Nasdaq 100 is 6.52% below its high, and a 10% correction…

A 10% correction in the Nasdaq 100 has not printed. Friday’s official close was 29,187.55, which is 6.52% under the 6 October closing high of 31,224.69, and the session low of 29,073.29 was still 6.89% under that high. A correction means 10% from the peak. Ten percent under 31,224.69 is 28,102.22, another 1,085.33 points, or 3.72%, from Friday’s close.

The missing fact is that this was not one tape. The Nasdaq-100 fell 5.49% from the 9 October close of 30,883.15. The Nasdaq Composite fell 4.48% over the same days, from 27,366.17 to 26,140.03, and is 5.29% under its own 6 October high of 27,599.89. On the Nasdaq-100 quote list, 47 of 101 symbols rose from 9 October to 16 October. The 10% damage is in chips: SMH fell 10.53% on the week and is 14.66% under its 6 October close. That split is the story the closing files support.

Key facts

• The Nasdaq-100 closed Friday 16 October 2026 at 29,187.55, down 2.50% on the day and 6.52% under the 6 October close of 31,224.69. Source: Nasdaq historical prices, file retrieved 18 October 2026.
• The 6 October intraday high was 31,361.37. Friday’s close is 6.93% under it, and Friday’s low of 29,073.29 was 7.30% under it. A 10% break from the closing high is 28,102.22. Same Nasdaq file.
• Thursday closed at 29,935.59, only 4.13% under the high. Thursday’s low, 29,408.41, was 5.82% under it, so a “more than 5%” print that day was intraday. Friday was the first close more than 5% under the high. Same file.
• Invesco QQQ closed Friday at $709.80, down 6.56% from its 6 October close of $759.66 and 5.52% from its 9 October close of $751.27. A 10% line on the ETF is $683.69. Source: Nasdaq QQQ history.
• SMH closed at $539.79, down 10.53% on the week from $603.33 and 14.66% under its 6 October close of $632.50. XLK closed at $183.47, down 7.70% on the week and 9.17% under $202.00. Nasdaq ETF files, same retrieval.
• Of 101 symbols on Nasdaq’s Nasdaq-100 quote list, 47 rose and 54 fell from 9 October to 16 October. Alphabet’s two classes both fell 3.15%. Counting Alphabet once leaves 47 gainers out of 100. FinanceFeeds count of the saved daily files.
• From the 2 June close of 30,660.60 to the 29 July close of 27,192.31, the Nasdaq-100 fell 11.31%. That correction was from the prior high. The 6 October close is a new high in the file, which runs back to 2 January 2024.
• The Kobeissi Letter posted on 16 October that the week was down 5.5% and the index was down 7% from its high. The week matches 5.49%. The 7% does not match the 6.52% close-to-close gap. Source: the post, 20:02 UTC.

What just happened, and why a 10% label is the wrong one

The highest close in the saved Nasdaq-100 history is Tuesday 6 October, at 31,224.69, with an intraday high of 31,361.37. A circulating close of 31,224.47 is 0.22 of a point under the table used here, so it is not the base for the percentage. Wednesday 14 October did close at 29,950.82, and that was 4.08% under 31,224.69.

The week’s 1,695.60 point decline, from 30,883.15 to 29,187.55, was two sessions. Wednesday took 618.66 points, from 30,569.48 to 29,950.82. Friday took 748.04 points, from 29,935.59 to 29,187.55. Together that is 1,366.70 points, or 80.6% of the week. Thursday fell 15.23 points, or 0.05%.

The Kobeissi Letter said the Nasdaq 100 “officially closes the week down -5.5%, posting its largest weekly decline since April 2025,” and that it was “down -7% from its high seen just 10 days ago.” The week is right if 5.49% is rounded to 5.5%. In this file it is also the largest Friday-to-Friday drop since the week ending 4 April 2025, when the index fell 9.77%, from 19,281.40 to 17,397.69. The 7% is not the close-to-close math. From 31,224.69 to 29,187.55 is 6.52%. From the intraday high to Friday’s close is 6.93%.

