Buy Crypto Markets Spot FuturesBTC Earn Event Center Rewards Hub

The last session was Monday, September 14. All three indices closed lower, with the S&P 500 down 0.48% at 7,620.25, the Nasdaq Composite down 0.56% and the Dow Jones Industrial down 0.29%. Rates did mThe last session was Monday, September 14. All three indices closed lower, with the S&P 500 down 0.48% at 7,620.25, the Nasdaq Composite down 0.56% and the Dow Jones Industrial down 0.29%. Rates did m

U.S. Pre-Market Briefing on Sept. 15: Semis Slide 4% as Tenable (TENB) Surges 16.51%; Trip.com Reports Tonight — Why Was 17% Growth Guided Down to 3%?

#US Stocks

2026/09/15 16:40

The last session was Monday, September 14. All three indices closed lower, with the S&P 500 down 0.48% at 7,620.25, the Nasdaq Composite down 0.56% and the Dow Jones Industrial down 0.29%. Rates did most of the work: the 10-year Treasury yield reached about 5.0%, its highest since October 2023. The gap between sectors was far wider than the index move. A public dispute over the pace of AI development set off a selloff in chip names, the semiconductor ETF fell more than 4% in a single day, and money moved into cyber security. Today's stock in focus is Tenable Holdings (TENB), up 16.51% at $35.08 on 3.05x its average volume. Tonight brings the New York Fed manufacturing index for September at 12:30 UTC, and Trip.com Group (TCOM) reports after the 20:00 UTC close. The real gate of the week comes tomorrow with the FOMC decision. Today's Academy is the Tuesday metric class on 52-week range position, the same ruler used in the last session, read from a different angle. Data in this article is based on the September 14, 2026 US close, and all times are UTC.

Today's Market: All Three Indices Fell and Rates Did Most of the Work

All three indices closed lower on September 14. The Nasdaq Composite fell 0.56%, the weakest of the three; the S&P 500 fell 0.48% to close at 7,620.25; the Dow Jones Industrial fell 0.29%, the smallest decline.

Rates were the main reason. The 10-year Treasury yield reached about 5.0%, its highest since October 2023. Expectations of a hike ahead of the meeting lifted the discount rate and pressed valuations down a notch. Every index moved around half a percent, which reads as an ordinary down day.

The real story that day was not in the indices. It was in the distance between sectors.

A public dispute over the pace of AI development set off a selloff in chip names, and the semiconductor ETF fell more than 4% in a single day, over eight times the index decline. On the same day money moved clearly into cyber security, and that whole line rose double digits.

Both moves happened in one session, under the same technology label, in opposite directions. The index averaged them out, which is why it only fell 0.48%.

Oil moved as well. Brent reached about $106 a barrel after a supply disruption in the Middle East. That variable lands tonight on Trip.com Group's transportation ticketing line, covered further below.

Stock in Focus, Tenable (TENB): The Real Counter-Move in a Down Market

Tenable Holdings (TENB) sells cyber security within infrastructure software and carries a market value of $3.87B. On September 14 it closed at $35.08, up 16.51%, adding about $550M of market value on the day. Volume reached 11.79 million shares, 3.05x its recent average and the heaviest in the group. Beta is 0.96, which means that when the market moves 1% it moves 0.96%. This is not a high-volatility name to begin with.

Why it rose: two AI lab chief executives published safety warnings over the weekend, and the market read them as a case for larger enterprise security budgets. Tenable sells vulnerability management, which means finding a company's own gaps first. It does not sell a firewall. It sells the list of where the holes are.

This day can be called a counter-move, and it should be. In the previous edition, writing about HP Enterprise (HPE), we made a point of the opposite: that was a sector rally rather than a counter-move, because the whole computer hardware group averaged +8.29% and HPE simply ran at the front. Today is different. All three indices fell, the neighbouring semiconductor group fell more than 4%, and this one line rose double digits on its own. The market and the adjacent sector went down together while this group went up. That is what a counter-move looks like.

Two qualifiers come first.

First, those were public statements, not orders. How much enterprise security budgets actually rise will only show up in the order numbers next quarter. What was priced that day was an expectation rather than realised performance.

