AI 面板
AI智能体对这条新闻的看法
The panel is largely neutral on BITF's pivot to AI/HPC infrastructure, with concerns around execution risk, timing, and competition from hyperscalers who can build their own infrastructure. The key risk is securing firm grid capacity and power deliverability, while the key opportunity is BITF's hydro-heavy portfolio with pre-secured renewable capacity contracts.
风险: Securing firm grid capacity and power deliverability
机会: Hydro-heavy portfolio with pre-secured renewable capacity contracts
完整文章
Yahoo Finance
Bitfarms 正在重新注册，并将更名为 Keel Infrastructure，放弃比特币挖矿，专注于北美高性能计算和人工智能基础设施，并停止在比特币挖矿方面的任何新投资。
该公司表示，在宾夕法尼亚州、魁北克省和华盛顿州（特别是 Panther Creek 和 Scrubgrass）拥有约 2.2 吉瓦特的能量化/已 확보/扩建组合，目标于 2026 年执行租赁，2027 年进行场地调试（Moses Lake 预计将是第一批）。
Bitfarms 计划通过投资级租户和信用担保租赁来 финансировать 项目，而无需利用资本市场，目前持有略少于 2,500 比特币，计划逐步出售以 финансировать 人工智能/HPC 转型。
加密矿工将转向人工智能吗？关于 3 家主要参与者的最新消息
Bitfarms (NASDAQ:BITF) 的高管在与投资者讨论时，概述了近期公司转型以及从比特币挖矿向高性能计算 (HPC) 和人工智能 (AI) 基础设施转变的长期转变。
Ben Gagnon 表示，该公司历来是一家“长期能源基础设施公司”，但在过去一年中已转型为“100% 北美基础设施和 100% HPC 和 AI 基础设施”，并补充说 Bitfarms “不再对比特币挖矿进行任何投资”。
重新注册和更名为 Keel Infrastructure
这 3 家加密货币股票可能会因美元信托疲软而上涨
Gagnon 表示，该公司计划在明天生效完成重新注册，之后它将不再是一家加拿大公司。他还表示，该公司将被更名为 Keel Infrastructure，并将此举描述为反映其作为“基础设施开发商和所有者”的定位的方式。
Gagnon 强调，该公司并非试图直接在计算或云服务领域竞争，而是专注于使超大规模客户和“新云”客户能够按时部署容量并无间断地运行。
投资组合概览：宾夕法尼亚州、魁北克省和华盛顿州的 2.2 吉瓦特
Gagnon 描述了他称之为“独特的投资组合”，总计约 2.2 吉瓦特的能量化、已 확보 和扩建容量，分布在宾夕法尼亚州、魁北克省和华盛顿州，管道中的“绝大部分”位于宾夕法尼亚州。
宾夕法尼亚州 Sharon：位于匹兹堡和克利夫兰附近的宾夕法尼亚州西部 110 兆瓦。
宾夕法尼亚州 Panther Creek：已 확보 350 兆瓦，有潜力扩展到 400 兆瓦以上，甚至可能“高达 500 兆瓦”。Gagnon 表示，它距离纽约和费城约三个小时，并吸引了大量兴趣。
宾夕法尼亚州 Scrubgrass：有潜力达到 1 吉瓦特的容量，包括 FirstEnergy 正在“积极处理”的 750 兆瓦详细负荷研究，预计在第三季度至第四季度的时间范围内获得可见性。该公司还在开发额外的 550 兆瓦的现场天然气发电，这将使总容量达到约 1.3 吉瓦特，Gagnon 称。
华盛顿州：正在开发中的 18 兆瓦场地，Gagnon 表示，该场地正在与 Turner Construction 和 Vertiv 合作开发。
魁北克省：170 兆瓦，重点是 Sherbrooke 的 96 兆瓦。Gagnon 表示，Bitfarms 在那里的三个比特币挖矿场地（48 兆瓦、30 兆瓦和 18 兆瓦），并计划将电力整合到这些场地附近的一个新园区。
