On Friday, June 5, 2026, the Nasdaq Composite fell 1,121.53 points, or 4.18%, to close at 25,709.43 — its worst single session since the tariff-driven turmoil of April 2025. The Nasdaq-100 fell harder still, down 4.77%. This was not a slow bleed. The index opened at 26,536.59, touched a session high of 26,572.25 in the first minutes, then capitulated continuously through the morning, the afternoon, and into the final hour, bottoming at 25,648.47 before a marginal recovery into the close. The day’s low marked a decline of nearly 4.5%. The composite spent the entire session grinding lower and never found its footing.
The number that matters is not the percentage but the catalyst behind it: this was a double detonation. A semiconductor earnings signal that questioned the trajectory of AI spending collided with a labor report so strong it resurrected the case for Federal Reserve rate hikes. Either alone would have stung. Together they produced the most violent technology selloff of the year, wiped roughly a trillion dollars from equity markets, and pushed the CBOE Volatility Index up 34% to close above 20 — a regime change in sentiment from the complacency that had governed the tape for weeks. CNN’s Fear and Greed gauge, in “greed” since the April record highs struck during the war with Iran, flipped to “fear” in a single session.
The Anatomy of the Selloff
The chip complex was the epicenter, and the damage was indiscriminate. Marvell and Micron each fell more than 10%. Intel dropped over 9%, AMD fell 10%, and Broadcom — the proximate trigger — slid roughly 5% on the heels of Thursday’s post-earnings decline. The iShares Semiconductor ETF lost more than 10% across Thursday and Friday, its worst two-day stretch since April 2025’s tariff shock. The catalyst, as several desks acknowledged, was not perfectly clean: Broadcom’s failure to raise its full-year AI chip outlook on Wednesday night was enough to turn the group lower on Thursday, but Friday’s intensity exceeded what the earnings alone justified. This is the signature of a crowded trade unwinding, not a considered repricing of fundamentals.
The macro accelerant arrived with the May payrolls report. Nonfarm payrolls rose 172,000 against an 85,000 consensus, and revisions added a net 93,000 jobs to the prior two months. Unemployment held at 4.3%. Layered on top of April CPI running at 3.8% year over year — the hottest since May 2023 — the data flipped the bond market’s posture entirely. Money markets now almost fully price the Fed’s next move as a hike rather than a cut. The two-year Treasury yield jumped 11.3 basis points; the ten-year pushed above 4.5% and the thirty-year breached 5%. Higher yields are mechanical poison for long-duration growth equities, and the Nasdaq is the longest-duration index there is.
The rotation beneath the headline was textbook. As capital fled technology, it did not leave the market — it moved. Consumer staples were the best-performing S&P sector, up more than 2%. Procter & Gamble and Clorox each rose about 5%. Colgate-Palmolive added 4%, Coca-Cola more than 3%, Johnson & Johnson 2%. The Dow, with its lighter technology weighting, fell only 1.35% — painful, but a fraction of the Nasdaq’s loss, and only one session removed from a record close. This divergence is the single most important tell in the entire episode, and I will return to it.
Context: How Bad Is 4.18%, Really
A 4.18% single-day decline ranks among the most violent sessions in the Nasdaq’s modern history, but context disciplines the panic. From its 52-week low of 19,334.98 — struck during the depths of early-2025 tariff anxiety — to its 52-week high of 27,190.21 set on June 1 at the peak of AI exuberance, the Nasdaq had more than doubled. Friday’s close sits just 5.4% below that all-time high. An index does not surrender a doubling in a single afternoon, and a 5.4% pullback from a record, however ugly the candle that produced it, is the kind of move that bull markets generate routinely on their way higher. In the arc of the cycle, this may yet prove to be noise. In the immediate term, it was the most painful day of 2026 for anyone long technology.
The Comparable Days
April 10, 2025 — the tariff shock (the direct analog). Friday’s loss was explicitly the worst since this episode. In early April 2025, President Trump’s sweeping tariff announcement — a proposed 10% minimum on all imports and a 34% increase on Chinese goods — sent the Nasdaq Composite to a 52-week low of 15,267.91 on April 8, down 13.4% year to date. The Nasdaq-100 logged its fourth-worst two-day loss since 2007. Then, on April 9, the administration announced a 90-day pause on reciprocal tariffs for all countries except China, and the index posted a roughly 12% single-day rally in the Nasdaq-100 — the third-largest one-day gain in its history. The recovery from there was relentless: the Nasdaq-100 surged 17.62% in the second quarter, and both the composite and the S&P 500 printed fresh all-time highs by the end of June. Time from bottom to new record: roughly three months. The lesson is brutal and specific — the selloff was policy-driven, and it reversed the instant the policy reversed.
