
Michael Burry sent readers to a BIS report on AI's poorly disclosed deals. Oracle (ORCL) reports $638 billion in one line.
Michael Burry spent July 23 pointing his readers at a Bank for International Settlements report on how AI deals get financed. The clearest illustration sits in an Oracle (ORCL) filing: a $638 billion backlog reported as one aggregate line, with the customer split left to S&P to estimate.

Everyone's asking if AI is the next dot-com bubble. Nvidia and 25 companies just gave a clue.
Nvidia rallied 25 companies behind open AI models, and OpenAI and Anthropic sat it out. Here's why I think their coming IPOs, not this fight, mark the real top of the AI bubble.

Jim Cramer's tech calls are still lagging the rest of his book. Intel is 1 for 5, and IBM crashed the day after his second buy call.
Jim Cramer's tech calls are still his weak spot. Intel is 1 for 5 since June, IBM is down nearly 30% since his back-to-back buy calls, and his tech picks trail the Nasdaq-100 by more than three points a call even after Tuesday's chip bounce.

Michael Burry cut his Oracle (ORCL) short in half, and says reports he closed it are fake news
Michael Burry sold half his Oracle (ORCL) puts on July 17 after a 63% stock plunge, then called reports that he'd fully exited the short "fake news." The same post added Nvidia (NVDA) puts, more DraftKings (DKNG), and a skeptical read on Netflix (NFLX).

Nvidia (NVDA) and Microsoft (MSFT) don't just sell OpenAI chips and cloud time. They also fund the bill.
Nvidia, Microsoft, Oracle, Broadcom, AMD, and CoreWeave have promised OpenAI close to a trillion dollars combined, and some of them are OpenAI's investors as well as its vendors. Here's how the money actually loops, and where it's already showing strain.

New Mexico blocked the pipeline feeding Oracle's (ORCL) biggest Stargate site. Bloom Energy (BE) has 2.45 GW riding on it.
New Mexico's land commissioner denied the Green Chile gas pipeline for Oracle's Project Jupiter Stargate site a second time, and SemiAnalysis now sees a 1 to 2 year delay risk. That puts Bloom Energy's record 2.45 GW fuel cell order, and a $61 billion market cap built on AI power deals, on the clock.

Kimi K3 just gave Nvidia (NVDA) its second "DeepSeek moment." The first one made chip demand go up, not down.
Moonshot AI's Kimi K3 sent Nvidia and Oracle lower this week, reviving the "cheap AI kills chip demand" argument. History, and the actual spending numbers, say the opposite is true.

Bullish Watch, week of July 13: the market panicked over a Meta headline, and got it wrong
This week's watchlist, updated through Tuesday: Nebius gave back its early bounce on a data-center headline unrelated to its business, Kraken Robotics is holding up best of the five, Intel swung from a 6% drop to a 4.5% rally in two days, and Rocket Lab and Harmonic are still finding their footing.

Oracle (ORCL) stock crashed 59%: is it near the bottom?
Oracle keeps beating earnings and keeps falling anyway, down from $345.72 to around $140. Here is why, the debt and cash burn behind it, the $638 billion bull case, and whether the stock is bottoming.

The AI stocks that led and lagged the first half of 2026
The first half of 2026 split the AI trade in two. Memory and chip-supply names like Micron (MU, +304%) and Intel (INTC, +278%) soared, while marquee names lagged and AI software fell, with Palantir (PLTR) down about 34%. Here is the full scoreboard and what it means.

Smart money is piling into AI neoclouds: Nebius NBIS, SharonAI SHAZ and Cerebras CBRS
Several top funds disclosed 5%+ stakes in AI compute names this spring: Leopold's Situational Awareness in Nebius (NBIS) and SharonAI (SHAZ), Tiger Global in Cerebras (CBRS), and Lone Pine in miners turned AI data centers TeraWulf (WULF) and Hut 8 (HUT).

Green in a sea of red: why Microsoft is bucking the tech selloff today
Tech is bleeding on a memory-chip rout, but Microsoft is green. A deep dive into the rotation lifting MSFT and asset-light software while chips and Oracle get sold.

The AI stock map: every layer of the trade, from sand to software
A top-to-bottom deep dive on the AI supply chain, from the machines that make the chips down to the hyperscalers writing the checks, and how every layer pushes and pulls on the others.