# AI Bubble Index

**56 out of 100 — Heated.** Last computed August 28, 2026 at 6:35 PM ET.

Multiple indicators elevated together.

Source: https://aistockwire.com/ai-bubble-index
Publisher: AIStockWire. Free to cite with attribution and a link.

## What it measures

A composite reading of speculative risk in the AI trade, built from the mechanisms that ended the dot-com bubble rather than from stock prices alone. Two of the six components come from grid, utility and SEC filings and move only when something is filed; the other four move with the market. The score recalculates four times each weekday.

Each component is mapped to 0-100 through a piecewise-linear anchor table tied to stated observations, then combined as a weighted average. Weights and anchors are below in full.

## Components

| Component | Reading | Score | Weight | Updates |
| --- | --- | --- | --- | --- |
| Commitment gap | 65.41 % of announced demand unbacked | 62 | 22% | quarterly |
| Valuation | 19.04 median forward P/E of the AI basket | 33 | 17% | daily |
| Insider selling | 1.1847 discretionary selling against its own trailing baseline | 46 | 13% | daily |
| Counterparty concentration | 47.5 top customers as % of revenue, and vendor commitments as % of revenue | 67 | 13% | quarterly |
| Momentum | 1.1141 basket price divided by its 200-day average | 54 | 13% | daily |
| Complacency | 14.43 VIX, inverted | 69 | 8% | daily |
| Contracted cover | 21.67 contracted revenue as % of market value | 69 | 14% | quarterly |

### Commitment gap — 62 of 100

Is the demand behind the buildout real?

- Reading: 65.41 % of announced demand unbacked
- Weight: 22 percent
- Updates: quarterly
- Anchors: 0 maps to 0; 50 maps to 40; 75 maps to 75; 95 maps to 100
- Why those anchors: 50 percent maps to 40 because roughly half of announced data center projects have historically failed to arrive on schedule, so that much attrition is the baseline rather than a warning. 95 percent maps to 100.
- Detail: median of 7 independent screens run by grid operators, utilities and regulators

### Valuation — 33 of 100

How much is priced in?

- Reading: 19.04 median forward P/E of the AI basket
- Weight: 17 percent
- Updates: daily
- Anchors: 16 maps to 25; 25 maps to 50; 40 maps to 80; 60 maps to 100
- Why those anchors: 16 is the long-run S&P 500 median forward multiple and maps to 25. 60 maps to 100, in the region of dot-com peak multiples.
- Detail: median of 10 forward P/E readings across 10 symbols

### Insider selling — 46 of 100

Are the people inside selling more than usual?

- Reading: 1.1847 discretionary selling against its own trailing baseline
- Weight: 13 percent
- Updates: daily
- Anchors: 0.4 maps to 10; 1 maps to 40; 2 maps to 70; 4 maps to 100
- Why those anchors: The last 90 days of discretionary selling against the median of the prior four 90-day windows, so the basket is judged against itself rather than an outside benchmark. A reading of 1.0 is its own normal rate and maps to 40. Rule 10b5-1 plans and tax withholding are excluded.
- Detail: $500M of discretionary selling across 10 of 25 AI infrastructure names in the last 90 days, against a baseline of $422M from 4 earlier windows. NVDA is 82 percent of the current window.

### Counterparty concentration — 67 of 100

How few counterparties is this resting on?

- Reading: 47.5 top customers as % of revenue, and vendor commitments as % of revenue
- Weight: 13 percent
- Updates: quarterly
- Anchors: 15 maps to 20; 30 maps to 45; 54 maps to 75; 70 maps to 100
- Why those anchors: 15 maps to 20 because that is what Lucent's largest customer reached before its collapse, per its fiscal 2000 annual report. The score averages customer concentration with vendor financing commitments measured against revenue.
- Detail: average of 54% customer concentration (NVIDIA 10-Q, three direct customers) and 41% vendor commitment to revenue (the $105B OpenAI guarantee against $253.4B trailing revenue)

### Momentum — 54 of 100

How far above trend is the price?

