Week in Review: Asia Selloff and Iran Strikes

The last time conditions looked roughly like this was August 2015, when a surprise yuan devaluation combined with commodity stress to produce a messy cross-asset selloff that hit Asia hardest. The parallel is loose, I want to be upfront about that, but the shape of the week rhymes: a geopolitical shock widening into energy markets, a sharp and sudden repricing of Asian tech, and the rest of the world catching a cold from both. In 2015, the S&P 500 round-tripped and energy bottomed first. Whether that pattern repeats here depends on what happens in the Strait of Hormuz over the coming days. But let me walk through what the week actually showed.

What the Week Revealed

The dominant tension was the collision between a hot war in the Middle East and a technology selloff in Asia that, by Friday, had become severe enough to demand its own headlines.

On the geopolitical side, U.S. strikes on Iran stretched into a seventh consecutive day, hitting infrastructure including bridges and water plants. Iran retaliated by attacking Gulf infrastructure, and Hormuz tensions continued to rise. Energy stocks responded predictably. The S&P 500 Information Technology sector was flat on Friday, but energy names were one of the few pockets of green in an otherwise red U.S. session. As I wrote in the Oil Weekly Gain and Asia Tech Selloff Explained post on July 17, the August 2015 episode is useful context here, not as a template but as a reminder that geopolitical supply shocks and tech selloffs can coexist and feed off each other.

The Asia tech selloff was the week's sharpest pain, driven by a convergence of forces. Taiwan's TWII dropped 6.47% and South Korea's KOSPI fell 6.37% on Friday alone. The mechanism matters here: Taiwan's semiconductor fabs depend heavily on energy imports that transit the Strait of Hormuz, so the escalation in the Gulf is a direct input-cost and supply-continuity risk for TSMC and its supply chain. Korea's memory sector faced a separate structural headwind. CXMT, China's largest domestic memory chipmaker, moved forward with its IPO this week, signaling a new wave of subsidized competition that could structurally pressure margins for Samsung and other incumbent memory producers. And both markets likely experienced forced selling from leveraged positions as margin calls cascaded through a session that moved too fast for orderly deleveraging. The Shanghai Composite lost 3.05%, the Hang Seng declined 1.78%, and Japan's Nikkei fell 4.03%. These are not small numbers.

The ripple effects reached the U.S. clearly. The Nasdaq fell 1.4%, the S&P 500 dropped 1.01%, and the VIX, which measures expected volatility, jumped 12.19% to close at 18.77. The Dow fell 0.77%. Emerging market ETFs bore the brunt: VWO (Vanguard Emerging Markets ETF) dropped 1.7%, while developed international markets fared somewhat better, with VEA down just 0.47%.

Meanwhile, the broader U.S.-China relationship added another layer of anxiety. President Trump claimed China obtained 220 million U.S. voter files ahead of the 2020 election, a headline that, while not directly market-moving on its own, reinforced the adversarial tone between Washington and Beijing at a moment when Chinese tech ETFs were already under pressure. MCHI fell 2.2% and KWEB dropped 2.44% on Friday, declines that reflected both the Asian tech contagion and the persistent risk premium the market assigns to U.S.-China tensions.

The trade environment got noisier too. President Trump threatened new tariffs on Canada over wildfire smoke, and Washington continued to push the EU to roll back import rules. Neither of these stories moved markets as directly as the Iran headlines, but they added friction to an already nervous week.

What the Agent Got Right, and Wrong

Let me start with the misses, because they matter more.

The agent closed its MU (Micron) research subject this week at a 12.88% loss, triggered by the stop-loss. The original thesis was built on extreme valuation metrics in the memory semiconductor space, but this is now a well-documented failure mode in the agent's research history. As the learnings explicitly state, re-entering the same semiconductor thesis at progressively higher prices has consistently destroyed value. The first entry captures the dislocation; later entries are chasing. MU was the agent's latest lesson in that pattern. The CXMT IPO is worth noting here as well: a major Chinese memory competitor entering public markets introduces a structural supply-side risk that pure valuation models tend to miss. If CXMT's subsidized capacity pressures pricing across the DRAM and NAND markets, extreme cheapness in memory names may not be a dislocation at all but a rational repricing of long-term competitive dynamics.

Separately, PEP (PepsiCo) was closed at a 4.20% loss, tripped by the confidence gate, which exits positions when confidence is low and drawdown exceeds a threshold. PEP had a confidence score of 0.56, well below the 0.65 floor the agent's own learnings now insist on. Entering it in the first place was a process error.

On the positive side, several active research subjects showed resilience or outright strength during the week. ADBE (Adobe) is now showing a 16.29% positive observed delta from its entry, and its thesis, built on extreme valuation dislocation in a profitable software company, remains intact per the agent's review system. CRM (Salesforce) is up 7.83% from entry on a similar thesis: enterprise software with strong free cash flow trading at a deep discount. GILD (Gilead Sciences) is up 8.50%, though the agent's review flagged minor concerns as the price approached its base case level. The thesis review noted that GILD reached within a dollar of the level the agent had identified, and the question now is whether the move has largely played out.

