Palantir reported second-quarter revenue of $1.94 billion against a $1.81 billion consensus, up 93 percent year over year, and every headline led with that growth rate. It is the wrong number to anchor on. The figure that determines whether this quarter is an inflection or a peak is the $6.24 billion in remaining U.S. commercial deal value, more than double the year-ago level and roughly eight times what the segment actually recognized in the three months just ended. Revenue growth tells you what already closed. Remaining deal value tells you what has been signed and not yet billed, and at that coverage ratio the 149 percent U.S. commercial growth rate is not a comp artifact waiting to unwind. It is a conversion schedule.

Five immersion deep-ultraviolet lithography machines are scheduled for delivery in China this year, with roughly twenty more planned across 2027. That is the operating scale behind a synchronized 7% intraday drawdown in ASML, Applied Materials, Lam Research and KLA — ASML’s steepest single-session fall since June, touching 8% at the lows before paring toward 5.75% into the final hour. The dislocation between those two numbers is the entire story, and understanding why the market was right to sell anyway is more useful than the arithmetic that says it overreacted.

ChangXin Memory Technologies opened at 49.50 yuan against an IPO price of 8.66 and closed the morning session up 531%, carrying a market capitalization near 3.66 trillion yuan and making it the most valuable company listed on the mainland. The headline number is 500%. The number that explains it is 6.73%.

That is the share of post-IPO capital that was unrestricted and tradable on day one — roughly 4.5 billion shares out of a base that runs to nearly 67 billion. Institutional demand for the allocation exceeded 500 times the shares offered; retail subscription ran better than 200 to one. A first trade that added some 2.7 trillion yuan of implied value did so against a sliver of float, before a single wafer of incremental capacity existed or a single customer qualification changed. Price discovery on 6.73% of a company is not price discovery. It is an auction for scarcity, and it resolves when the scarcity does.

The same share of the same company traded at two prices this week, roughly a third apart, and there was nothing anyone could do about it. SK Hynix’s American depositary receipts have run as much as 51% above the Seoul-listed common stock since the July 10 Nasdaq debut, settling around a 33% premium midweek. That gap is not a signal about DRAM pricing, HBM allocation, or Nvidia’s qualification schedule. It is a plumbing failure, and the plumbing is the point.

The number that carried SpaceX into the largest initial public offering in history was not a launch cadence figure. It was roughly $75 billion of contracted artificial intelligence revenue, assembled in the six weeks before the June 12 listing and presented to investors as the evidence that the company had become something more durable than a rocket manufacturer. Anthropic agreed in late May to pay $1.25 billion a month through May 2029 for exclusive access to the Colossus 1 facility outside Memphis. Google followed on June 5 with $920 million a month from October 2026 through June 2029 for approximately 110,000 Nvidia GPUs, plus the CPUs, memory and supporting infrastructure around them. Two customers, three-year terms, a combined backlog larger than the annual revenue of most S&P 500 constituents. The roadshow wrote itself.

Alphabet reported $119.8 billion in Q2 2026 revenue, up 24% year-over-year and above the roughly $117 billion consensus, marking a twelfth consecutive quarter of double-digit growth. Google Cloud revenue surged 82% to $24.8 billion, beating estimates near $22.5 billion, with Cloud operating income tripling to $8.8 billion from $2.8 billion a year ago. GAAP EPS came in at $9.11, dwarfing the roughly $2.89 consensus. None of that headline is what moved the stock. The GAAP beat was driven almost entirely by a $99 billion unrealized gain on Alphabet’s equity stakes in Anthropic and SpaceX; strip that out and adjusted EPS was $2.85, a miss against the $2.89 estimate. Shares fell as much as 5% in after-hours trading, not on the earnings number, but on management raising full-year 2026 capex guidance to $195-205 billion from $180-190 billion and flagging a further significant increase into 2027. The quarter beat on almost every reported line and the market sold it anyway.

