Today I'm publishing a new essay, Policy on the AI Exponential. AI is progressing extremely fast—much faster than the policy process was built to handle. The essay lays out where I think the technology is now, and the action needed to close the gap: darioamodei.com/post/policy-…

Burry's thesis: customer concentration receivables slope = channel stuffing risk. The data he cites is real. Top 3 customers = 64% of Nvidia's receivables, up from 33% in 2020. That's a steep concentration curve. But the thesis has a structural blind spot: GPU depreciation economics. Burry treats AI accelerators like commodity hardware that loses value linearly. That's not how this works. An H100 or Blackwell cluster isn't just silicon. It's a revenue-generating engine tied to proprietary software stacks, trained datasets, and network topology. A "used" GPU in a functioning AI cluster retains value as long as it powers a profitable LLM. The economic floor is far higher than traditional depreciation models assume.

Make sure your business is ready to accept AI agents as customers!

Bitcoin was sold as the asset that survives geopolitical chaos. The chaos arrived. Bitcoin bled. The US bombed Iran. Iran fired ballistic missiles at US forces in Kuwait. Israel pushed into Lebanon. The Strait of Hormuz has been effectively closed for ninety-four days. Oil ripped above $90. Bitcoin slid below $70,000 with the longest ETF outflow streak in its history. Ten days. $2.97 billion out. A $1.3 billion BlackRock IBIT block sale through a dark pool. The S&P 500 closed at an all-time high. The Nasdaq closed at an all-time high. The ninth straight green week on the S&P, the longest streak since 2023. Nvidia ripped 6 percent on a new PC chip. Dell rose 11. HP rose 9. Berkshire anchored Alphabet’s $80 billion equity raise with $10 billion at a discount. That divergence is the trade. Gold rose. Bitcoin fell. That is not the verification trade Bitcoin promised. That is the risk trade Bitcoin became. The wrapper is the giveaway. The ETF that was supposed to bring Bitcoin to institutional grade has become its exit ramp. Institutional money does not buy Bitcoin to ride a war hedge. It buys the wrapper for exposure, sells it for liquidity, and rotates the proceeds wherever the duration story is paying best. This week, that meant out of Bitcoin and into AI compute. Same week, Michael Saylor’s Strategy sold 32 Bitcoin. First sale in four years. The most religious never-sell balance sheet in the asset just sold. The narrative break matters more than the size. Now look at where the rotation lands. Alphabet is raising $80 billion in equity to fund the AI buildout. Berkshire just anchored at a roughly six percent discount. Nvidia just pledged $150 billion a year in Taiwanese suppliers. The capex flows into chips. The chip orders flow into the same special purpose vehicles, private credit pipes, and insurance balance sheets that have been routing residual GPU risk into American retirement annuities for months. The institutional dollar leaving Bitcoin is, in real time, rotating into the AI capex cycle that ends inside the same Athene-Bermuda pipe Burry traced last week. Two stories. One verification regime. The story that wins is the one that lets the slowest-clock holder absorb the residual. War did not interrupt the rally. War accelerated the rotation. The S&P at 7,600 and Bitcoin at $70,000 are the same data point read in two languages. open.substack.com/pub/shanak…

Anthropic is too expensive and will either lose customers or cut prices

DTCC and the Stellar Development Foundation announced today plans to enable the tokenization of DTC‑custodied assets on the @StellarOrg network. This collaboration advances DTCC’s multi chain strategy and expands how traditional assets move across digital ecosystems. DTC‑tokenized assets are expected to be made available on the Stellar network in the first half of 2027, supporting the evolution of a more open, interoperable, and efficient financial ecosystem. Get the full story: dtcc.com/news/2026/may/27/to…

