Thursday 5 November 2026 was the worst session for US equities since the April 2025 tariff crash. The S&P 500 lost 4.4% to 7,354. The Nasdaq Composite fell 5.5% to 24,315, which put it more than 10% below the October record — the technical definition of a correction. The Nasdaq 100 dropped 6.3%, the Dow gave up 1,833 points, all eleven S&P sectors closed lower, and roughly USD 3 trillion of market value disappeared in a single day.

Friday brought a partial rebound: Nasdaq +1.09% to 24,579, S&P +0.31% to 7,376. But the week still closed with the Nasdaq down 3.4% — a fourth consecutive weekly decline — and the S&P down 2.7%, its worst week since 5 June.

The question every reader is typing into a search bar is simple: is this the AI bubble popping, or a buying opportunity? The honest answer is that nobody knows, and this article will not pretend otherwise. What it can do is separate what actually changed from what did not, name the levels, and show the arithmetic of trading it.

What Actually Broke — Earnings or Multiples?

The trigger was political: a Democratic midterm sweep and the read-through that a Democrat-controlled Congress slows data-centre buildout and tightens AI regulation. That is a policy story, not an earnings story.

Bank of America’s post-mortem makes the distinction explicit. The move was multiple compression, not estimate cuts — FY2028 earnings estimates were left unchanged. Investors did not conclude that AI companies will earn less; they concluded they will pay more for that stream of earnings. Those are very different claims, and they imply very different trades.

This matters for anyone holding a long NAS100 position. An earnings problem resolves when the numbers are printed. A multiple problem resolves when the discount rate or the narrative changes — and neither has a scheduled date.

The Three-Way Transmission

AI capex to index concentration

The Nasdaq 100 is not a diversified index in any practical sense. A handful of hyperscalers dominate its weight, so a single view on AI capital expenditure moves the whole benchmark. When that view wobbles, there is no offsetting sector to cushion it — which is exactly what Thursday demonstrated.

A hot payroll to long-duration tech

October non-farm payrolls came in at +165,000 against a +62,000 consensus, with unemployment at 4.1% and wages up 0.4% month on month. December Fed hike odds jumped to roughly 77%, and the dollar index pushed back above 102.

Higher-for-longer rates hit long-duration assets hardest, and AI equities are the longest-duration assets in the market: their cash flows sit far in the future, so they discount more heavily at every rate step. A strong jobs report is therefore bearish for Nasdaq longs, which is counter-intuitive to anyone who learned that “good news is good news.”

Rates to the dollar to everything else

The dollar firming above 102 tightens global conditions at the same time as equity risk premia widen. That combination tends to drag gold, emerging-market FX and cyclical commodities with it, even when their own fundamentals are intact. If you are trading more than one instrument this week, check whether you are actually holding the same trade three times.

The Levels That Matter

  • 24,315 — Thursday’s Nasdaq Composite close, and the level that defines the correction. A daily close back above it starts to repair the technical damage.
  • 24,000 — the round number and the obvious psychological shelf beneath the market. Losing it on a closing basis invites momentum sellers.
  • The October record high — the upside reference. Reclaiming it would invalidate the correction label, not the bubble debate.
  • VIX — the regime gauge. Elevated and rising VIX means gaps and wider spreads; falling VIX means the panic is being priced out.

On NAS100 specifically, treat these as zones rather than lines. Index cash opens gap, and a stop placed exactly at a round number is a donation to whoever is hunting it.

How to Trade It

Buying the dip versus shorting the bounce

Both are legitimate, and both fail in the same way: entering at the open, with no plan for a second leg. If you are long, your thesis needs a reason beyond “it always comes back.” If you are short, remember that short squeezes in a concentrated index are violent and fast.

Position size is the whole argument

Here is the arithmetic that decides whether you survive the week. On an ASIC-regulated retail account, major indices are capped at 20:1 leverage. An offshore entity will advertise up to 1:500. The same NAS100 position therefore requires 25 times more margin on the ASIC account.

That is not a marketing detail; it is the difference between a position that absorbs a 3% adverse move and one that is closed out by it. Work backwards from the margin you are willing to lose, not from the largest lot size the platform will accept.

UZFX lists 7 index CFDs including NAS100, US500 and US30, with a 0.01-lot minimum and a USD 10 minimum deposit on the Web Terminal, H5 and the iOS, Android, Windows and Mac apps. There is no MT4 or MT5. Negative balance protection applies. Leverage depends on the entity you contract with: 20:1 on major indices for ASIC retail clients, up to 1:500 offshore. Read the entity line before you fund, not after.

Stops belong outside the noise

With average true range expanded after a 5.5% session, a stop measured in points from your entry on a normal-volatility basis will be taken out by ordinary noise. Size the position so the stop can sit beyond the recent range, then accept the smaller lot that follows.

What Would Confirm or Kill the Thesis

Confirms the bubble thesis: hyperscalers cutting or delaying capex guidance; credit spreads widening for AI-adjacent issuers; a failed retest of 24,315 followed by a lower low; December hike delivered with hawkish guidance.

Invalidates it: FY2028 estimates holding while multiples stabilise; capex guidance reaffirmed on Q3 calls; Nasdaq reclaiming the October high; the Fed holding in December.

FAQ

Q: Is the AI bubble bursting right now?

Nobody can say with certainty. What is documented is a 10% correction driven by multiple compression rather than earnings cuts. A correction is a price event, not a forecast, and it can end at 24,000 or extend much further.

Q: Should I buy the Nasdaq dip?

Only with a defined invalidation level and a position size that survives another 5% move. The next scheduled catalysts are US CPI on 10 November and the FOMC on 8-9 December, so the gap risk is dated and knowable.

Q: How much margin does a NAS100 position need?

Divide the notional by your leverage. At 20:1, a position worth USD 10,000 locks USD 500 of margin; at 1:500 it locks USD 20. That is a 25x difference for the same exposure, and it is set by your account’s entity, not by the platform.

Q: Why did the Nasdaq fall when jobs data was strong?

Because strong data raises the odds of a December hike. Higher rates reduce the present value of long-duration cash flows, and AI equities are the most long-duration assets listed.

Q: Is a 0.01-lot Nasdaq trade enough to matter?

It is enough to learn with real money and small enough that a 3% adverse gap does not end your account. Most retail index losses come from lot size, not from direction.

Risk Warning

Trading CFDs on indices carries a high risk of rapid loss because of leverage. This article is market analysis and education, not investment advice, and no level named here is a recommendation to buy or sell. Corrections are not forecasts. Size every position as though the next 5% move goes against you.