AI Hype Debt Bubble Detector
Inference-to-Capex Risk Scanner 2026
$3T AI boom hiding inference-to-capex collapse. Scan models, labs, providers -- reveal if your career pivot, investment, or agent stack sits on hype debt.
Tool #4 of the Algorithmic Survival Hub. Economic Survival anchor: Compute Trinity (CDIR+ACS+CBBP) + Bubble Detector (HDRS). Part of 117+ free tools.
AI Hype Debt Bubble Scanner
Enter model, provider, and economic signals. Scan for inference-to-capex collapse risk.
Model Type
Provider
Model Size
70BModel parameter count in billions (1B-1T)
Inference Cost
$2.47Cost per million tokens for inference calls
Training Cost
$247MModel training cost in millions USD
Annual Capex
$4.7BProvider annual capital expenditure in billions USD
Utilization Rate
47%Percentage of deployed compute actually utilized (industry avg ~47%)
Efficiency Risk
4/5Risk of efficiency breakthrough destroying pricing (1=low, 5=DeepSeek-level disruption)
Circular Financing
3/5AI companies funding AI companies (1=organic revenue, 5=circular dependency)
Sentiment Volatility
5/5Narrative-driven valuation risk (1=fundamentals, 5=pure hype)
HDRS
100/100
Risk Tier
CRITICAL
Collapse Prob.
95%
Capex Ratio
0.01x
AI Hype Debt Bubble Detector -- Frequently Asked Questions
What is AI hype debt?
AI hype debt is the gap between what AI companies spend on infrastructure (capex) and the actual revenue their AI models generate. When $3T+ flows into AI capex but utilization rates average only 47%, the difference represents hype debt -- infrastructure built for demand that may never materialize at current pricing.
What is the inference-to-capex ratio?
The inference-to-capex ratio measures how much revenue AI inference generates relative to the capital spent building infrastructure. A ratio above 2x means the provider is spending more on infrastructure than inference revenue can justify -- a classic bubble indicator similar to telecom capex ratios in 1999.
What are circular financing risks in AI?
Circular financing occurs when AI companies fund other AI companies, creating artificial demand loops. When Microsoft invests in OpenAI, then pays OpenAI for Azure AI services, revenue appears organic but is actually circular. This inflates real demand metrics and masks the true utilization gap.
What are utilization cliff warning signs?
Warning signs include: GPU utilization below 50%, declining inference-per-dollar metrics, increasing idle compute capacity, providers offering deep discounts, and efficiency breakthroughs (like DeepSeek) that prove current pricing is unsustainable. The cliff occurs when utilization drops below the breakeven threshold.
How can I reduce hype debt exposure?
Diversify across providers (no single provider should handle more than 25% of spend), migrate inference to efficient alternatives like DeepSeek, build sovereign compute reserves for self-hosted inference, stress-test your AI stack against 50% pricing collapse, and build income streams that benefit from cheaper AI rather than depending on expensive AI.
How do CDIR, ACS, CBBP, and HDRS relate?
CDIR measures compute debt-to-income ratio (affordability). ACS measures agent swarm creditworthiness (reliability). CBBP measures compute collateral value (borrowing power). HDRS measures systemic bubble risk (market stability). Together they form the complete Algorithmic Survival diagnostic -- personal compute health plus market-level risk.
Does model size impact bubble risk?
Yes. Larger models require more inference compute, higher training costs, and more capex to serve. However, model size alone does not determine risk -- efficiency (cost per useful token), utilization rate, and circular financing patterns matter more. DeepSeek proved that smaller, efficient models can disrupt larger models overnight.
Which providers have the highest bubble risk?
Providers with high capex concentration, low utilization rates, and significant circular financing carry the highest risk. The calculator weights each provider differently based on public capex data, pricing trends, and dependency chains. Generally, providers spending aggressively on custom silicon with unproven utilization face higher HDRS scores.