David

Update on $IREN Investment Position Following Q2 FY2026 Results I believe the Iren sell off is overblown and below I will describe why and how we should actually look at this business. With a sell off this hard you should always ask yourself "why did I invest in this company in the first place". Remember we own shares of a company, not a number on a screen. My Investment Thesis: $IREN represents a compelling opportunity to invest in critical AI infrastructure at an inflection point, with the characteristics of an early-stage company despite its public listing. The firm has completed an eighteen-month strategic pivot from Bitcoin mining to AI cloud services, securing 4.5GW of power capacity and a $9.7 billion contract with $MSFT. Revenue from the new AI cloud business line is expected to commence at scale in CY Q2 2026, with a path to $3.4 billion in annualized run rate revenue by year-end 2026. The investment case centers on IREN's position as a vertically integrated infrastructure provider serving the world's most sophisticated technology companies in a capital-intensive industry with formidable barriers to entry. Overview & Timing: IREN should be evaluated as an early-stage infrastructure company currently in its pre-revenue phase for its core AI cloud business. The fiscal Q2 2026 results (ended December 31, 2025) reflect the transitional period during which the company is deliberately decelerating Bitcoin mining operations and constructing AI-optimized data centers. The AI Cloud Services segment generated $17.3 million in the quarter, representing initial deployments at the Prince George, British Columbia facility. The meaningful revenue ramp is expected to begin in CY Q2 2026 as the Microsoft contract becomes operational at the Childress, Texas campus and additional British Columbia capacity comes online. By CY Q4 2026, the company targets 140,000 deployed GPUs supporting $3.4 billion in annualized revenue. This timing structure is critical to the investment thesis. Current period losses of $116 million in operating income reflect the strategic build-out phase, not operational underperformance. The $720 million invested in data center infrastructure during the six months ended December 31, 2025 represents capital deployed ahead of revenue recognition. Investors should expect continued transitional losses through fiscal Q3 2026, with the business model proving itself in the subsequent three quarters as contracted capacity becomes revenue-generating. My Individual Thoughts: This AI infrastructure industry is undergoing a fundamental transformation that will reshape technology capital allocation over the next 3 years. What began as hyperscale companies building their own data centers has evolved into a recognition that compute infrastructure, particularly GPU clusters at gigawatt scale, represents a strategic asset class requiring specialized capabilities that most enterprises cannot replicate internally (not yet at least). The supply-demand imbalance in AI compute capacity will persist through at least 2028, driven by the exponential growth in model training requirements and inference workloads that show no signs of moderating (see Claude’s most recent drop). This dynamic creates exceptional pricing power for infrastructure providers who can deliver certified, reliable capacity at scale, particularly those serving tier 1 customers where uptime and performance are non-negotiable. The industry structure will consolidate rapidly around a small number of vertically integrated providers. The capital intensity, requiring billions in equity and debt financing, combined with the multi-year lead times for power interconnection agreements creates natural oligopoly conditions. Companies that secured power rights in 2023 and 2024, as IREN did, possess assets that cannot be replicated by new entrants at any reasonable timeframe. I expect three to five dominant players to emerge in North America by 2027, with IREN positioned as one of them based on its secured power portfolio and demonstrated ability to access capital markets. The competitive landscape will bifurcate between infrastructure owners like IREN who control the full stack from power to cooling systems, and asset-light resellers who will face margin compression as hyperscalers increasingly prefer direct relationships with infrastructure operators (as seen with $NBIS starting to scale data center builds and moving away from their pure play asset-light model). The investment opportunity exists precisely because IREN sits at an inflection point where the infrastructure risk has been substantially retired while the revenue risk remains unproven to public market investors. The company has already deployed $1.1 billion in data center capital over the past six months, secured firm power interconnection agreements, and signed a $9.7 billion anchor contract with Microsoft. The heavy lifting of site development, utility negotiations, and construction execution, the areas where infrastructure projects typically fail, are largely complete or de-risked. What remains is customer contract conversion, which carries materially lower execution risk when serving sophisticated counterparties conducting thorough technical and commercial diligence. (If you aren't aware technical and commercial diligence takes 3-6 months). The market has not yet priced in the revenue that will emerge in the next two quarters because investors struggle to underwrite pre-revenue infrastructure businesses and lack the operational frameworks to evaluate construction progress and contract pipeline quality. According to Dan, they are in "advanced negotiations" which I believe is LOI stages, but I could be wrong. From an operational perspective, I believe IREN's vertically integrated model will prove decisively superior to competitors relying on third-party contractors or leased power arrangements. The company's ability to bring Horizon 1 at Childress from groundbreaking to operational status in under 18 months, while simultaneously managing construction at four other sites, demonstrates execution capabilities that are genuinely scarce in not only this industry but many others. The 2,000 operational team represents an intangible asset that will compound in value as institutional knowledge accumulates around GPU cluster optimization, cooling efficiency, and rapid deployment methodologies. These operational advantages will translate into higher gross margins and faster time-to-revenue compared to peers, though this will only become evident in the financial results over the next four quarters. Not immediately. The compelling aspect of this investment at the current entry point is the asymmetric risk-reward profile created by the timing mismatch between capital deployment and revenue recognition. Investors who wait for revenue proof will pay substantially higher multiples once quarterly results demonstrate the business model's viability, for a scarce infrastructure asset serving strategic customers with high margins and deep revenue visibility (multi-year). In my opinion, the downside case is protected by the liquidation value of the power rights and constructed data centers, which would attract acquisition interest from hyperscalers and/or infrastructure funds at valuations near or even above the current market cap. Looking forward, I envision IREN as a multi-billion dollar revenue business serving numerous tier-one customers with industry-leading gross and EBITDA margins. The business model will have transitioned from capital-intensive growth to steady-state cash generation, enabling debt paydown and shareholder returns. The company's remaining power capacity will provide optionality for further expansion or partnerships, while the operational expertise will create opportunities for adjacencies such as purpose-built training clusters or inference-optimized architectures. Most importantly, the business will have demonstrated that founder-led, vertically integrated infrastructure operators can compete effectively for the most demanding workloads in technology, establishing a valuation premium to both traditional data center REITs and asset-light cloud resellers. This business is still a baby, they are purposely slowing down $BTC mining production and swapping ASICs for GPUs. I believe for the future cashflows and extremely high barriers to entry in the future, big money will pay strong premiums. In our new AI world, I believe this industry will become a safe haven asset. Along with other AI adjacent industries I see. With that being said, I can tell most of you have never operated a business, nor worked with or invested in early stage companies. We are lucky because typically you wouldn't get your money back for YEARS because of how illiquid private markets are and genuinely most early stage investments go to $0. Just my thoughts, not financial advice, and do your own research.

We could be on here

Every seminar ends with great questions and not enough time. Next Thursday in Sydney, we’re flipping the format. Markets Unfiltered is built around the questions you submit when you register. If you want something covered, this is how you make it happen.

- 3 year pay back on GPUs - the last 2 years of MSFT contract then pay off the data centers too - year 6 onwards = free data centers (for colo or cloud) + any residual value in the GB300s Or, you can sign a colo deal and wait 10++ years to get your capital back on the data centers. I'll take option 1 thanks, and keep compounding shareholder returns into the AI thematic.