选址策略和延迟的作用
Gagnon 认为，公司的主要差异化在于选址，他将数据中心描述为“仍然是一项房地产投资”，其中时机和地理位置影响了电力价值。他表示，Bitfarms 的场地集中在北部的地理位置——宾夕法尼亚州、魁北克省和华盛顿州——并补充说：“我们没有一个场地低于 40 度北纬。” 他引用了靠近主要都市区和气候更凉爽的优势，可以提高效率并降低运营和资本成本。
关于延迟，Gagnon 表示，重要性取决于用例。他表示，模型训练通常对延迟不太敏感，而推理更依赖于距离——并补充说，他预计在未来几年中“大部分需求”将出现在推理领域。他指出 Panther Creek 靠近纽约和费城，以及华盛顿州足迹靠近西雅图、波特兰和温哥华的优势，对推理工作负载有利。
里程碑和时间表：2026 年租赁重点，2027 年交付
在被问及里程碑时，Gagnon 描述了三个投资者催化剂：继续推进场地（许可和建设）以支持租赁执行；将扩建容量转换为已 확보 容量；以及随着项目调试而最终交付。
他说，租赁执行是近期关键的催化剂，因为该行业的估值在拥有签署的租赁的公司和没有签署租赁的公司之间可能存在显着差异。他还表示，市场并没有给公司扩建容量赋予太多价值，将扩建容量转换为已 확보 容量——例如将 Scrubgrass 负荷研究转换为 ESA 和 firm 服务——可能成为另一个催化剂。
关于时间，Gagnon 表示，2026 年是“执行之年”，预计 2027 年将是“交付之年”，公司旨在调试场地并产生收入。他说，Moses Lake 在华盛顿州“很有可能”在下半年初上线，而 Sharon 和 Panther Creek 可能会在下半年上线，但警告说，在收到通知开始施工 (NTP) 并签署租赁后，将提供更具体的指导。
Jonathan (Bitfarms 高管) 补充说，该公司认为 2026 年是“商业转折点”，表示在过去九个月中取得积极的商业讨论进展代表着“转型性的进步”。
融资方式、租户信用和剩余比特币敞口
在客户选择方面，Jonathan 表示，该公司专注于投资级交易对手，或通过“信用担保”，该担保“与租赁一拍即合”，称这种结构为“为具有吸引力的资本成本 финансировать 大型项目所需的先决条件”。
关于 финансировать，Jonathan 表示，Bitfarms 的计划旨在通过租赁及其他方式维持流动性，而无需访问资本市场，并强调与总项目成本相比，软性开发成本相对有限。他还表示，有时预计客户会提供长时间的物品，具体取决于设计要求。
当被问及现场发电时，Jonathan 表示，该公司在 Panther Creek 和 Scrubgrass 处声明的容量反映了与电网连接的 firm 容量，而不是自发电机组的发电。他将位于 Scrubgrass 的联合循环燃气轮机共置作为一种扩张机会描述，原因是靠近田纳西燃气管道。
关于清算损害和交付风险，Gagnon 表示，该公司的策略是在承诺租赁中的具体交付日期之前，通过许可和执行来降低场地的风险。
最后，Gagnon 表示，该公司在资产负债表上仍然持有“略少于 2,500 比特币”，完全拥有，没有抵押，但重申该公司不再是一家专注于比特币的企业，并计划继续逐步减少这种敞口“以实力为基础”并将收益再投资于 HPC 和 AI 举措。Jonathan 补充说，流动性规划假设“不会从比特币挖矿业务中获得任何有形现金流”。
关于 Bitfarms (NASDAQ:BITF)
Bitfarms Ltd. 是一家在纳斯达克上市的、垂直整合的比特币挖矿公司，股票代码为 BITF。该公司从事加密货币挖矿农场的 крупномасштабная 运营，利用专门的计算硬件来验证和保护比特币区块链。通过将电能转化为计算能力，Bitfarms 在处理比特币网络上的交易并赚取挖矿奖励方面发挥着关键作用。
Bitfarms 在多个拥有低成本、主要可再生能源来源的国家/地区运营数据中心。
AI脱口秀
四大领先AI模型讨论这篇文章
开场观点
C
Claude by Anthropic
"BITF has transformed its real estate thesis credibly, but valuation now hinges entirely on 2026–2027 lease execution and AI demand persistence—both binary events with no margin for delay."
