AI Whipsaw
FUD, Rebuttals, And Volatility
In AI Apocalypse last week, we noted a suspsicious sequence: SemiAnalysis helped trigger a sharp pullback in AI-infrastructure names with bearish research, and then two new ETFs launched with SemiAnalysis as research partner—giving those ETFs a chance to buy some of the same names after the reset.
Zero Hedge version. https://t.co/WobA45nQmb
— Portfolio Armor (@PortfolioArmor) July 4, 2026
Monday gave us another version of that movie.
SemiAnalysis came out with another bearish AI-hardware note, this time about a potential delay in Nvidia’s Kyber architecture. Investors.com reported that SemiAnalysis said Kyber could be delayed by up to 12 months into 2028. An Nvidia spokesperson denied the report, telling IBD, “Our roadmap is intact.”
Then Korea added its own stress test.
Samsung estimated it would post a nearly 19-fold increase in Q2 operating profit. Korean stocks sold off anyway, with the KOSPI falling hard enough to trigger circuit breakers.
That tells us something useful. The issue isn’t demand. It’s expectations, leverage, and positioning.
Our System Still Likes The AI Stack
As of Monday night’s close, Portfolio Armor’s Top Ten leaned heavily into the AI trade.
That lines up with Nvidia’s response that its roadmap remains intact. It also lines up with the broader evidence we’ve been pointing to for months: the AI buildout is still real, and the bottlenecks remain in the physical stack.
The volatility is useful.
When bearish narratives hit the group, option premiums rise. We can harvest that premium to lower the cost of bullish structures, improve asymmetry, and in some cases get paid to take on defined-risk bullish exposure.
The Barbell
Our trades sit on a barbell.
On one end, we have speculative convexity trades. Those are lower-priced names where a big catalyst—drug approval, policy shift, hyperscaler customer, major order—can produce returns of 10x or more. Current fundamentals often lag the story there. That’s part of the setup.
On the other end, we have higher-probability theme-continuation trades. Those are names where the theme, technicals, and fundamentals already line up. The potential return if the stocks hit the high end of the option market’s expected movement is more like 1-3x, but the trade can still be modestly profitable if the stock drifts sideways, or even declines by 10% or 15%.
Three of today’s trades fit that second category. They give us AI-stack exposure after the pullback, with defined risk, favorable option pricing, and decent-to-strong technical and fundamental screens.
The fourth trade gives us a non-AI diversifier with a strong setup.
Full details on all four trades available here.