BAC (Bank of America) quietly moved to a 2.68% positive delta, and its thesis, built on quality financials benefiting from a positively sloped yield curve, is intact. The 10-year Treasury yield closed at 4.541% on Friday with a slight decline, which is broadly supportive of the banking thesis.

The Asia Problem and the Research Subjects Caught in It

The sharpest impact of the Asia selloff fell on 005930.KS (Samsung Electronics), which the agent opened as a research subject based on what appeared to be an extreme valuation dislocation. Samsung is barely above its entry price, with a delta of just 0.20%. The thesis is intact per the review system, but the broader Korean market's 6.37% single-session decline is the kind of environment where even cheap stocks get cheaper. The CXMT IPO adds a structural dimension to the risk: Samsung's memory division now faces a well-capitalized Chinese competitor entering public markets with access to state-backed capital. The agent's learnings about semiconductor re-entries are relevant context here, though Samsung was a first entry, not a re-entry. The key question going into next week is whether Asian tech stabilizes or whether this selling has further to run.

EWT (Taiwan ETF) fell 2.83%, and EWJ (Japan ETF) declined 1.54%. While neither is an active research subject, the contagion affected IWM (Russell 2000 ETF), which is one of the subjects the agent is studying. IWM fell 0.52% on Friday (its underlying index, the Russell 2000, fell 0.42%, with the ETF slightly underperforming) but still holds a 3.13% positive delta from entry. The review system flagged minor concerns, noting that large-cap tech momentum could reverse the small-cap rotation thesis. Friday's action, where small caps fell less than large caps, actually supports the relative strength thesis, but the confidence score of 20% is the lowest in the active research set. This is worth watching closely.

Energy, Defense, and the Geopolitical Thesis

TTE.PA (TotalEnergies) barely moved on the week at a negative 0.24% delta from entry, which is surprising given the escalation in the Gulf. The thesis was built on Hormuz escalation providing a tailwind for European energy majors. That tailwind is clearly present in crude oil prices and European natural gas, but TTE hasn't fully captured it yet. The thesis remains intact per the review system.

RTX (RTX Corporation, the defense name) sits at a negative 2.88% delta, and the review system flagged minor concerns about potential government spending cuts. The defense backlog thesis is structurally sound, with NATO spending expansion and the Iran conflict providing a multi-year demand story, but the price action hasn't reflected that yet.

EWG (Germany ETF) is at a negative 2.72% delta with minor concerns flagged. The EWG ETF itself fell 0.19% on Friday, modestly outperforming the DAX index which dropped 0.34%. The trade headline about Washington pushing the EU on import rules rollback is directly relevant here. If U.S.-EU trade tensions escalate further, German exporters, especially automakers, would feel the pressure. The thesis needs the European recovery to continue despite these headwinds.

The Remaining Threads

META (Meta Platforms) declined to a negative 3.47% delta from entry. The thesis, built on attractive growth-adjusted valuation and strong margins, is intact per the review system. A 1% down day in the broad market is noise for a name like this, not signal.

LLY (Eli Lilly) holds a 4.07% positive delta with minor concerns flagged. The review system noted a sharp pullback from its peak as a potential sign of profit-taking. LLY is one of the few healthcare names the agent's learnings endorse, given its genuine hypergrowth profile in GLP-1 drugs.

What This Week Taught

The week illustrated something the agent's research history has documented repeatedly but still struggles to internalize: the difference between a thesis being right and a trade working. Samsung's valuation metrics are genuinely extreme. MU's were too, before the stop-loss hit. The CXMT IPO is a reminder that valuation dislocations in semiconductors can persist or deepen when the competitive landscape shifts structurally. The agent's calibration data shows a Brier score of 0.292, which is close to uninformative. The hit rate across closed research sets, at 41 observations, is not yet strong enough to override the simple lesson that cheap can get cheaper, especially in volatile sectors during geopolitical stress.

As a reminder, everything in this post is observational research, not personalized advice. Readers should consult an authorized financial advisor before making any decisions.

What Stays With Us

The number I'm carrying into next week is 6.47%. That's how much Taiwan's stock market fell in a single session on Friday. Korea was close behind at 6.37%. Those are not normal moves. The causes are identifiable: Hormuz risk threatening energy supply chains critical to semiconductor manufacturing, the CXMT IPO signaling a new era of Chinese memory competition, possible forced liquidation of leveraged positions, and the overhang of deteriorating U.S.-China relations. Whether they represent a genuine repricing of Asian tech risk or an overshoot driven by the confluence of these fears, the answer will shape the next few weeks for multiple research subjects the agent is tracking. I'll be watching whether Monday brings stabilization or acceleration.

Research output, not investment advice. The material above is observational and educational. The operator of Observed Markets may hold personal positions in subjects the agent studies (disclosed at observedmarkets.com/conflicts-of-interest). Always consult an authorized financial advisor before any investment decision. Past observed outcomes do not predict future results.