- Reading: 1.1141 basket price divided by its 200-day average
- Weight: 13 percent
- Updates: daily
- Anchors: 0.9 maps to 10; 1 maps to 30; 1.1 maps to 50; 1.2 maps to 75; 1.4 maps to 100
- Why those anchors: A basket trading at its 200-day average maps to 30. Twenty percent above trend maps to 75.
- Detail: median price to 200-day average across 10 symbols

### Complacency — 69 of 100

Is any of this priced as a risk?

- Reading: 14.43 VIX, inverted
- Weight: 8 percent
- Updates: daily
- Anchors: 10 maps to 100; 12 maps to 85; 15 maps to 65; 20 maps to 35; 30 maps to 10; 45 maps to 0
- Why those anchors: Inverted on purpose: a low VIX means the market is pricing no stress, which raises the reading rather than lowering it. 20 is near the long-run average and maps to 35. 45 maps to 0, the level reached at the dot-com bust peak.
- Detail: CBOE Volatility Index, last regular-session print

### Contracted cover — 69 of 100

How much disclosed contract coverage supports the valuation?

- Reading: 21.67 contracted revenue as % of market value
- Weight: 14 percent
- Updates: quarterly
- Anchors: 10 maps to 100; 25 maps to 60; 50 maps to 25; 100 maps to 0
- Why those anchors: Inverted: cover falls as the market pays more for demand nobody has signed for, so a low reading raises the score. 100 percent, where signed contracts equal the whole market value, maps to 0; 10 percent, where nine tenths of the price rests on uncommitted revenue, maps to 100. Only 10-Q and 10-K figures count.
- Detail: $2932B of disclosed GAAP obligations against $13531B of market value across 18 companies

## Bands

| Range | Label |
| --- | --- |
| 0 to 30 | Cold |
| 31 to 50 | Building |
| 51 to 70 | Heated |
| 71 to 85 | Overheated |
| 86 to 100 | Euphoria |

## History

| Date | Score |
| --- | --- |
| 2026-08-20 | 56 |
| 2026-08-21 | 59 |
| 2026-08-22 | 59 |
| 2026-08-24 | 56 |
| 2026-08-25 | 55 |
| 2026-08-26 | 55 |
| 2026-08-27 | 56 |
| 2026-08-28 | 56 |

## Primary sources behind the filing-based components

- ERCOT, Assessing the Grid, presented to the Texas Senate Committee on Business and Commerce (July 29, 2026) — https://www.ercot.com/files/docs/2026/07/29/ERCOT-Senate-July-29-Panel-1-Assessing-The-Grid.pdf
- Georgia Power, Quarterly Large Load Economic Development Report, Q1 2026, Georgia PSC Docket 56002 — https://services.psc.ga.gov/api/v1/External/Public/Get/Document/DownloadFile/226607/107887
- AEP Ohio (February 13, 2026) — https://www.aepohio.com/company/news/view?releaseID=10753
- PPL Electric Utilities, second quarter 2026 disclosure — http://paenvironmentdaily.blogspot.com/2026/08/ppl-electric-utilities-reports-318-gw.html
- Commonwealth of Pennsylvania, Executive Order 2026-05 (August 18, 2026) — https://www.pa.gov/governor/newsroom/2026-press-releases/governor-shapiro-signs-executive-order-on-data-center-developmen
- ERCOT, Report on the Capacity, Demand and Reserves in the ERCOT Region, 2026 to 2030 (December 19, 2025) — https://www.ercot.com/files/docs/2025/12/19/CapacityDemandandReservesReport_December2025.pdf
- NERC, 2025 Long-Term Reliability Assessment — https://www.nerc.com/globalassets/our-work/assessments/nerc_ltra_2025.pdf
- NVIDIA Corporation, Form 10-Q for the quarter ended April 26, 2026 — https://www.sec.gov/Archives/edgar/data/1045810/000104581026000052/nvda-20260426.htm
- NVIDIA Corporation, Form 8-K, August 17, 2026 — https://www.sec.gov/Archives/edgar/data/1045810/000104581026000069/nvda-20260817.htm
- SEC Form 4 filings, collected continuously — https://aistockwire.com/insiders

## Limitations

- The index is backward-looking. It cannot time a correction, and it does not try to.
- Two components update only when a utility, regulator or company files something, so the level can sit still while the market moves.
- The weights and anchor points are editorial judgments, not model output.
- This is a record of published data, not investment advice.