The Composite is a different series, and it is the one in FinanceFeeds’ Friday note on the Nasdaq Composite. That piece reported a close of 26,140.03, which matches this file. The week there, 27,366.17 to 26,140.03, is down 4.48%. The day was down 2.15%, from 26,713.03. The Composite is 5.29% under 27,599.89. A 10% break from that high would be 24,839.90. Folding the Composite, the Nasdaq-100 and a chip ETF into one “Nasdaq” percentage is how 6.52% gets told as almost 10%.

The 10% break is in chips, not in the index

QQQ did the index’s move in dollars. It closed 6 October at $759.66 and Friday at $709.80, down 6.56%. The week, from $751.27, was down 5.52%. Friday’s low was $707.23, or 6.90% under the 6 October close. Volume was 66,926,660 shares, against 26,513,570 on the high day. The 10% line is $683.69. The chart is that ETF. It is not a plot of 29,187.55.

SMH is through its own 10% line. Ten percent under the 6 October close of $632.50 is $569.25. Friday was $539.79, down 14.66% from that close and 10.53% from $603.33 on 9 October. The week itself was a 10% drop. XLK is close and not through: $202.00, then $198.78, then $183.47, down 7.70% on the week and 9.17% from 6 October. Its 10% line is $181.80, and Friday finished $1.67 above it. SOXX, the iShares Semiconductor ETF, closed at $493.39, down 11.80% on the week from $559.40 and 16.30% under $589.45. SOXX is not the Philadelphia Semiconductor Index.

FinanceFeeds’ Friday chip note reported the Philadelphia index at 11,137.29, down 3.83%, with all 30 constituents lower, and AMD at $528.75, down 5.95%. The member files here match the share prices in that account: AMD $528.75, Intel $89.85, down 5.79%, Micron $878.84, down 5.63%, and Arm $227.80, down 6.94%. The Philadelphia index level belongs to that report. The Nasdaq-100 close is a different series.

Twenty-nine of the 101 symbols were down more than 10% from their own 6 October closes, and 26 fell more than 10% on the week. Rocket Lab fell 25.47% to $50.84 and CoreWeave fell 25.21% to $61.39, including an 11.91% Friday drop. Nvidia, one constituent and not this article’s price call, closed at $214.13, down 6.61% on the week from $229.28 and 10.50% from its 6 October close of $239.24. That stock is through 10%. The index is not.

Who is exposed

The large lines did not move together. From 6 October to Friday, Apple rose from $333.63 to $343.46, up 2.95%, and was up 2.03% on the week. Microsoft went from $529.30 to $529.05, down 0.05% from the high day. Walmart rose from $107.20 to $118.81, up 10.83% from 6 October. Costco rose from $935.68 to $989.11, up 5.71%. Amazon fell from $256.29 to $250.00, down 2.45%. Alphabet Class A fell from $347.68 to $340.59, down 2.04%. Meta fell from $738.88 to $709.325, down 4.00%. Tesla fell from $380.68 to $354.72, down 6.82%.

Where a 10% break has printed, it is in the stocks, not the index. From 6 October, Broadcom fell 12.52% to $328.75, Micron 15.95% to $878.84, AMD 18.58% to $528.75, and Sandisk 21.75% to $1,299.30. An index holder is long Apple’s gain and Micron’s loss. An SMH holder is not.

The quote list’s market-cap column is not Nasdaq’s published weight. Dropping Alphabet’s Class C line, the eight largest names are about 63% of that sum, with Nvidia about 12.4% and Apple 12.0%. A 22-name set of chip, equipment and memory lines in the same file, including Nvidia, Broadcom, AMD, Micron, ASML and Applied Materials, sums to about 29.7%. That is why SMH can fall 14.66% while the index falls 6.52%.