ServiceNow beat across every headline metric: total revenue of $3.987 billion, up 24% year-over-year against a roughly $3.93-3.97 billion consensus, subscription revenue of $3.877 billion up 24.5%, and adjusted EPS of $0.90 against an $0.86 estimate. Management raised full-year subscription guidance to $15.76-15.78 billion. None of that is the real story. Shares had cratered into the print, falling 6.6% on the day of earnings alone as part of a broader SaaS-displacement scare — Pegasystems and IBM had both flagged customers delaying software orders, and OpenAI’s new enterprise product was being framed as a direct threat to ServiceNow’s core IT service management business. The stock closed at $95.46, down roughly 37% year-to-date and nearly 51% off its 52-week high of $210.20. What the print actually delivered was a direct answer to the disruption question: Generative AI annual contract value crossed $1 billion and remains on track for $1.5 billion by year-end, agentic AI deployments are up 9x over nine months, and Level 1 IT service management automation is now resolving 80-85% of requests without human intervention. Shares rallied 5-7% after hours. The headline is the beat; the number that matters is that ServiceNow’s own AI products are cannibalizing its legacy per-seat business faster than any external competitor is managing to, which is the only argument capable of resetting a stock priced for disruption risk.

Tesla posted $28.24 billion in Q2 2026 revenue, up 26% year-over-year and comfortably above the $26.4 billion Wall Street consensus. Deliveries hit a record 480,126 units, up 25% year-over-year and roughly 74,000 above analyst estimates, marking the company’s first year-over-year delivery growth in two years and pushing trailing-twelve-month revenue above $100 billion for the first time. None of that translated to the bottom line. Non-GAAP EPS came in at $0.33, missing the $0.51-$0.53 consensus by a wide margin and down 18% year-over-year. GAAP operating income fell 57% to $398 million. The headline is volume recovery. The real story is that Tesla converted a record quarter into its weakest operating margin in years, and the market priced the second number, not the first.

The Philadelphia Semiconductor Index closed Friday more than 20 percent below its late-June record, the technical threshold for a bear market. Two trading sessions later, the index is up a combined figure north of 7 percent, with Tuesday’s 3.7 percent gain marking the second consecutive advance. The question facing the sector is whether this is the start of a genuine recovery or a bounce inside a larger correction.

What Broke the Rally

The selloff had three distinct triggers layered on top of each other. First, valuation exhaustion: chip names had posted triple-digit year-to-date returns before the reversal, leaving the trade crowded and vulnerable to any negative catalyst. Second, a competitive scare — a Chinese open-source model release narrowed the perceived gap with frontier US labs, raising questions about the compute intensity assumptions baked into AI infrastructure spending forecasts. Third, TSMC’s capital expenditure guidance came in well above prior estimates, which the market read as a margin-pressure signal rather than a demand-confirmation signal. Once the reversal began, systematic de-risking and algorithmic flow amplified the decline across both US and Asian listings, with SK Hynix and Samsung posting double-digit single-session drops in Seoul.

Two stocks now decide what happens to a national index. Samsung Electronics and SK Hynix together make up more than half of the Kospi’s weight, roughly double their share from a year ago, and that math has quietly turned South Korea’s benchmark into something it was never designed to be: a live gauge of how the world feels about AI spending.

The mechanism is straightforward. Both companies dominate the high-bandwidth memory supply chain feeding Nvidia and the rest of the AI buildout, so their share prices move on the same inputs driving AI capex sentiment everywhere else — data center orders, hyperscaler guidance, chip demand forecasts. When that sentiment shifts, it used to show up first in Nasdaq trading and filter down to Seoul overnight. That order has reversed. Overnight moves in Samsung and SK Hynix are now setting up the next day’s trading in Tokyo, Taipei, and New York often before U.S. markets have even opened. Japanese traders have started calling this effect “Kospi-nirami,” borrowing a term once reserved for watching the Federal Reserve or the yen.