Second, Tenable was not the top name of the day. Zscaler (ZS) closed at +16.52%, finishing 0.01 points ahead of Tenable's 16.51%, and the two share the front rank. Writing Tenable up as the day's largest gain would be wrong.

So why is Tenable the stock in focus? Because the turnover behind the two differed a great deal: Tenable ran 3.05x its average volume against 1.90x for Zscaler. The next section takes that apart.

The five-dimension score fills in the character of the day. Relative Strength is a full 100, with Trend Position at 69, Peer Ranking at 80 and Volatility Control at 82, so the other four corners all sit at 69 or above and the shape is unusually even. Valuation Temperature at 40 is the lowest of the five and the only one below a passing mark.

A score of 40 there says something direct: what the market bought that day was an expectation, not an order already on the books. Expectation has moved out in front, and orders have to arrive behind it to hold this position.

Peer Comparison: One Infrastructure Software Group, Cyber Up Double Digits and AI Infrastructure Down

Put six infrastructure software names from the same session side by side and the split is immediate: Zscaler (ZS) rose 16.52%, Tenable (TENB) 16.51%, CrowdStrike (CRWD) 13.85%, Palo Alto (PANW) 13.09% and Fortinet (FTNT) 9.04%, while Oracle (ORCL) fell 3.70%. That leaves 20.22 points between the ends.

The five that rose all sell cyber security, and the one that fell sells AI infrastructure. One classification was cut cleanly in two, and the cut sits in a different place from the computer hardware split of the previous session, where systems rose and drives fell. Money is changing lanes this week.

One more figure belongs next to those: the Nasdaq infrastructure software group averaged only +2.06% that day. Five of six names rose double digits while the group average came to 2%, which says the classification did not rise. One line inside it did. Tenable finished 14.45 points ahead of the group average.

A supporting name, Coinbase (COIN), rose 9.24% on an unrelated line: Senate Republicans released a revised draft of the crypto market structure bill, which took one notch off regulatory uncertainty for exchanges. It has nothing to do with security budgets and should not be read alongside them.

One-Minute Concept: Before Reading a Gain, Read the Volume Behind It

The volume multiple is the day's volume divided by the average daily volume over a recent period. A gain tells you how far the price moved. The multiple tells you how many people took part.

The two get read as one thing, yet they measure different quantities. Three marks are enough.

First, around 1x. Volume sits close to the daily average and the price mostly follows the broad market. There is no need to read the day as an event of its own. It is closer to an ordinary reading of the whole market than to anything the company did.

Second, above 2x. The day was not priced only by the people who already held the name. New capital moved the price to another level. 2x is a meaningful line: past it, the mix of participants has changed rather than the same holders simply trading more often.

Third, equal gains still differ. The clearest example sits right here: Tenable and Zscaler both rose about 16%, yet Tenable ran 3.05x its average volume and Zscaler only 1.90x. Same price move, different depth of participation.

That is why the stock in focus is Tenable rather than the name 0.01 points ahead of it. A gain is the outcome. The multiple is the width of the process. A 16% day on 1.9x volume can be existing holders trading among themselves. The same 16% on 3.05x says a fair amount of new money arrived in that session.

⚠️ The multiple also has limits: it does not separate buyers from sellers, and it does not say whose money arrived. A high multiple means many people took part. It does not mean they were right. Treat it as a filter for whether a day is worth reading on its own, not as a judgement of direction.

What to Watch Tonight: Trip.com Reports, and Why 17% Growth Was Guided Down to 3%

The New York Fed manufacturing index for September lands at 12:30 UTC tonight. Trip.com Group (TCOM) reports after the 20:00 UTC close, with the call at 00:00 UTC on Sept 16.

The real gate of the week comes tomorrow: the FOMC decision at 18:00 UTC, published together with the rate dot plot and the economic projections, followed by the chair's press conference at 18:30 UTC. The current policy range is 3.50% to 3.75%, and the market prices a 25 basis point hike at this meeting at about 85%. Today is the first day of the meeting, and no statement is released.

(The previous edition said the weight of the week sat on the Wednesday FOMC. Today is the second square on that board, and hike pricing has moved from above 80% then to about 85% now.)

First, what Trip.com does. It is the largest online travel platform in China and runs four brands: Ctrip, Qunar, Trip.com and Skyscanner. It owns no hotels and no flights, and earns booking commissions, which makes its revenue close to a live reading of Chinese travel demand. Tonight covers the second quarter, April to June.

