自2008年雷曼兄弟倒闭以来，政策制定者在1929年与2008年的经验下，透过如 Basel III 等规则加强监管。文章指出，2000年的网路泡沫在市场上极为剧烈，但社会层面的破坏性远低于「大萧条」与「金融海啸」，因为资产下跌主要集中在如401(k)等可承受损失的金融资产，而非住房。系统性危机常见两类成因：一是高杠杆（如次级房贷、部分首付仅5%）；二是原本被视为低风险的资产其实高风险，两者在2008年同时出现。文中将AI话题作为比较：NVIDIA股价自10月高点回落超过20%，抹去超过10亿美元? wait，实际为超过1兆美元（1 trillion）的市值，与2006年美国次级房贷高峰约6000亿美元（600 billion）相比，更像是有序修正而非毁灭性崩盘。

主要关切在于私募信贷(private credit)风险已部分移向银行监管外部。文章称，私募信贷可动用「不罕见」的十亿美元（1 billion）级交易，且贷款来源常有约一半向银行借入；更麻烦的是标的票据流动性差、难以做市价标示，利率和价值难及时发现。借款人大多为私募股权(PE)企业，且常以高杠杆与高配息策略加大脆弱度。Ludovic Phalippou指出，真正风险不在「突然爆炸」，而是隐性损失持续缓慢外显，体现在回报下降与保险公司资产负债表走弱。保险机构是主要资金来源之一，但缺乏单一强力中央监管；好在其通常不易出现挤兑，且在必要时可被要求增资。银行也有保护：即使已向私募信贷基金放贷，仅在基金净资产值(NAV)跌至0美元时才可能受损，目前仍与0有明显距离。

对「恐慌性撤资」的讨论，主要来自零售导向基金的赎回压力。这些基金通常允许每季赎回最多5%的基金价值，近期涉Tricolor、First Brands的争议与「SaaSpocalypse」情绪使部分投资人提前离场。因为BDC可公开交易，当其价格低于报告净资产值（常见在每美元65到80美分）时，个人投资人会产生卖出诱因，外界容易误解为系统危机。文中认为尚未形成全面恐慌：撤资比例离100%仍远，且许多产品仍有可观净流入，表示估值分歧而非一致踩煞车。3月末Jay Powell表示当局正监测向银行体系外溢，但尚未看到传染机制，意味私募信贷或出现部门性回报下滑与个别损失，而非即将的广泛系统性事件。

Since the 2008 collapse of Lehman Brothers, policymakers have tightened rules such as Basel III based on lessons from 1929 and 2008. The summary notes that the 2000 dot-com bust was severe in markets but less socially damaging than major crises, because losses were largely concentrated in financial assets like 401(k) portfolios rather than homes. Systemic crises usually arise from two channels: heavy leverage (for example, subprime borrowers with as little as 5% down) and the reversal of assets once believed to be safe; both were present in 2008. By contrast, what is now called an AI bubble appears more orderly in this framing: NVIDIA shares fell more than 20% from the October peak and wiped out over US$1 trillion in market value, while the U.S. 2006 subprime mortgage peak was around US$600 billion, which the article treats as a less extreme signal than a crash dynamic.

The principal concern is that private credit risk is concentrated outside stricter bank regulation. Deals can be large (US$1 billion is not unusual), and lenders often fund about half of lending with bank borrowing; loan notes are often illiquid and hard to price in real time. Borrowers are frequently private-equity-linked firms, often associated with high leverage and heavy dividend extraction. Ludovic Phalippou is cited for the key risk pattern: not a sudden implosion but slow, hidden losses materializing through lower returns and weaker insurer balance sheets. Insurers, a large source of capital, are less visible to a single central regulator, but they are less prone to runs and could be forced to raise capital if needed. Banks remain relatively insulated because many have lent to private-credit funds, yet these funds’ net asset value would have to fall to zero before bank losses appear.

The “panic” story centers on retail-linked funds. These funds for individual investors often allow up to 5% quarterly redemption, and concerns over Tricolor, First Brands, and the SaaSpocalypse episodes pushed some investors to take exits. Because business development companies (BDCs) are publicly traded, their shares trading below reported net asset value at about 65 to 80 cents on the dollar created a clear incentive to sell, which can look like a panic. The article argues this is not system-wide panic: redemption rates are much farther from 100% than near-total flight, and some funds still receive substantial inflows even amid outflows. As of late March, Fed Chair Jay Powell said authorities were watching for contagion to the banking system and had not seen it, suggesting private-credit pain may be largely gradual return compression and selective losses rather than a new global systemic event.