Jim Rickards has been pointing at July 29, 2026 since March. Four months of press releases, a free video presentation, and a sustained ad campaign all converge on one date and one event: Meta Platforms reporting Q2 earnings after the closing bell.

The prediction is specific. Meta reports, and if the numbers disappoint, the AI debt thesis Rickards has been building gets its first real test. He compares it to the dotcom moment when Cisco warned it could not see demand in December 2000 and the Nasdaq’s unwind accelerated.

As of July 15, Rickards had released three new press angles on July 14 alone.

The prediction in plain terms

Rickards says $200 billion in AI-related debt was raised in 2025. The broader Jim Rickards AI debt warning lays out the full case. By his updated figures from July 2026, AI companies issued $236 billion in debt in the first five months of this year alone. The pace is accelerating.

He argues that much of this debt sits in off-balance-sheet structures: special-purpose vehicles, synthetic leases, and data-center lease securitizations that bundle cash flows into bonds. The structures resemble what Enron used to hide debt in the 1990s and what Lehman used to hide leverage before 2008. The full AI Black Paper Jim Rickards presentation walks through those structures in detail.

The July 29 prediction is about timing. Rickards says Meta is the most leveraged of the AI hyperscalers. The company carries roughly $87 billion in total debt including lease obligations, against 2026 CapEx guidance of $125 to $145 billion. Reality Labs has lost close to $80 billion cumulatively since 2020. If Meta misses on revenue or guides Q3 below consensus, the market reprices the AI spending cycle and the debt structures underneath it face scrutiny. The prediction comes down to a specific company, a specific date, and a specific mechanism linking them together.

The new angles from July 14

Rickards dropped three fresh press releases on July 14, each adding a layer to the thesis as the date approaches.

The first angle: AI companies are spending billions on worker training and reskilling. Rickards calls this paradoxical. If AI replaces workers, why invest so heavily in the people it supposedly replaces? His answer is that adoption is harder than the technology. Companies need employees who can actually use AI, and without them, the productivity gains investors are pricing in may not materialize. Stock prices assume full-speed adoption. If adoption takes longer, those prices get reconsidered.

The second angle: the biggest AI companies are investing in each other. Money flows from one hyperscaler to another and back again. Rickards asks whether this reflects genuine outside demand or cash moving in a circle. Circular spending inflates revenue figures without proving that real customers are paying real money for real products.

The third angle: one major technology company spent $55.7 billion on capital expenses in its most recent fiscal year, blowing past a $50 billion target. Next year’s guidance reached as high as $95 billion. Rickards frames this as the AI boom entering a phase where spending discipline matters more than spending promises.

All three angles point back to July 29. The earnings reports will show whether spending is producing returns, whether revenue comes from outside customers or circular deals, and whether the borrowing pace is sustainable.

Why this date and not another

Rickards chose Meta deliberately. Among the Magnificent Seven, Meta combines the highest debt load, the largest CapEx guidance, the biggest non-core cash burn through Reality Labs, and a stock already down roughly 15% year-to-date. No other AI hyperscaler carries that combination of leverage and spending commitment. He is one voice in a wider AI bubble debate where four camps are converging on the same destination from different angles.

Meta reports after the bell on July 29. The after-hours window matters because that is when guidance gets released. Revenue beats can mask guidance weakness, and guidance is what moves the AI spending narrative. If Meta beats Q2 revenue but cuts Q3 guidance or raises CapEx again, the market reads that as spending outrunning returns. If Meta misses revenue entirely, the “dotcom moment” frame gets traction.

Rickards on dated calls

The July 29 prediction fits a pattern in Rickards’ career. He attaches dates and mechanisms to his forecasts.

In 2006, he sent a warning to CIA officials that a financial crisis was coming within two years. Lehman collapsed in September 2008. In January 2020, he published a note through his newsletter warning of a contagion event. The S&P 500 peaked on February 19 and crashed 34% over the next 23 trading days. In 2016, he published The New Case for Gold when gold traded near $1,200 and called for $10,000. The same dated-call structure shows up in his gold-mining thesis pitch, where he attaches a number and a timeline to a single mining claim. Gold crossed $5,000 in early 2026 and Rickards maintains the $10,000 target.

The calls are not all perfectly timed. Rickards was early on the gold call by several years. His currency war predictions have played out in slow motion rather than the acute crisis he described. The 2008 warning was directionally correct but the timing was approximate.

The July 29 prediction is different because it is tied to a calendar event. Meta’s earnings date is fixed. The results are public. Within hours of the report, the thesis either gets validated, deferred, or weakened. There is no ambiguity about when the test happens.

What to watch

The prediction resolves on July 29 after the bell. Four numbers matter.

Meta’s Q2 revenue against the $58 to $61 billion guidance. Q3 revenue guidance against consensus. CapEx commentary for the back half of 2026. And free cash flow, which shows whether the spending is generating cash or burning it.

If all four come in strong, Rickards’ thesis waits for the next earnings cycle. If revenue misses or guidance disappoints, the AI debt structures he has been documenting get their first stress test. The $236 billion in AI debt issued in five months becomes the number everyone suddenly wants to understand.

Rickards has been early on calls before, and he has also been right. July 29 is the date he chose, and the earnings report will give the thesis its first real test. The revenue numbers, the guidance, and the free cash flow will tell the rest of the story.