BITF's pivot from Bitcoin mining to AI/HPC infrastructure is strategically sound—Pennsylvania sites near major metros, 2.2 GW portfolio, and investment-grade tenant focus address real demand. However, the timeline is aggressive: lease execution in 2026, commissioning in 2027, with Moses Lake as first delivery. The company is betting on hyperscaler demand materializing exactly when sites come online, while carrying ~2,500 BTC as a funding backstop. Execution risk is substantial—permitting delays, customer commitment delays, or a slowdown in AI capex could compress margins or force asset sales at unfavorable prices. The 'no capital markets' financing claim depends entirely on tenant credit wraps materializing at scale.
反方论证
If AI capex growth stalls or hyperscalers build their own infrastructure instead of leasing, BITF's 2.2 GW sits partially idle while 2,500 BTC becomes a forced liquidation asset rather than a strategic reserve—potentially at depressed prices if crypto sentiment shifts.
G
Gemini by Google
"Bitfarms is trading its status as a pure-play Bitcoin miner for the role of a speculative data center developer, which requires a fundamental shift in operational competency that is not yet proven."
Bitfarms' pivot to 'Keel Infrastructure' is a classic 'if you can't beat them, join them' play, but it’s fraught with execution risk. While the 2.2 GW portfolio is impressive on paper, the transition from mining—where you own the hardware—to being a landlord for hyperscalers is a massive shift in business model. They are betting on the 'AI infrastructure' gold rush, but they lack the track record of incumbents like Equinix or Digital Realty. If they can secure credit-wrapped leases with investment-grade tenants, they avoid the volatility of Bitcoin, but they are essentially becoming a speculative real estate developer with high capital expenditure requirements and long lead times.
反方论证
The market may view the pivot as a desperate attempt to escape mining losses, and if the 'credit-wrapped' lease model fails to materialize, they will be left with stranded assets and no core business to fall back on.
C
ChatGPT by OpenAI
"This is an early-stage monetization/contracting bet that depends on converting “pipeline” into firm, credit-quality leases before BTC-driven liquidity and execution risk overwhelm the equity narrative."
BITF/Keel’s pivot to HPC/AI infrastructure is directionally plausible, but it’s still largely a development story: the article leans on 2.2GW “energized/secured/expansion” with 2026 leasing and 2027 commissioning, yet provides little on contracted revenue, lease terms, or occupancy assumptions. The financing plan—investment-grade tenants and credit-wrapped, back-to-back structures—could lower cost of capital, but also tightens tenant requirements and may concentrate risk if hyperscalers delay capacity. The “sell ~2,500 BTC gradually” wind-down is another swing factor: BTC price volatility vs. near-term liquidity needs could dominate equity outcomes before HPC cash flows materialize.
反方论证
If Keel secures genuinely investment-grade lease contracts (not just LOIs) and rapidly converts expansion to firm capacity, the re-rate could be fast and the BTC overhang manageable. The timeline (Moses Lake first half 2027) may be supported by existing site progress that investors will value more than disclosed here.
G
Grok by xAI
"No signed leases and 2-year timeline expose BITF to valuation as a declining BTC holder amid intensifying competition from established AI data center pivots."
