发布时间:2026-08-28 一、AI总结内容 一、核心要点 1. AI基础设施需求增长迅猛,大语言模型周度代币用量8个月内增长近17倍,物理基础设施建设滞后形成结构性瓶颈 2. IREN构建全栈AI基础设施,涵盖数据中心、GPU存储等硬件层、NVIDIA认证Marantis软件层,实现各层价值联动 3. 2026财年核心成果:新签多份长期合同(含前沿AI实验室、Cohere等),2026年容量已锁定40亿美元ARR,运营ARR达10亿美元;为微软交付首个50MW部署项目Horizon One,英伟达认证通过;近3个月完成65亿美元GPU融资,超100%覆盖相关GPU资本支出,28亿美元无需投资级担保且利率仅个位数 4. 2027年计划交付300MW IT负载,2028年增至约1.2GW gross容量,布局美国德克萨斯、加拿大不列颠哥伦比亚、澳大利亚、西班牙等多地站点,2027资本支出预计250-300亿美元 5. 定价方面,三年期合同价较2023年11月上涨125%,五年期涨70%,近期合同IT负载价格超2000万美元/兆瓦,投资回收期约2年,客户 prepayments覆盖GPU资本支出的45%-55% 二、风险与关注 1. 资本支出规模大,需平衡数据中心建设与GPU采购的资金分配,需持续高效利用资金飞轮支撑增长2. AI基础设施需求与供给的动态变化,需跟踪客户需求、资本融资环境、GPU供应的波动3. 站点建设面临电网、水资源等地方政策约束,需确保项目合规运营三、投资线索 1. 全栈布局优势:硬件与软件结合,Marantis提供编排层,扩大客户覆盖范围,包括小型AI企业、企业客户等 2. 融资模式成熟:投资级客户融资利率约6%,非投资级客户约9%, prepayments与GPU融资形成资金循环,剩余资金可反哺数据中心建设3. 长期增长逻辑:AI应用从 chat 向物理世界延伸(机器人、自动化),算力需求持续旺盛,IREN锁定多地电网与土地,具备差异化竞争优势 二、音频原文(转写) Good day and thank you for standing by. Welcometo Iron FY 2026 Results Call. After this, at this time all participants are in a listen only mode. After the speaker's presentation, there willbe question and answer session. To ask a question during the session, you will need to press star one one on your telephone. You will then hear an automated message advising your hand is raised. To withdraw your question, please press starone one again. Please be advised that today's conference is being recorded. I'd like to hand theconference over to your first speaker today, Mike Power, Vice President, Investor Relations. Please go ahead. Good afternoon and welcome to IRON'S FY 2026 RESULTS presentation. I am Mike Power, VP of Investor Relations, and with me on thecall today are Daniel Roberts, co-founder and coCEO, Anthony Lewis, CFO and Kent Draper, ChiefCommercial Officer. FOUR Begin, please note that this call is being broadcast live with a presentation for those dialed in by phone. You can elect to ask a question through the moderator afterour prepared remarks. I would like to remind everyone that certain statements made during thiscall. They constitute forward-looking statements. Those statements are based on current expectations and assumptions and the subjects. to risksand uncertainties that could cause actual result s to differ materially. Please refer to Slide 2of the accompanying presentation. and o s six efilings for more information in that regard. Um, d. cheaper compute It doesn't make existing things more efficient. It makes new things economic.Infrastructure enables applications. Applications create demand for more infrastructure. Everybuild out in history has worked this way. And that's the structural disconnect, and it's only getting wider. So let me walk through how we're set up. against that backdrop So we operate acrossthree layers. from the bottom up So first of all, the data centers. the land the power The substations, the cooling. Arguably the hardest layerto build. And that's where the shortage begins.In the compute. the gpus storage networking that go inside the data centers. And then finally software on top. The manager said. services and enterprise support That's where Marantis lives for us. And just today, Marantis was named an inaugural NVIDIA Certified Hypervisor. So we've nowgot NVIDIA validation at the software layer. aswell as the hardware Why own all three? Because each layer makes the one underneath it worth more. A grid connection is worth more witha data centre on it. Worth more again with GPUs.inside More again with services wrapped aroundthe customer. Most of this market rents at leastone of those layers. we own the entire stack Sohere's how the year went. Just move on to highlights, please operate it. Thank you. Four thingsto take away from this update. And I'll be quick because there's more detail coming a little bit later. So firstly customers New multi year cloud contracts including Cohere. Prometheus perplexity figure i o i Foul AI. Kingsfield AI. and separately A leading frontier AI lab. whose name We're not able to disclose just yet. Revenue fourbillion dollars at ARR. is now contracted for our 2026 capacity. and one billion All that is operating today. This does not include revenue expected to ramp in 2027. such as the seven hundredmillion dollars of ARR associated with our NVID IA cloud contract. Delivering Horizon One was delivered to Microsoft this month. The first of four 50 megawatt deployments. With horizon two through four. Targeted for the December quarter. And finally funded. six and a half billion dollarsof GPU financing now in the past. three monthswith prepayments that's more than one hundred percent of the associated gpu capex funded and twopoint eight billion of it needed no investmentgrade off take and still priced in the single digits So let me start with customers because everything else. follows from them Our 2026 capacityis largely sold out. So the questions we get now are all about 2027. and 2028. We're continuingto contract future capacity deliberately. everycontract Opportunity gets weigh