Nasdaq 100 Forecast This Week: Down 7% in July, and This Week Is the Rematch

TL;DR

• The Nasdaq 100 closed 28,274.20 on Friday, up 167.89 points or 0.60%, and the Nasdaq Composite rose 1% to 25,373.85.
• July was the worst month since March 2025. The index fell roughly 7%, against a 7.69% drop in that earlier month, and it was the only major index to lose ground.
• It sits 8.1% below its June 3 record of 30,762.20, having been 10% below as recently as July 29. The bounce started before the earnings did.
• This week is the rematch. The names that broke it in July report in the next four sessions: AMD, Palantir, Datadog and AppLovin.
• The index does not need good news to rally from here. It needs the absence of bad news, which is a much lower bar and a much more fragile one.

Where Is the Nasdaq 100 Headed This Week?

The short answer: this is the most asymmetric week the index has had all summer, because it enters oversold with the exact catalysts that oversold it.

July did real damage. The Nasdaq 100 lost about 7%, its steepest monthly fall since March 2025, while the Dow rose and the S&P 500 finished essentially flat. The causes were not mysterious: a semiconductor selloff that erased a trillion dollars, a hawkish Fed, volatile oil, and a growing suspicion that AI capital spending would not produce returns fast enough to justify the multiples paid for it.

Then, in the last two sessions of the month, it stopped falling. On July 29 the index was 10% below its record. By Friday's close it was 8.1% below. That is a bounce of roughly two percent in two sessions with no fundamental news attached to it, which is what positioning unwinding looks like rather than conviction returning.

Conviction, or its absence, arrives this week.

The Board

Worst month since March 2025, an 8% hole, and the four names that dug it all report inside four sessions.

The Levels That Matter

The record: 30,762.20, set June 3. The index needs 8.8% from Friday's close to reclaim it. That is not a one-week move in normal conditions, and it frames the realistic bull case as a recovery rather than a breakout.

The correction line: roughly 27,686. Ten percent below the record is the level the index traded at on July 29, and it is the one that matters psychologically. Losing it again turns a bounce into a lower low, which is the first genuine trend break this index has produced in 2026.

The gap between those two numbers is the whole week. At 28,274.20 the index sits closer to the correction line than to the record, with roughly 2% of cushion below and nearly 9% of work above.

The rule for the week: in an oversold index, the first level to watch is the one underneath. Bounces that fail at obvious resistance are common. Bounces that break to new lows are the ones that change the regime.

The Four Names That Decide It

This is not a macro week for the Nasdaq 100. It is a company week, and four reports carry it.

AMD reports Tuesday after the close, and it is the single most important print for this index all month. Consensus is $1.62 on about $11.3 billion, up 47%, and it sits inside management's own guided range, so the beat is close to pre-agreed. What matters is the data centre line. Every megacap that has reported is a buyer of AI compute describing what it intends to spend. AMD is a seller reporting what was actually ordered. If the promised spending has not converted into orders, that shows up here first.

Palantir opens the week Monday evening with consensus sitting above the company's own guidance, which is an unusual position for the sell side to take and a nervous one to hold.

Datadog reports Thursday before the open and is the highest-variance name of the week: options price 13%, the last print moved the stock more than 30% against an 11.5% implied move, and full-year guidance implies the company earns in the second half roughly what it earned in the first while revenue grows near 30%.

AppLovin closes Wednesday guiding 52% growth at an 84% EBITDA margin, with a stock down about a third this year on a third-party data point rather than a company one.

Scenario 1: The Squeeze

Clean numbers from AMD, supported by Datadog and Palantir, into an index that is 8% below its record with positioning already flushed. Oversold indices do not rally politely. This is the scenario that produces a 4% week and leaves everybody who rotated into defensives in July looking slow.

Scenario 2: The Confirmation

AMD's data centre line disappoints, or the guide is soft. The index breaks the correction line, prints its first meaningful lower low of 2026, and the July selloff stops being a rotation and becomes a de-rating. This is the scenario with the longest tail, because a broken AI capex assumption does not repair itself in a quarter.

Scenario 3: Right on Earnings, Wrong on Friday

The prints are fine, the index rallies for four days, and then payrolls on Friday comes in hot, long yields push higher, and the highest-multiple index in the world gives it all back before the weekend. The Fed already held with three members voting to hike, and the 30-year is above 5.19%. Long-duration assets do not get to ignore the long bond.

The Options Angle

• Do not sell volatility into this index this week. Realised moves beat implied repeatedly through July, and this week's calendar contains a name that has already produced a 30% single-session reaction inside the last quarter.
• The index is the wrong instrument for the view. If the thesis is "AI capex is real", the expression is the reporting companies, not the index that averages them against Costco and PepsiCo. The index dilutes exactly the signal you are trying to trade.
• Tuesday evening is a two-event window. AMD and SpaceX both print, with correlated tape risk and no ability to manage between them.
• If you want defined-risk upside on the squeeze, a call spread expiring after Friday survives being right on Tuesday and wrong on payrolls, which an outright call does not.

The One-Line Read

The Nasdaq 100 closed 28,274.20 after losing about 7% in July, its worst month since March 2025, and it enters this week 8.1% below the June record having already bounced two percent off the correction line on nothing but positioning: the four names that broke it all report inside four sessions, which makes this the rematch rather than the recovery, and the index will be told whether July was a rotation or a diagnosis by Thursday morning.