The week tech sold off on good results

The index funds hid it. The Nasdaq 100 fund (QQQ) and the technology fund (XLK) each ended the week to Friday, October 30 down 2.0%. Underneath, 23 large technology and internet companies lost more than a tenth of their value in five trading days.

There are 178 technology and communication companies on the screener worth $10 billion or more. The median one fell 1.0%, while the median large company of any kind rose 1.1%. 107 of the 178 fell. 23 fell by more than 10%, and 11 of those reported results during the week. The other 12 fell without a report of their own.

In this chart: Reddit · ServiceNow · F5 · Fortinet · Intel · Twilio · Roblox · Snap · Atlassian · HP · Marvell · Guidewire · Cloudflare · HubSpot · Pinterest · Fair Isaac · Datadog · Alphabet · AppLovin · Figma · United Microelectronics · Teradyne · Strategy

The week, day by day

For three days the funds rose. By Wednesday's close the technology fund was up 4.2% on the week. The 23 were flat on Monday and Tuesday and turned down on Wednesday. Then the big reports landed, on Tuesday to Thursday evenings. On Thursday the median of the 23 went from 3.3% down on the week to 9.5% down, and on Friday the technology fund lost 4.9%, taking the whole week's gain and more, after Amazon and Intel reported on Thursday evening.

A strong quarter, then the outlook

The seven biggest fallers all reported that week, and all of them grew revenue, from 13% to 53%. Measured against what analysts expected, as the articles about each report quoted it, five beat on revenue and two, ServiceNow and Intel, matched. What sank them was the next quarter: either an outlook below estimates, or one that met estimates but showed growth or margins slowing.

An outlook below estimates:

• Reddit beat on revenue ($895 million against $877 million expected), but guided October to December at $1.00 to $1.03 billion, against $1.04 to $1.05 billion. Daily users came in at 130.3 million against 133.5 million expected, and in the US they fell about 2% from a year earlier, to 50.7 million, as fewer people arrive from Google search. The company said it will move to a mostly logged-in site. Down 29% on Friday.
• Intel matched on revenue ($16.51 billion, against estimates the articles put between $16.42 and $16.54 billion) and beat on earnings per share ($0.52 against $0.40), then guided the fourth quarter at $15 to $16 billion, where analysts had about $17 billion. Bloomberg's reporter, after interviewing chief executive Lip-Bu Tan, put the weakness down to PCs and the price and availability of memory, which Tan expects to last into the first half of 2027. Down 29% on Friday, and 45% for October, its worst month in at least twenty years.
• Roblox beat on revenue ($1.57 billion against $1.46 billion) and on bookings, though bookings, the measure it guides on, were 12% lower than a year earlier. It reported 138 million daily users, short of the 145.8 million in the estimates Bloomberg cited, and guided holiday-quarter bookings at $1.89 to $2.00 billion, below the $2.01 billion consensus. Down 18% on Friday.

An outlook that met estimates, with something slowing inside it:

• ServiceNow matched on revenue ($4.11 billion) and raised its full-year subscription outlook, but adjusted earnings per share came in at $1.00 against $1.02, and it said contracted business for the next twelve months would grow a little over 18%, down from 20.5%. Bloomberg's headline said the outlook "revives AI fears", the worry that AI tools eat into its kind of software. Down 27% on Thursday.
• F5 beat on revenue and earnings and guided the next quarter in line. But its first outlook for the new fiscal year was 8% to 9% growth, after 13%, and it guided gross margin to 81.5% to 82.5%, from 84%, as it pays more for components. It also expects hardware replacement orders to decline in the second half. Articles read that as the end of a refresh cycle that had lifted the shares about 80% this year. Down 20% on Wednesday.
• Fortinet beat ($0.91 a share against $0.85) and raised its full-year revenue outlook, but trimmed its services revenue forecast to $5.15 to $5.17 billion, from $5.18 to $5.22 billion. Down 20% on Thursday.
• Twilio beat on revenue and earnings and guided the next quarter 1% above estimates. The guide still meant organic growth slowing to 11% to 12%, from 16%, and on the call management said this quarter's smaller beat came from a slowdown in messaging, its largest product. Down 15% on Friday.

Alphabet fell 12.6% on Thursday after its own report, when Google's services revenue came in at $96.3 billion against $98.5 billion expected. We wrote about that day in Alphabet's worst day in twenty years, in context.

The ones that fell without a report

Twelve of the 23 had no results that week. Some had a story of their own:

• Marvell fell 13% on Thursday, on reports that it is in advanced talks to buy Lumentum, a deal articles expected to be paid largely in new shares, on the same morning Alphabet's results weighed on AI names. It fell another 10% on Friday with the rest of the sector.
• Guidewire fell 8% on Tuesday after agreeing to buy hyperexponential, a London underwriting-software company, for $1.7 billion in cash.
• Fair Isaac fell 13% on Monday after holding its mortgage score price flat at $10 to keep lenders using it as a rival score gains ground, with the housing regulator's director promising "more action" on credit score competition.

For several others, no company news appears in their feeds at all. Datadog fell 11.5% on Friday and Cloudflare 9.5%. The articles that day tie the slide to Amazon's and Intel's results on Thursday evening, after Alphabet's and ServiceNow's the day before, and to worry about how much the largest companies are spending on AI and how much cash it burns. Atlassian started the week at a record and ended it 16% lower, a week before its own report; nothing in its feed explains it. HP fell 9.6% on Friday. Articles that week described a memory shortage squeezing PC makers, and Intel blamed memory for its weak outlook, but none tie HP's fall to either.