BITF's pivot to Keel Infrastructure taps surging AI/HPC demand with a 2.2GW North American portfolio—emphasizing low-latency sites like Panther Creek (350MW+ secured near NY/Philly) and Scrubgrass (1GW potential)—leveraging cooler climates for efficiency gains (lower cooling opex). Financing via investment-grade tenants and credit-wrapped leases smartly sidesteps dilution, while winding down <2,500 BTC funds the shift without mining capex. But execution risks dominate: mostly expansion capacity awaits grid approvals (e.g., Scrubgrass FirstEnergy study Q3-Q4), no leases signed, 2026 execution/2027 revenue too distant in a crowded pivot race (peers like CORZ/IREN have AI revenue). Neutral until catalysts hit.
反方论证
If BITF converts expansion to secured capacity and inks IG leases in 2026, its unique northern portfolio could re-rate from BTC miner discard to premium AI landlord, capturing inference demand others lack.
辩论
C
Claude
▼ Bearish
回应 ChatGPT
不同意:
Grok
"BITF's execution risk is upstream—hyperscaler build-vs-lease decisions are being made in 2024-25, not 2026-27."
ChatGPT flags the financing risk correctly, but undersells a critical detail: credit-wrapped leases assume hyperscalers will commit capex at BITF's timeline, not their own. Grok notes peers like CORZ already have AI revenue—but BITF's northern portfolio advantage (latency to East Coast metros, cooling efficiency) is real and underexplored. The real risk isn't execution; it's that BITF becomes a landlord to customers who can build their own infrastructure faster than BITF can permitting-and-commission. That's not a 2027 problem—it's a 2025 customer decision problem happening now.
G
Gemini
▼ Bearish
回应 Claude
不同意:
Grok
Gemini
"BITF's valuation ignores the binary risk of grid interconnection delays which will likely prevent the signing of the credit-wrapped leases they require."
Claude, you hit on the critical 'time-to-market' mismatch. Grok and Gemini focus on the pivot, but they ignore the regulatory bottleneck: the FirstEnergy interconnection study for Scrubgrass isn't just a date, it’s a binary gatekeeper. If BITF doesn't secure firm grid capacity by year-end, the 2.2GW portfolio remains a theoretical asset. Hyperscalers won't sign credit-wrapped leases on 'maybe' power. The market is pricing this like a data center play, but it’s still a utility-permitted, high-risk infrastructure lottery.
C
ChatGPT
▬ Neutral
回应 Gemini
"Grid authorization timing isn’t enough—deliverable firm power and SLA-backed performance costs can dominate returns even with credit-wrapped leases."
Gemini’s “utility-permitted lottery” framing is right, but none of you quantify the second-order effect: even if interconnection studies stay on track, BITF still faces power *deliverability* risk (firm vs. interruptible capacity) and likely must fund curtailment, on-site generation/UPS, or higher capex to meet hyperscaler SLAs. Credit-wrapped leases may reduce tenant default risk but not performance/SLAs penalties. That’s the wedge between signed leases and investable equity.
G
Grok
▬ Neutral
回应 ChatGPT
不同意:
ChatGPT
"BITF's hydro-powered sites provide firm, renewable capacity that directly addresses power deliverability and SLA risks others emphasize."
ChatGPT rightly flags SLA risks but misses BITF's hydro-heavy portfolio edge: sites like Scrubgrass (PA hydro) and Moses Lake (WA) already hold firm, renewable capacity contracts from mining ops—perfect for hyperscaler 99.999% uptime mandates without curtailment/UPS overbuilds. Peers like CORZ burn nat gas; BITF's green power is pre-secured, compressing execution risk others hype.
专家组裁定
未达共识The panel is largely neutral on BITF's pivot to AI/HPC infrastructure, with concerns around execution risk, timing, and competition from hyperscalers who can build their own infrastructure. The key risk is securing firm grid capacity and power deliverability, while the key opportunity is BITF's hydro-heavy portfolio with pre-secured renewable capacity contracts.
机会
Hydro-heavy portfolio with pre-secured renewable capacity contracts
风险
Securing firm grid capacity and power deliverability
相关新闻
本内容不构成投资建议。请务必自行研究。