Market-structure
Jordi Visser: the May-June AI selloff was a speed crash, not a bust, and market structure is why
The strategist argues that emotionless AI agents and passive benchmarks systematically underweighting high-growth names are producing sharper, faster corrections inside a still-intact structural bull market, rather than the start of one.
Macro strategist Jordi Visser's read on the May-June 2026 selloff, which erased over $2.2 trillion from the largest technology names, is that it was a "speed crash" rather than the beginning of a genuine bust: leveraged speculative positioning got flushed out fast and violently, but the underlying demand drivers for AI capacity were left intact. His central thesis is that AI token demand is still in an early stage, and that every new model capability threshold, coding agents, consumer assistants, whatever comes next, opens a larger consumption surface than the physical infrastructure supplying compute can keep pace with, which keeps structural demand ahead of supply even through sharp corrections.
The market-structure argument underneath this is the more interesting part for anyone building or operating trading systems. Visser points to a market increasingly dominated by three forces: AI trading agents that do not experience fear or fatigue the way human traders do, passive index benchmarks that mechanically underweight high-growth AI names because of how they are constructed, and a physical buildout being funded directly by the small number of corporations with the balance sheets to do it. His claim is that this combination produces a distinct pattern, parabolic rallies followed by sharp, transient corrections, rather than the multi-year bear markets that characterised prior tech cycles, because none of the three structural forces actually reverses course during a drawdown.
His evidence for reading the correction as bullish rather than the start of a bust is specific: Apple raised handset prices by 20 percent citing surging memory costs, and its stock rallied to all-time highs rather than falling, which Visser reads as the market pricing a future where consumer AI agents are coming and premium hardware to run them is a foregone conclusion rather than a risk. He also points to daily RSI divergences, lower price lows paired with higher RSI lows, appearing for the first time since July 2025, occurring specifically on bad news, which is a textbook technical signal of exhausted selling pressure rather than accelerating decline.
The caveat worth attaching to any "emotionless AI agents produce a new kind of market structure" thesis is that it has not been tested across a full cycle that includes a genuine, sustained demand shock rather than a leverage-driven flush. Agents that do not panic also do not obviously know when a structural break has actually occurred versus a temporary air pocket, and a market structure argument built on three forces that have only operated together for roughly two years is a thesis, not an established pattern. Readers should treat this as one credible, well-argued view on why AI-era corrections may look different from prior cycles, not as confirmation that they will.
Read the original: BigGo Finance - Jordi Visser: the May-June AI selloff was a speed crash, not a bust, and market structure is why. Commentary is the independent editorial view of Share Trading; the original article is credited to its publisher.