AI: The Biggest Capital Misallocation in History | Julien Garran

Feb 5, 2026

Julien Garran, macro strategist and founder at Macro Strategy, warns AI is a historic capital misallocation and argues much of the AI build is fragile. He discusses narrow real‑world utility of LLMs, vendor financing and data‑center risks, scaling limits, and why a major market rotation into resources and emerging markets may follow.

Ask episode

AI Snips

Chapters

Books

Transcript

Episode notes

INSIGHT

AI Capital Misallocation Is Systemic

Julien Garran estimates AI-driven capital misallocation is ~17x worse than the dot-com era, posing systemic macro risk.
He projects AI investment could subtract 3–6% from GDP if the cycle reverses, forcing major policy response.

INSIGHT

Vendor Financing Is A Canary

Most AI businesses are loss-making and reliant on vendor financing, creating fragility when funding tightens.
NVIDIA's receivables surged ~770% and act as a leading indicator of stress, like Cisco's receivables in 2000.

INSIGHT

LLMs Correlate, They Don't Understand

Large language models correlate patterns but lack causal understanding, causing unpredictable failures when applied beyond simple Q&A.
Academic tests show top models complete paid Upwork-style tasks only ~2.5% of the time, effectively near zero for real work.

Get the Snipd Podcast app to discover more snips from this episode

Get the app

1. Strategic Actions and Decisions

* Assess capital allocation: Julien Garran states the current AI-driven capital misallocation is 17x worse than the dot-com era, indicating severe systemic risk. [1:50]

* Model macroeconomic impact: Prepare for scenarios where a slowdown or reversal in AI investment could reduce GDP by 3-6%, necessitating macro intervention.[2:35]

* Evaluate AI vendor financing risk: Monitor “circular vendor financing” (exemplified by NVIDIA’s 770% receivables growth) as a leading indicator of market stress.[10:00]

* Stress-test AI ROI assumptions: Challenge business cases built on generative AI, citing studies showing failure rates of 65-99.7% in real-world applications.[14:00]

* Shift portfolio allocation: Consider a strategic pivot from overvalued AI and tech equities into underinvested resources and select emerging markets.[49:45]

2. Executive Summary

This discussion with Julien Garran presents a critical view of the AI investment boom, framing it as a capital misallocation crisis 17x larger than the dot-com bubble. The argument is that generative AI has fundamental technical limitations—relying on correlation, not causation—which constrain its commercial usefulness. With most players (except NVIDIA) deeply loss-making and reliant on unsustainable vendor financing, a market correction is anticipated. The macroeconomic risk is significant, potentially shaving 3-6% off GDP if the cycle reverses. The proposed strategic response is a major rotation away from AI/tech and into hard assets and emerging markets.

3. Key Takeaways and Practical Lessons

1. Extreme Capital Misallocation: The AI investment frenzy represents a bubble of historic scale compared to previous cycles.

* Practical Lesson: Immediately pressure-test the ROI and capital efficiency assumptions for any AI-related project or investment against stricter, fundamentals-based criteria.

2. Technical Utility vs. Hype: Generative AI’s commercial utility is narrow due to its reliance on probabilistic correlation rather than understanding causality.

* Practical Lesson: Restrict generative AI pilot projects to low-stakes, internal efficiency tasks (like drafting or summarization) and avoid building complex operational workflows on it in the near term.

3. Vendor Financing Red Flags: Rapidly rising receivables in the AI infrastructure sector (notably NVIDIA’s 770% growth) serve as a primary indicator of impending market stress.

* Practical Lesson: Add the receivables and vendor financing activities of major AI infrastructure companies to your financial dashboard as leading risk indicators for the broader tech sector.

4. Data Center Viability: The massive data center build-out carries high execution risk and may be fundamentally unprofitable due to extreme power costs and unsustainable debt.

* Practical Lesson: Scrutinize investments in data center operators and REITs, modeling scenarios where compute demand falls short and rental prices collapse.

5. Imminent Market Inflection: A major shift in market leadership is expected, moving away from tech and into commodities and specific emerging markets like India.

* Practical Lesson: Initiate a strategic review to rebalance portfolios, reducing exposure to cash-burning AI equities and beginning a staged allocation to mining, energy, and emerging market assets.

Visit Macro Strategy here - https://www.macrostrategy.co.uk/team

Watch on Youtube below -

Get full access to The Noble Update at georgenoble.substack.com/subscribe