AI ECONOMICS: The bill for “tokenmaxxing” just came due.
Meta’s employees ran up a staggering internal AI bill in a single month. The number is reshaping how Big Tech thinks about the cost of intelligence.
In case you missed it...Samsung Electronics, SK Hynix, and Micron Technology were sued on June 25 in federal court over allegations of collusion and price-fixing in the commodity memory market.
Filed in the Northern District of California, the complaint names 17 plaintiffs — a mix of individuals and small businesses — who claim the three chipmakers conspired to keep commodity DRAM artificially scarce in order to inflate prices, according to Wccftech Source. Yahoo Finance
Hedge funds are abandoning US tech stocks at the fastest pace in over a decade:
The Technology, Media, and Telecom sector has seen the largest and most sustained outflows of any US sector in 2026. The net selling accelerated sharply in June to nearly -3% of total US gross market value, according to Goldman Sachs. Last week, hedge funds sold the most US Information Technology stocks in more than 10 years, in both dollar and percentage terms. This was driven by long and short sales at a ratio of roughly 1.3 to 1. Semiconductors and semiconductor equipment accounted for more than half of the total tech sales, having now been net sold for 8 consecutive sessions. Hedge funds are running to the exit in US tech. Source: Global Markets Investor, Goldman Sachs
In case you missed it... US MegaCap Tech is facing its worst month since April 2022
(and 3rd worst month ever)... Source: zerohedge
The software squeeze
At range highs, it is difficult to get greedy. Source: TME, LSEG Workspace
Nvidia will launch a PC “superchip” this year as the semiconductor giant goes head-to-head with Apple, Qualcomm, Intel and AMD for the first time.
Computer makers including Dell, Asus and HP will use what Nvidia claims is “the most efficient PC chip ever built”, paired with Microsoft’s Windows software, Nvidia chief executive Jensen Huang announced on Monday. The launch marks a major competitive shift in the consumer PC industry and a new business line for $5.1tn Nvidia. The company, best known for its dominance of semiconductors for AI infrastructure, is pushing beyond its traditional graphics cards into integrated chips that power the whole PC. Lenovo, Microsoft, Acer and Taiwan’s MSI will also use the product. Source: FT
Jevon-paradox Token usage (blue bars) is exploding higher. It started in January when Agentic AI went mainstream with Claude Cowork and Moltbook (OpenClaw).
AI users are creating agents and code, leading to exponential growth in AI usage. It's just starting. Source: a16z, Jim Bianco
44% of every dollar companies spend on AI goes directly to fixing bugs that the AI itself created.
A report from Entelligence AI across 2,444 companies shows that for every $1 spent on AI tokens, $0.44 goes to bug fixes, $0.27 to rewriting AI-generated code, and $0.11 disappears into review and merge delays. Companies spending $100,000 on AI tokens and only $18,000 worth is reaching production. The other $82,000 is overhead generated by the tool itself. Lightrun's 2026 report found that 43% of all AI-generated code still requires manual debugging in production even after passing every quality test. Not a single engineering leader surveyed said they were fully confident AI code would behave correctly once deployed. Wall Street is pricing AI as a productivity tool and the data says 82% of the spend never reaches the actual product. SOURCE: @Aiswarya_Sankar Sam Boboev
