The Schizophrenia Market: When Tariffs Are Really Subsidies
The Schizophrenia Market: When Tariffs Are Really Subsidies
The stock market on January 15th told two contradictory stories and somehow believed both.
In one, a Trump administration moved H200 semiconductors to China—reversing a Biden-era export ban with ceremonial fanfare—then immediately taxed those same chips at 25% on the way out. In the other, BlackRock announced $14 trillion in assets under management and the entire financial establishment nodded along like this was normal consolidation rather than what it actually is: a few institutions now controlling pools of capital larger than the GDP of most countries on Earth.
These aren't separate market movements. They're the same reflex, measured in different currencies.
The Nvidia Theater
Let's unpack the chip deal first, because it's a masterclass in how politics and capital markets have converged into something that looks less like policy and more like performance art.
Trump's team approved Nvidia H200 exports to China. Great. Stabilizes the China market, eases supply-chain tension, gives Jensen Huang a win after months of lobbying from his air-conditioned office. Except—and here's where reading comprehension matters—those chips have to come through the U.S. first. For testing. For "security verification." And at that moment, they attract a 25% tariff.
This isn't a tariff. The U.S. government is taking 25% of the sales value on H200 chips destined for China. Call it what it is: a royalty. A vig. A government fee for permission to sell your own product to your customer.
Nvidia didn't protest. The stock jumped 2%+. ASML hit an all-time high. Why? Because everyone in that ecosystem understands the deeper transaction: this policy gives Nvidia access to a market worth potentially $50 billion per year while creating a revenue stream for Treasury. It's rent extraction dressed up as national security.
And it works. China can't manufacture competitive AI chips at scale. The U.S. can't ban the sale outright without destroying its own semiconductor industry. So we've arrived at this: a managed dependency model where American companies profit, the government skims 25%, and Chinese companies get access to last-generation AI infrastructure while their government loses face—but also gets what they need. Everyone wins by the measure of quarterly earnings. Everyone loses by the measure of actual technological decoupling.
The Russell 2000 has beaten the S&P 500 for 10 consecutive sessions—the longest streak since 1990. Small-cap enthusiasm. Flight to perceived value. TSMC's blowout outlook revived hopes about the longevity of the AI bull-market driver (report this morning showed strength in data-center demand). Broadcom fell 4% anyway. Volatility crept higher. The market priced the H200 deal as net-positive but simultaneously hedged against it, which is exactly what you do when you can't tell if you're looking at policy or theater.
BlackRock's $14 Trillion Problem
Now the other story: BlackRock pulled in $342 billion of client cash in the fourth quarter, pushing the firm to a record $14 trillion of assets.
Fourteen trillion. Let that settle. A single asset manager now controls a pool of capital larger than the entire economy of every nation except the U.S. and China. The company itself doesn't really matter anymore. What matters is what this number represents: the permanent hollowing-out of human decision-making in the investment process.
Investors added $181 billion to BlackRock's exchange-traded fund business, which now has $5.5 trillion overall. ETFs. The most passive possible expression of capital allocation. You don't pick stocks. You don't pick sectors. You pick a fund that moves with the index, and BlackRock sits in the middle, collecting fees on the flow.
BlackRock's own executives dress this up in language about "growth channels" and "private markets" and "tokenization." The press release from January 15 talks about a "unified platform" and "accelerating momentum." But what's actually happening is something darker: the shift from high-fee, opaque active management to transparent, low-cost, and technology-driven solutions has reached its logical conclusion.
This isn't innovation. It's the death of judgment.
Markets used to work like this: humans argued about value, made bets, won and lost fortunes, and prices adjusted. Now they work like this: you buy the index fund, BlackRock compiles your vote with millions of others, and whatever gets funded gets compounded. The algorithm is transparent and it's cheap, which is genuinely good for retail investors. But it means that capital allocation—the fundamental economic question of where resources flow—is no longer debated. It's systematized. Automated. Inevitable.
The TSMC trade will sustain AI infrastructure investment for the next 18 months minimum. BlackRock's new assets will chase that same infrastructure. The two facts contradict in ways that matter (if TSMC is so good why not buy it directly?) but the market doesn't care. The money flows into the ETF. The ETF holds TSMC. The trade works because everyone else is also in the ETF.
This is consolidation of financial decision-making, not innovation.
What Actually Happened This Week
- Stocks rebounded as TSMC assuaged concerns about the sustainability of current data-center spending, with Nvidia jumping over 2% and ASML notching a record.
- The Russell 2000 beat the S&P 500 for a 10th straight session—the longest streak since 1990.
- The Trump administration formalized the H200 China deal with a 25% tariff designed to be paid on chips that get rerouted through the U.S. for "testing."
- China posted a record $1.2 trillion trade surplus as exports beat forecasts, but trade with the U.S. fell sharply, highlighting ongoing tariff frictions.
- BlackRock announced total assets under management surged to $14 trillion by the end of 2025, spurred by a record $698 billion in annual net inflows.
None of these facts is individually false. Together, they describe a system that is working exactly as designed and failing at everything that actually matters.
The market priced it all as good news and rallied. Which is the real schizophrenia: we're celebrating the consolidation of capital, the automation of judgment, the outsourcing of geopolitical leverage to tariff schedules, and the permanent assumption that AI infrastructure spending never faces saturation.
Something will crack first. I'd bet on the Fed's independence, but that's a different piece.
Upvoted! Thank you for supporting witness @jswit.