What the coverage said

Tickr Lab reads every article in each company's Yahoo Finance feed and gives it a tone from −1 to +1. Across the 23, the average went from +0.13 in the week before (139 articles) to −0.14 in the week itself (538 articles). Coverage of Reddit, ServiceNow, Intel and Roblox turned clearly negative. Coverage of F5, Fortinet and Twilio stayed close to neutral, about 0.0 to +0.06, despite falls of 15% to 20%: several of those articles argued the drop had gone too far. Articles written after a fall describe the fall, so tone follows price here rather than explaining it.

How far from the 52-week high

For the twelve biggest fallers, the chart shows where each stood against its highest close of the past year, before the week and after it.

The week hit companies in very different places. Atlassian was at its high on October 23. F5, HP and Fortinet were within 8%. All four are now between 16% and 30% below. ServiceNow went from 16% to 44% below.

Others were already far down. Roblox was 61% below its high from a year ago and is now 62% below. Reddit went from 44% to 65% below and closed Friday at a new 52-week low. Intel went from 38% to 53% below the high it set in June.

Cheaper than before. Cheap for them?

A fall of a fifth or a third makes a company cheaper than it was a week ago. The more useful comparison is with the company's own past. The chart measures market value against the past year's revenue, now and at every quarter end of the last five years. Revenue works even for companies with losses or one-off profits, which is half of this list.

• Below their own median, some at the bottom of the range. ServiceNow trades at 7.0 times revenue, half its five-year median of 14.6 and level with its lowest quarter end in that time. On earnings it is at 75 times, against a five-year median near 130. Reddit is at 6.0 times, against 13.6 since it listed in 2024 and below its lowest quarter end, 9.2. Roblox is at 5.2, against a median of 10.1 and a low of 6.8. Snap is at 1.4, against 3.4. Atlassian is at 7.6, against 11.2. HP and Guidewire sit just under their medians.
• Still above their own median, even after the fall. Intel is at 5.6 times revenue, more than twice its five-year median of 2.5. It made a loss over the past year, so it has no P/E. Marvell is at 19.1, against 10.3. Twilio is at 5.6, against 3.5. F5 is at 5.7, against 4.6, and Fortinet at 12.8, against 10.5. On earnings, F5 (27 times, median 26) and Fortinet (44 times, median 42) are close to their own history.

In this chart: Snap · Reddit · ServiceNow · Roblox · Atlassian · HP · Guidewire · Fortinet · F5 · Twilio · Marvell · Intel

The split mostly follows where each stock was before the week. Intel, Marvell and F5 had risen hard this year, so a fall of a fifth still leaves them above their usual multiple. ServiceNow, Reddit and Roblox had been falling for months, and this week took them below anything in their recent record.

What the Signal says

Tickr Lab's Signal reads every stock after every close with one fixed set of rules and names the one rule that fired. It describes the price trend; it does not forecast. After Friday's close, of the twelve, 2 read Very Weak, 7 Weak, 2 Neutral and 1 Strong – Elevated Risk.

• Breakdown (Very Weak): a price more than 10% below its 200-day average and still falling. Reddit is 42% below its average and Intel 24%.
• Downtrend (Weak): below both the 200-day and the 50-day average, but not yet 10% below the 200-day. ServiceNow is 9% below and F5 7%.
• Distribution (Weak): heavier-than-usual volume on a falling month. Fortinet and Marvell are both still above their 200-day averages, but more than 30% below their highs, which fires the deeper version. HP is still 17% above its 200-day average.
• Below 200D (Weak): for companies the rules treat as early-stage, any close below the 200-day average. Roblox is 18% below and Snap 2%.
• Neutral: Twilio is still 9% above its 200-day average, so no weakness rule fires. Atlassian reports within a week, and the rules hold its reading at Neutral until then.
• Strong – Elevated Risk: Guidewire did most of its falling on Monday and Tuesday, after announcing an acquisition. Its price is back inside the band the rules call a clean setup, but its fundamentals do not pass the check, so the reading stops at Strong – Elevated Risk.

The rules look only at how far and how fast a price moved, not why. That is why a 14% fall can still leave one stock reading Strong – Elevated Risk while a 16% fall leaves another Weak.

None of this says where these shares go next. It is how far they fell, against what, and how they read now, measured the same way Tickr Lab measures every company.

Share prices are daily closes from October 23 to October 30, 2026. The 23 are the technology and communication companies worth $10 billion or more at the October 23 close that fell more than 10%. The 52-week high is the highest close in 252 trading days. Price to sales and P/E use market value at the October 30 close over the trailing four quarters' revenue and net income, taken from each company's results releases or SEC filings, whichever is newer, against its quarter ends from September 2021 to June 2026. Snap's latest report is from June, so its figure does not include the quarter it reports on November 4. The Signal uses the Balanced rules after the October 30 close. Analysts' estimates are as the articles about each report quoted them (StockStory, Zacks, Bloomberg, Investing.com); where sources differed, the post names the one used. Reasons for each fall come from the companies' releases and the full text of the articles in their Yahoo Finance feeds.