Goldman Sachs CEO, David M. Solomon on nytimes "A.I. won’t eliminate 25% of jobs. What’s more likely is that people will find more productive ways to spend their time. When I was a first-year banking analyst, something as simple as making a graph of a stock’s performance took six hours of looking up prices in back issues of The Wall Street Journal on microfiche. Today, a first-year analyst can do it in seconds, and we have employed more people than ever in recent years. With more sophisticated tools, the complexity of our work naturally expands. Do any of us feel like we have less to do these days despite the convenience of Excel, email or Zoom?" --- nytimes .com/2026/05/22/opinion/ai-job-crisis-goldman-sachs.html?smid=nytcore-ios-share

A single country currently controls approximately eighty-five to ninety-one percent of global rare-earth refining capacity. “Single-source monopolies that could deny us these things, not just in a time of conflict, but as a leverage point contrary to our sovereign national interests.” That is the verbatim statement issued on Tuesday May 26, 2026, by United States Secretary of State Marco Rubio at Hyderabad House in New Delhi, at the signing of the bilateral India-US Critical Minerals Framework after the Quad Foreign Ministers meeting. On the same day, in the same city, the Quad announced up to twenty billion dollars in coordinated mobilization across mining, processing, recycling, and related investments in the rare earth and critical minerals supply chain. The bilateral framework was signed by Rubio and Indian External Affairs Minister Subrahmanyam Jaishankar. It is the second such bilateral the United States has signed in approximately seven months. The first was the US-Australia framework of October 20, 2025. On February 4, 2026, the State Department hosted the Critical Minerals Forum in Washington. On February 20, 2026, India joined the US-backed Pax Silica initiative. India holds significant rare-earth and monazite reserves and produces less than one percent of global rare-earth output. The framework signed this week is not aimed at the deposits. It is aimed at the processing bottleneck. The Secretary of State’s single sentence at the podium is the doctrine the entire sequence of bilateral, plurilateral, and multilateral frameworks is built to escape. The variable that prices access to critical minerals is no longer ownership of the deposit. The variable is admissibility to the processing network that converts the deposit into refined material. This is the same logic that now governs the Strait of Hormuz, where the Persian Gulf Strait Authority and the United States blockade impose mirror admissibility conditions on the same chokepoint from opposite sides. It is the same logic governing the Taiwan arms pause, the Federal Reserve payment-account proposal, the regulated dollar stablecoin compliance regime, the silent fragmentation of the equity narrative on the Friday the S&P touched a new all-time high, and the USCIS memorandum that ended a seventy-four-year green card pathway. One week. Six surfaces. One variable. The variable is admissibility. This is not a story about a diplomatic visit. It is a story about the conversion of every commitment in the global system, including the commitment to supply a foundational material, from ownership of the resource to admission of the resource. The framework is architectural, not accusatory. open.substack.com/pub/shanak…

Looking ahead, forecasts for tokenized assets vary a lot but they all point in the same direction: growth. McKinsey: $2–4T by 2030. Ark Invest: $11T by 2030. BCG/Ripple: $9.4T by 2030, $18.9T by 2033. Standard Chartered: $30T by 2034. The gap between $2 trillion and $30 trillion is more about definitions than adoption. Different institutions are measuring different things. McKinsey focuses mostly on bonds, loans, funds, and equities. Standard Chartered adds commodities and trade finance. BCG and Ripple include deposits and stablecoins alongside more traditional asset categories. Despite these differences, the broader trend is consistent: Asset tokenization is expected to expand.

They say AI will kill the middle class. I say, the middle class was already dead. Trading 40 years for a pension that inflation eats alive isn’t security. It’s slow surrender. AI didn’t break the system. It just turned the lights on.

Wealth *is* zero-sum in any closed financial system—like a national balance sheet—because net financial wealth (assets minus liabilities) must sum to zero across all agents. Income is flow; wealth is stock. You can earn $100,000/year for decades and still hold *zero or negative net worth* if debt offsets assets. The Fed’s 2023 Survey of Consumer Finances confirms the bottom 50% of U.S. households holds *negative net worth*: median net worth = –$2,500, with auto loans, medical debt, and credit card balances exceeding value of vehicles, used furniture, and bank accounts. Their income services debt held by others—not wealth accumulation.