Skip to content

There's Not Enough Money In The World: AI Capex as Inverse QE

English Summary

Plur_daddy argues that AI capital expenditure is draining global financial capital and triggering a market regime change from abundance to scarcity, forcing a fundamental revaluation of all assets as the deepest pockets in the world run dry and speculative investments face systematic liquidation.

The AI Capex Mechanism: Inverse QE

AI capex operates like fiscal stimulus in reverse. When hyperscalers issue bonds or sell treasuries to fund AI infrastructure, the private sector absorbs duration—but crucially, once dry powder is exhausted, every dollar flowing into AI must come out of something else. This creates a convex scramble for capital.

The parallels: - Fiscal stimulus: Government issues treasuries → private sector absorbs duration → cash circulates with multiplier effect → net positive for asset prices - AI capex (early phase): Hyperscalers issue bonds → private sector absorbs duration → cash circulates with multiplier effect → net positive for asset prices - AI capex (current phase): No more dry powder → forced selling of existing assets → negative portfolio balancing effects → inverse QE

Unlike fiscal stimulus (where the Fed becomes the end absorber), AI capex forces direct competition between uses of capital.

The Capital Shortage Cascade

The deepest pockets are tapped out. When Altman approaches the Saudis and Softbank for commitments, they no longer have dry powder—they must sell assets to free up capital:

  1. Portfolio liquidation: BTC (underperforming), SaaS stocks (disruption risk), hedge fund redemptions
  2. Forced selling by funds: Hedge funds sell to meet redemptions
  3. Cascading effects: Falling prices → tightening margin → more selling → confidence collapse

This is why press reports indicate the world's deepest pockets (Saudis, Softbank) are "largely tapped out."

The Warsh Problem: Wrong Diagnosis

Trump's selection of Warsh accelerates the crisis. Warsh believes the problem is "too much money," when the actual problem is capital shortage. This misdiagnosis drives policy in exactly the wrong direction, tightening conditions when loosening is needed.

The timing of market acceleration since Warsh's selection confirms this dynamic.

Asset Bifurcation: Winners and Losers

When capital becomes scarce, discount rates rise and markets force a brutal evaluation of where capital is most useful:

Winners (near-term cash flows): - ✅ DRAM/HBM/NAND manufacturers (SNDK, MU): Overearning now despite cyclical long-term outlook - Cash flows today > speculative future value

Losers (long-duration speculation): - ❌ Crypto: "Tip of the spear for liquidity conditions" → feels bottomless - ❌ Speculative retail momentum stocks: Unable to hold gains even with improving fundamentals - ❌ SaaS software bags: Face AI disruption risk

Market-wide effects: - Sovereign and credit yields rise (demand for money > supply) - Highly speculative assets disproportionately punished

The Liquidity Analogy: Water in the Tub

Money/liquidity = water. Financial assets = rubber ducks floating in the tub.

To raise asset prices (higher water level), you can: 1. Increase total volume (rate cuts/QE) 2. Unclog pipes flowing in (plumbing actions like RRP) 3. Decrease water draining out

The current problem: Too much demand for money, causing crowding out. Discussion focuses on supply, but demand is equally important.

Investment Strategy: Defense and Patience

Author's position: - Mostly cash after selling gold/silver at the top - Not in a hurry to buy anything - Not a trading call — context for sense-making

Recommendations: - Don't stay complacently super long - Play defense and be selective - Manage risk actively - Be patient: Extraordinary opportunities will emerge this year

Key Insight

This is not a temporary dip or correction—it's a structural market regime change. The capital-light web 2.0/SaaS paradigm that powered the 2010s allowed excess capital to flow into speculation. The capital-intensive AI paradigm reverses this, forcing capital out of speculation and into infrastructure.

The paradigm shift: - 2010s: Capital abundance + productivity shortage → speculation thrives - 2020s: Capital shortage + AI capex demand → speculation dies

繁體中文總結

Plur_daddy 主張 AI 資本支出正在吸乾全球金融資本，引發市場從「資本過剩」到「資本短缺」的結構性轉變，迫使所有資產重新估值，因為全球最深的口袋已經見底，投機性投資面臨系統性清算。

AI Capex 機制：反向量化寬鬆

AI 資本支出的運作方式像財政刺激的反向操作。當超大規模雲端業者發行債券或出售國債來資助 AI 基礎設施時，私人部門吸收了存續期間（duration）——但關鍵是，一旦乾糧耗盡，每一塊錢流入 AI 就必須從其他地方抽出來。 這造成資本的凸性爭奪（convex scramble）。

類比關係： - 財政刺激： 政府發行國債 → 私人部門吸收存續期間 → 現金循環產生乘數效應 → 資產價格淨正面影響 - AI capex（早期階段）： 超大規模業者發債 → 私人部門吸收存續期間 → 現金循環產生乘數效應 → 資產價格淨正面影響 - AI capex（現階段）： 沒有乾糧了 → 被迫出售現有資產 → 負面投資組合平衡效應 → 反向 QE

與財政刺激不同（聯準會成為最終吸收者），AI capex 迫使資本直接競爭不同用途。

資本短缺的連鎖效應

最深的口袋已經見底。 當 Altman 向沙烏地和軟銀尋求承諾時，他們不再有閒錢——必須出售資產來釋放資本：

  1. 投資組合清算： BTC（表現不佳）、SaaS 股票（面臨顛覆風險）、對沖基金贖回
  2. 基金被迫賣出： 對沖基金為了滿足贖回而賣出資產
  3. 連鎖效應： 價格下跌 → 保證金緊縮 → 更多賣出 → 信心崩潰

