The $8.5B Signal: Wall Street's Quiet Exit from Crypto Mining Finance85亿美元信号：华尔街悄然撤离加密挖矿金融
CoreWeave's $8.5B AI-backed loan, backed by Meta, shows Wall Street has found a better model than crypto mining finance: active compute with contracted cash flows.CoreWeave 获得 Meta 支持的85亿美元AI贷款表明，华尔街找到了比加密挖矿融资更好的模式：主动计算+合同现金流。
The Setup
CoreWeave just closed an $8.5 billion GPU-backed loan, co-signed by Meta. It’s being called “ComputeFi” — and it’s everything that “MinerFi” tried to be, but wasn’t.
The difference matters structurally. Crypto mining finance used ASIC hardware as collateral. When BTC dropped, both revenue and collateral value collapsed together. Lenders learned to stay away.
ComputeFi works differently: GPUs must be deployed, operational, and already generating revenue before capital is extended. The loan sits on top of a live cash-flow engine, not a speculative asset.
Key Takeaways
- Miners failed the finance test: ASIC-backed lending models imploded alongside crypto prices. The collateral was only valuable when the market was hot.
- ComputeFi works because it’s not crypto: Active AI contracts, real enterprise customers, predictable revenue. Wall Street understands this.
- CoreWeave’s backlog = $67B. That’s more than 7x IREN’s $9.7B and ahead of Nebius at $47B. Early pivot from crypto to AI = structural first-mover advantage.
- Bernstein rates CoreWeave top among “neocloud” peers on commercial model: software depth + contracted/on-demand mix + backlog + customer diversification.
Why It Matters
If you’re still thinking about crypto mining companies as infrastructure plays, you’re reading last cycle’s playbook. The real infrastructure story now is GPU cloud — and it’s being financed by the same banks that once wrote mining loans, now competing for AI compute exposure.
For crypto-native investors, the implication is uncomfortable: the smart money has already left the building.
What to watch:
- Whether IREN and other hybrid miners can complete the pivot before their BTC revenue meaningfully deteriorates
- Whether neocloud valuations compress as more GPU supply comes online in 2026-2027
- Whether any crypto mining finance players attempt to rebrand as “AI compute lenders” — and whether anyone believes them
背景
CoreWeave 刚完成了一笔 85 亿美元的 GPU 担保贷款，Meta 参与联署。这被称为”ComputeFi”——正是”MinerFi”曾经试图成为、但未能实现的东西。
两者之间的区别在结构上很重要。加密挖矿融资以 ASIC 矿机作为抵押物。当 BTC 下跌时，收入和抵押物价值同时崩盘。贷款方学聪明了。
ComputeFi 的逻辑不同：GPU 必须已部署、已运行、已产生收入，资本才会到位。贷款建立在实时现金流引擎之上，而非投机资产之上。
关键要点
- 矿商没能通过金融测试：以 ASIC 为抵押的贷款模式随加密货币价格一起崩盘。抵押物只有在市场火热时才有价值。
- ComputeFi 行得通，因为它不是加密货币：企业级AI合同、真实客户、可预测的收入。华尔街理解这个。
- CoreWeave 的订单储备 = 670 亿美元。比 IREN 的 97 亿高出7倍多，领先 Nebius 的 470 亿。早于市场从加密货币转向 AI = 结构性先发优势。
- Bernstein 在”新云”同类中给予 CoreWeave 最高评级，在商业模式上：软件深度 + 合同/按需组合 + 订单储备 + 客户多元化。
为什么重要
如果你还把加密挖矿公司当作基础设施标的来思考，你还在用上一个周期的 playbook。真正的基础设施故事现在是 GPU 云——而它正被那些曾经撰写挖矿贷款的银行提供资金，如今它们争相获取 AI 计算敞口。
对于加密原生投资者来说，这个暗示令人不安：聪明钱已经离开这个领域了。
值得关注：
- IREN 和其他混合矿商能否在 BTC 收入实质性恶化之前完成转型
- 随着 2026-2027 年更多 GPU 算力上线，新云估值是否会压缩
- 是否有加密挖矿融资玩家试图转型为”AI 计算贷款方”——以及市场是否买账
Related Essays相关文章
Join Rex's Lab on Telegram 加入 Rex's Lab
Crypto · AI · Investing — raw thinking, before it becomes a tweet. 加密 · AI · 投资 — 推文之前，更原始的思考。