March 10, 2025 — the correction open (-4%). A month before the tariff bottom, the Nasdaq-100 fell 4% in a session, its worst day since 2022, driven by the same tariff inflation fears compounded by recession signals; JPMorgan put recession odds at 40% at the time. Tesla fell 15.4%, Nvidia 5.1%. The index closed in correction, 10.4% below its December high. This was the leading edge of the same storm that bottomed in April — a reminder that 4% days often cluster rather than arrive alone.
July 17, 2024 — China chip sanctions (-2.8%). The Nasdaq logged its worst day since December 2022 on a Bloomberg report that Washington was weighing new semiconductor restrictions on China. Nvidia fell 6.6%, AMD 10.2%. Critically, the Dow rose 0.6% that same session and closed above 41,000 for the first time — the identical Dow-up, Nasdaq-down rotation signature visible on Friday. The index reclaimed new highs within weeks.
September 3, 2024 — manufacturing miss plus Nvidia (-3.3%). A soft ISM manufacturing print collided with a 9% drop in Nvidia on slowing-growth fears. It was the worst day since the early-August meltdown and a fitting introduction to September, historically equities’ cruelest month. The lost ground was recovered in roughly six weeks.
August 2, 2024 — the recession scare (correction territory). The mirror image of Friday’s problem. A weak payrolls number — 114,000 against a 179,000 forecast — pushed the Nasdaq more than 10% below its July high and raised fears the Fed had eased too late. Where Friday’s danger is an economy too hot, August 2024’s was an economy feared too cold. Both are macro shocks rather than demand breaks, and the resolution was the same: new highs within eight weeks.
The Pattern That Governs the Outcome
Across every analog, the catalyst — not the magnitude of the drop — dictated the recovery path. The distinction is the entire analysis. Selloffs born of valuation fatigue or external macro shock, against an intact underlying demand cycle, recovered in weeks to a few months: the 2024 episodes were each retraced inside two months, and the April 2025 tariff crash — far deeper than Friday’s, a genuine bear market in the composite — was fully recovered in roughly three once the policy trigger was removed. Selloffs born of an actual demand rupture — a real inventory glut, a real capex pause, the 2022 rate-driven bear — went deeper and lasted far longer, because the problem was earnings rather than sentiment.
Friday’s move is the first kind wearing the mask of the second. The proximate trigger was a single company declining to raise a forecast, amplified by a yield shock and a crowded-trade unwind. The underlying AI buildout — the capital expenditure, the data-center demand, the order books — has not been shown to be broken. It has been questioned, which is not the same thing. The bearish path requires Broadcom’s reticence to metastasize into actual Nvidia and AMD estimate cuts. Absent that confirmation, the structural case stands, and the history says this resolves upward.
The Position
A 4.18% session is a top-decile bad day, and the instinct to extrapolate it into catastrophe is precisely the instinct that the April 2025 tape punished most severely — nine of the ten largest Nasdaq-100 rallies in history have erupted during periods of maximum stress, and an investor who sold the April lows missed a 12% single-day rebound and a new record within a quarter. The economy is strong; strong economies do not produce the recessions that kill bull markets, even as hot data complicates the Fed’s path. The yield story is the real near-term overhang, and a break below Friday’s 25,648 intraday low would open the path toward 25,000 and the 100-day moving average. But the tape is broadcasting its own diagnosis in plain sight: the Dow closing within 1.35% of a record while the Nasdaq sheds 4.18%, capital stampeding into Colgate and Coca-Cola rather than into cash, is not the architecture of a market coming apart. It is the architecture of a market changing its mind about what to own. The next Nvidia print is the referendum. Hold the data-center growth rate and June 5 is a footnote by August. Crack it, and the analog stops being April 2025 and becomes 2022. Everything rides on which it is — and nothing in Friday’s selling answered the question.