QQQ is the cash vehicle. NQ futures reference the index, not SMH. A hedge struck at 28,102.22 was not paid, because that level did not trade. Friday’s QQQ low of $707.23 was still $23.54 above $683.69. From the 31 December 2025 close of 25,249.85, the index is still up 15.59%.

What the tape shows

Week means the 9 October close to the 16 October close. “Versus 6 Oct” uses Friday against that day’s close, not the intraday high. The 10% level is 90% of the 6 October close. QQQ, XLK, SMH and SOXX are ETFs. The first two rows are indexes. All figures are from the Nasdaq files saved for this piece.

The median weekly change across the 101 symbols was down 1.13%. The average was down 3.41%, pulled by the tail. An index down 5.49% while the median name is down 1.13% is a weight story. Forty-seven symbols finished the week higher.

The same ETF already had a real correction from an older high. QQQ closed 2 June at $746.16 and 29 July at $661.73, down 11.32%, which sits under the dashed line on the chart. It then made a new high at $759.66 on 6 October. The line is 10% under that new high. Price has not touched it.

What was said, and what was not

Joe Mazzola, head trading and derivatives strategist at Charles Schwab, wrote in the 16 October note: “AI and chip-related stocks struggled this week as investors reassessed whether heavily indebted data center, software, and semiconductor companies can turn today’s enormous capital spending into tomorrow’s profits.” The page is dated 16 October and was modified at 13:38 UTC, 9:38 a.m. New York, so it is a morning note, not the close. The same note said Deutsche Bank kept a buy on CoreWeave and that there was no company-specific reason to justify the sell-off. It also reported, citing Bloomberg, that Nvidia chief executive Jensen Huang said CoreWeave’s financial position was improving and did not require a backstop. Schwab had the stock down 5.6% early. The saved close was $61.39, down 11.91% on the day and 25.21% on the week from $82.08.

The last time the index itself stood near 10%, the quotes were from July and have to stay there. On 28 July, AdvisorHub’s Bloomberg account said the Nasdaq 100 was approaching a correction. Mark Luschini, chief investment strategist at Janney Montgomery Scott, said: “The AI trade is being looked at with a much greater degree of skepticism, and the shift in sentiment means it has become something of a one-way trade with things getting sold unmercifully.” Jim Awad, senior managing director at Clearstead Advisors, said: “People are worried about capital spending, the return on that spend, the level of free cash flow, and there’s some dis-investment going on ahead of the Big Tech reports we’re seeing this week.” That week was late July. The article put the index down 9.5% from the 2 June record. The file agrees: 27,763.13 on 28 July is 9.45% under 30,660.60. The next close, 27,192.31, finished an 11.31% correction.

Nasdaq, the exchange, and Invesco, the QQQ sponsor, did not have to comment for the level to be knowable. The historical table is the document. No Friday statement from either firm on this drawdown turned up in the pages fetched here, and none is invented. The public lines are Mazzola’s morning sentence, the July quotes, and a post that rounded 6.52% toward 7%.

Rates, July, and the week ahead

Schwab’s morning note is the rate context, with that clock attached. It said odds of a hike later this month were back to 17% on the CME FedWatch tool, from around 45% on Wednesday after the hot CPI, while a hike before year-end was still above 90%. The note gave no CPI index level, so none is added. It described a 10-year swing from near 5.33% to a low of 5.05%, and the page displays 5.10%. Those are morning figures, not a closing auction.

The note puts the first official third-quarter GDP estimate on 29 October, a day after the Fed meeting, and said Atlanta Fed GDPNow was still at 3.6% while the average analyst estimate was well below 3%. A year-end hike, if those morning odds hold, is a discount-rate argument about AI spending. It does not rewrite Friday’s close. The index was 6.52% under its high at the bell.

July is the precedent that did reach 10%, and then ended. The first close back above the 2 June print of 30,660.60 was 22 September, at 30,732.40. The record followed on 6 October. Luschini’s line described a market that broke 10% the next session. Reusing it as a description of Friday skips the new high. The spending doubt is still visible in SMH and in Mazzola’s sentence. It has not, on this peak, produced a 10% index correction.