這就是為什麼新聞報導指出全球最深的口袋（沙烏地、軟銀）已經「基本耗盡」。

Warsh 問題：錯誤診斷

Trump 選擇 Warsh 加速了危機。Warsh 認為問題是「錢太多」，但實際問題是資本短缺。 這個誤診驅使政策往完全錯誤的方向走，在需要放鬆時反而緊縮。

自 Warsh 被選定以來市場加速變化的時間點，證實了這個動態。

資產分化：贏家與輸家

當資本變得稀缺，貼現率上升，市場迫使對資本最有效用途進行殘酷評估：

贏家（短期現金流）： - ✅ DRAM/HBM/NAND 製造商（SNDK, MU）：現在賺超賺，儘管長期前景是週期性的 - 今天的現金流 > 投機性的未來價值

輸家（長期投機）： - ❌ 加密貨幣：「流動性條件的矛尖」→ 感覺無底洞 - ❌ 投機性散戶動能股： 即使基本面改善也無法守住漲幅 - ❌ SaaS 軟體包袱： 面臨 AI 顛覆風險

市場整體效應： - 主權債和信用債殖利率上升（資金需求 > 供給） - 高度投機資產受到不成比例的懲罰

流動性類比：浴缸裡的水

貨幣/流動性 = 水。金融資產 = 浮在浴缸裡的橡皮鴨。

要提高資產價格（更高的水位），你可以： 1. 增加總水量（降息/QE） 2. 疏通流入的管道（管道措施，如 RRP） 3. 減少流出的水

當前問題： 對貨幣的需求太大，造成排擠效應。討論焦點在供給，但需求同樣重要。

投資策略：防禦與耐心

作者立場： - 大部分持現金（在頂部賣出黃金/白銀後） - 不急著買入任何東西 - 不是交易建議 — 提供理解脈絡

建議： - 不要自滿地做多 - 採取防禦並精挑細選 - 積極管理風險 - 保持耐心： 今年會出現非凡機會

核心洞察

這不是暫時性的下跌或修正——這是市場結構性轉變。驅動 2010 年代的資本輕量型 web 2.0/SaaS 典範允許過剩資本流入投機。資本密集型的 AI 典範反轉了這一點，迫使資本從投機流出並進入基礎設施。

典範轉移： - 2010s：資本過剩 + 生產力短缺 → 投機繁榮 - 2020s：資本短缺 + AI capex 需求 → 投機死亡

Key Quotes

"We are experiencing a market regime change due to the shortage of financial capital due to the AI capex cycle."

"The web 2.0 and SaaS paradigms that powered the 2010s market boom were super capital light, allowing an excess of capital to flow into speculative assets."

"Once the dry powder is gone, any dollar going into AI capex needs to come out of something else. It forces a convex scramble for capital."

"AI capex has served as a form of inverse QE with negative portfolio balancing effects."

"There have been press reports that the deepest pockets in the world, the Saudis and Softbank, are largely tapped out."

"Altman goes to them with his hands out and begs them to follow through on past committments. Unlike past windows when they had dry powder, they now have to sell something in order to give him the money."

"Trump selected Warsh. This is especially problematic because he believes the problem is that we have too much money, when in actuality we are facing the opposite issue."

"When the cost of capital goes up, this raises the discount rate. Long duration speculative assets are punished, while assets with near term cash flows benefit."

"Crypto naturally gets annihilated in this environment, as the tip of the spear for liquidity conditions."

"This is not a time to stay complacently super long. It's a time to play defense, be very selective about what you own, and manage risk."

"I am mostly in cash after selling gold and silver at the top. I'm not in a hurry to buy anything. I believe there will be extraordinary opportunities this year if you are patient."

Personal Reflection

Why This Matters

This post reframes the current market volatility not as noise but as signal of structural change. Three reasons this analysis stands out:

  1. Mechanism clarity: The AI capex = inverse QE framing is brilliant. It explains why the market felt like incremental stimulus early on, then suddenly reversed. The "dry powder exhaustion" threshold is the inflection point.

  2. Falsifiable thesis: We can test this by watching:

  3. Continued struggles by large capital allocators (Softbank, sovereign wealth funds)
  4. Persistent bifurcation between near-term cash flow stocks and speculation

  5. Correlation between AI capex announcements and broader market weakness

  6. Actionable insight: The author isn't just describing the problem—he's positioning for the next cycle. "Extraordinary opportunities if you are patient" implies he's waiting for capitulation/forced selling to create mispricings.

Questions This Raises

  1. What breaks the cycle?
  2. Does AI capex plateau? (Demand saturation, project delays, ROI disappointment)
  3. Does the Fed pivot to accommodate? (Warsh's "too much money" diagnosis suggests no)

  4. Do new capital sources emerge? (Middle East tapping oil revenues again, China reopening capital flows)

  5. Investment implications:

  6. If crypto is "tip of the spear," does it bottom first when liquidity returns?
  7. Are DRAM/HBM manufacturers a value trap? (Overearning now, cyclical later)

  8. What's the timing on "extraordinary opportunities"? Q2? Q3? 2027?

  9. Broader implications:

  10. If AI capex is crowding out everything else, does this slow non-AI innovation?
  11. Does this accelerate consolidation? (Only hyperscalers can afford AI infra)
  12. What happens to startups in a capital-scarce environment?

What I'd Watch

  • Hyperscaler capex guidance: Any cuts = confirmation of capital constraints
  • Credit spreads: Widening = capital shortage intensifying
  • Middle East sovereign wealth fund flows: Are they net sellers? Of what?
  • Warsh policy signals: Does he double down on tightening or pivot?
  • Crypto capitulation: When does it stop going down? (Potential leading indicator of liquidity returning)

This is the kind of analysis that separates signal from noise. Plur_daddy isn't predicting prices—he's explaining the structural forces that will determine prices over the next 12-24 months.