets $3.4 billion in AI Cloud ARR by the end of CY26.

The Math: Wall Street’s Broken Spreadsheets

To understand how fundamentally mispriced this stock is, look no further than the valuation models. My Fair Value estimate on Simply Wall St remains at $95 per share, indicating the stock is 56% undervalued.

But here is the incredible part about that valuation: it is based on an ultra-conservative base case.

• The "Base Case" Model: To arrive at the $95.75 Fair Value, the valuator assumes IREN hits $8.7B in revenue and $2.9B in earnings by the year 2031 (representing a 63% p.a. growth rate), with a 10% discount rate, 33% profit margin, and a future PE of 25x.
• The Realistic Upside: However, based on the ramp of Childress, BC, and Sweetwater 1, IREN could very well hit $8.3B in revenue by 2027 and $12.7B by 2028.

Even if we assume it takes them 5 full years to reach $8.7B in revenue, the Fair Value still spits out $95.75. If they execute anywhere near my more bullish 2028 estimates, the upside is almost incalculable...

Risks

Dilution risk

While I remain highly convicted, we have to remain objective. If I am wrong and management taps into that $6B ATM sooner and heavier than expected, it will definitively impact the share price in the short-to-medium term. Ongoing dilution is a legitimate long-term risk for this stock. However, in an aggressive land grab for AI infrastructure, this capital is exactly what allows the company to keep scaling, secure massive power blocks, and capture market share while peers stall. It is a calculated tradeoff. This is why we've seen the more short term oriented investors and traders locking in profits in the past weeks.

Macro Uncertainty & The Energy Crisis

We can't analyse a power-hungry infrastructure play without addressing the macro environment. With the escalating war in Iran and the broader Middle East, the threat of a severe global energy crisis is very real. While IREN has deeply advantaged, contracted power, a prolonged structural spike in energy markets could impact underlying grid costs and temporarily squeeze operational profitability. Energy is their primary input, and global shocks matter.

Embracing the Volatility

Between the ATM headline fears, the anticipation of the next hyperscaler deal, and severe macro/geopolitical instability, the rest of this year is guaranteed to be highly volatile. But volatility is the price of admission for multi-bagger returns. If macro fears drag the broader market down and create strong dips in IREN's share price, I won't be panicking. Instead, I’ll be ready to buy more.

The Takeaway

IREN’s lead in secured power capacity continues to widen, punctuated by their massive new 1.6GW site in Oklahoma. Over time, the latent value of this >4.5GW portfolio will become obvious, even if it doesn't feel that way today amid the post-earnings volatility.

With Sweetwater energising in just a couple of months, it’s likely that IREN will sign a new hyperscaler deal in the next 3-6 months. When that happens, expect a massive re-rate as Wall Street scrambles to upgrade their idiotic projections.

Still holding, ignoring short-term noise, focused on the long term.