FinanceFeeds’ Friday tape had Tesla at $354.72, down 2.13%, and Netflix at $70.53, down 2.86%, the same closes as these files. Netflix was up 0.33% on the week, from $70.30 to $70.53, and up 2.68% from $68.69 on 6 October. It is not a name dragging the index through 10%. The Composite piece places Netflix, Tesla and Meta reports in the week of 19 October. Those reports can move large non-chip lines. They are not a substitute for a close at 28,102.22.

What happens next

By Friday 23 October the correction either prints or it does not. The level is 28,102.22, another 3.72% under 29,187.55, and $683.69 on QQQ. It gets there if the names already down more than 10% from 6 October, including Nvidia, Broadcom, Micron and AMD, keep falling and Apple, Walmart and Costco do not offset them. It misses if last week’s shape holds: 47 gainers, a median member down 1.13%, and the loss concentrated in two sessions. A 10% headline before a close at or under 28,102.22 is a label, not a print.

On 28 October the Fed meeting in that Schwab calendar sits a day before the GDP estimate. If the year-end hike odds the morning note put above 90% are still the price, the issue is the cost of capital on AI spending. That can extend the chip drawdown without pulling every member through 10%. SMH is already 14.66% under its 6 October close. Another 10.53% week in that ETF would still not, by itself, put the Nasdaq-100 at 28,102.22. The chip set above is about 30% of a market-cap sum, not the index.

Frequently asked questions

Is the damage still a chip story?

Yes. The Nasdaq 100 closed Friday at 29,187.55, 6.52% under the 6 October high of 31,224.69, so a 10% correction has not printed. SMH is 14.66% under its own 6 October close, and 29 of 101 Nasdaq-100 list symbols are down more than 10% from theirs. Apple, Walmart and Costco rose over that stretch. The index loss is the weight of the chip complex, not a uniform break across the list.

What close would count as a 10% correction?

A correction here means 10% under the peak close. Ten percent under 31,224.69 is 28,102.22. Friday’s close would have to fall another 1,085.33 points, or 3.72%, to reach it. Friday’s low was 29,073.29, still 6.89% under the closing high and 7.30% under the intraday high of 31,361.37. On QQQ the same test is a close at or under $683.69. Friday’s ETF low was $707.23.

Did the index fall 7% from the high?

Not on the close. The Kobeissi Letter said the week was down 5.5% and the index was down 7% from its high 10 days earlier. The week matches: 30,883.15 to 29,187.55 is 5.49%. The close-to-close drawdown from 31,224.69 is 6.52%. Friday’s close is 6.93% under the intraday high of 31,361.37, which is the gap a 7% line is rounding. That is still short of 10%.

How can 47 members rise while the index falls?

Nasdaq’s quote list has 101 symbols. From 9 October to 16 October, 47 rose and 54 fell. Alphabet Class A and Class C both fell 3.15%, so counting Alphabet once still leaves 47 gainers out of 100. The median weekly move was down 1.13%. The index fell 5.49% because large lines such as Nvidia, Broadcom and Micron fell much more than the median stock. Weight, not breadth, made the week.

Was the Nasdaq 100 already in a correction in July?

Yes, from the prior high. The 2 June close was 30,660.60 and the 29 July close was 27,192.31, down 11.31%. On 28 July the index was still 9.45% under that June close, which matches the “just shy of 10%” account published that day. The next session finished the break. The first close back above the June high was 22 September, and a new high followed on 6 October. This 6.52% is a new drawdown.

Is the chart the Nasdaq-100?

No. The chart is the Invesco QQQ Trust. Its 6 October close was $759.66 and its Friday close was $709.80, down 6.56%, with a 10% line at $683.69. The index points, 31,224.69 and 29,187.55, stay in the text because an ETF chart cannot carry them. QQQ’s week, from $751.27 to $709.80, was down 5.52%. The ETF tracked the index. It is not the index.

This is not financial advice.