Argument and analysis from tek54's columnists.
Reading Beijing's H200 rationing as a market reopening misses what actually happened: export control has become industrial allocation, run on both ends by discretion no one can audit. A cap set at "less than half what they asked" is a dial, not a policy.
SK hynix is a great company that just sold $26.5 billion of new shares into the greatest memory shortage on record. The record is a fact about the moment, not the company — and the proceeds are the tell.
Together AI raised $800 million and tripled its valuation in sixteen months renting out Nvidia's chips. Whether the company is good was never the question. The question, as always, is the price — and who's already long.
Four senators from both parties agree the CREATE AI Act should make NAIRR permanent. Treating that quiet consensus as proof the loud AI fights are just as solvable would be a category error of its own.
Amazon is spending $1 billion to put forward-deployed AI engineers inside its customers. Strip away the language of partnership and it's the oldest power move in enterprise tech, running one layer above the chips.
A $269 billion single-session drop wasn't a verdict on Google's chips. It was the market repricing the assumption underneath the whole AI trade: that the people who make the compute worth anything come with the building.
Jalapeño is being read as independence from Nvidia. Read it again as a map of who can still say no to OpenAI.
The robotaxi recall is being read as a safety story. It's something more important: a quiet answer to the only question that matters about autonomous systems — who decides when they're fixed.
The two popular framings — block the data centers, or clear the runway for them — are both category errors. What the commission actually did on June 18 is harder to chant and more likely to work, if its own deadlines don't wreck it.
The cash-rich chipmaker just sold its biggest-ever slug of bonds. The more interesting debt is the borrowing it isn't on the hook for — the loans its lease commitments quietly make safe. We have seen a vendor underwrite its own demand before.
Everyone is reading the Cursor deal as Musk's late entry into the AI coding race. Read the structure instead: four days after going public, he turned his own stock into an acquisition engine. That is a power move, not a product one.
On June 22, every S&P 500 fund on earth must buy Marvell, a chip designer whose stock has roughly tripled this year. Forced buying is not the same as conviction — and the gap between them is where the AI trade keeps its risk.
The moratorium fight treats "AI regulation" as one switch to flip on or off. The laws actually moving — on chatbots, on hiring, on kids — are narrow, old, and mostly about who is accountable. Missing that is the category error.
OpenAI caught China-linked accounts amplifying America's data-center backlash. The comfortable lesson is that the backlash is manufactured. Learn that lesson and you lose both the information fight and the siting fight.
Morgan Stanley expects AI debt issuance to nearly double to $570 billion this year. An equity mania costs the believers. A debt mania sends the bill to everyone else.
Strip away the farewell keynote and the new Siri demo, and WWDC 2026 confirmed a power shift: the most valuable company in the world now leases the model, the cloud, and the chips behind its flagship AI from its chief rival.
SpaceX, OpenAI and Anthropic are walking through the same door at once. The door is the tell. The trillions are the question.
The chips are real and the demand is real. That was never the question. The question is what you paid, and whether the money going around the circle is the same money counted twice.
Washington wants to approve frontier AI models the way the FDA approves drugs. The analogy is comforting, popular, and wrong about how software actually fails. I say this as someone who would have been the agency.
A trillion dollars of concrete and silicon isn't a means to a plan. It is the plan — a bet, made with the balance sheet rather than the slogan, about who gets to decide.
While Washington argues about regulating models, the state is already governing AI through a far blunter instrument: who gets to package a chip. Get that frame wrong and you mistake the sideshow for the policy.
Strip away the mission statements and the model demos. The only durable asset left in artificial intelligence is the compute — and the contracts for it are already signed.