The market has one question.

Net revenue grew 17% year over year last quarter, after 21% in 25Q4 and 16% in 25Q3. Three steady quarters. Yet the company guided this quarter to 3% to 8%, a midpoint of 5.5%, which cuts the growth rate by two thirds, or 67.6%.

⚠️ One basis has to be stated before the rest makes sense: the first three bars on the chart are reported results, and the fourth is the midpoint of the company's own guidance range. Guidance is not a result and it is not the market consensus either. It is a range management gave itself, and mixing the two inverts the conclusion.

On the last call, management attributed the slowdown to energy costs and geopolitics. Brent reached about $106 a barrel in the previous session, which means that thread lands directly on the transportation ticketing line tonight.

Why read growth rates rather than the revenue figure itself? Trip.com's revenue is strongly seasonal: the third quarter carries the summer travel peak and the first quarter carries Lunar New Year, so comparing absolute amounts points the wrong way. A year-over-year rate cancels the season and shows the real slope of demand.

Drill-Down: The Two Fastest Lines Together Are About 11% of Revenue

Lay last quarter's four business lines out by growth rate and a structural problem appears, with every figure on a renminbi basis: corporate travel management at RMB 0.69B and +20%, packaged tours at RMB 1.1B and +19%, accommodation booking at RMB 6.5B and +17%, and transportation ticketing at RMB 6.1B and +12%.

The two fastest lines come to about RMB 1.79B together, roughly 11% of total revenue. Corporate travel grew fastest and holds 4%; packaged tours grew 19% and hold 7%. Neither can lift the total, however fast it runs.

The total is set by the other two: accommodation booking at RMB 6.5B plus transportation ticketing at RMB 6.1B come to about 78% of revenue. Within that pair, transportation ticketing grew only 12%, the slowest of the four.

Put those together and tonight's arithmetic is plain: for the total to fall from 17% to somewhere between 3% and 8%, both large lines have to slow at once. The small lines cannot offset it.

The two large lines also come under pressure in different ways. Air ticketing earns a booking commission, so a higher fare is not a higher commission. A rising oil price may first hold down ticket volume, and fewer tickets means less commission. The hotel line tracks domestic consumption itself far more closely and relates to oil only at a distance.

Two specific things to watch tonight.

First, where total revenue growth lands inside the 3% to 8% range. Near 8% points mainly to a high base a year earlier. Near 3% points to demand itself slowing. One range, two very different meanings.

Second, the gap between the accommodation and transportation lines. If ticketing falls further than accommodation, that matches the oil thread. If both fall together, the question sits closer to overall consumption.

⚠️ There is also an antitrust fine of RMB 5.18B, which fell after the quarter ended and is not on this quarter's books. How it is recognised has to be read separately, so do not look for it in tonight's numbers.

On the stock itself, Trip.com closed at $39.08 on September 14, up 0.15%, with no earnings and no after-hours session that day. The price still sits against its 52-week low of $38.04 in a range of $38.04 to $78.99, a position of about 2.5%. What that 2.5% actually means is the subject of today's Academy.

Academy: Metric Class, What 52-Week Range Position Actually Measures

The previous lesson, the weekend catch-up, ran 52-week range position across six oil companies and asked whether the barrel was revenue or cost for each of them. That lesson showed how to use the ruler to compare different roles.

Today is the second lesson with the same ruler. The sample changes to five AI-chain companies and the question changes too: what is this ruler actually measuring?

⚠️ The two samples are entirely different and must never be quoted against each other. The positions of those six oil names have no bearing on these five.

How it is calculated. The formula is (latest close − 52-week low) ÷ (52-week high − 52-week low) × 100. Take Oracle (ORCL): the price is $144.72 against a one-year low of $114.50 and a high of $329.50, which places it at 14% of the range. Note that every name uses its own one-year endpoints. Two companies at 50% are not at the same kind of price. The number is relative to the company itself, never to its peers.

What it measures. Apple (AAPL) sits at 89.5%, and 89.5% does not mean expensive. It means the market barely changed its view of this business over the year, so the price stayed near the top of the range. Position measures how much the market revised its mind, not how good the company is.

That sentence is worth a second thought. The two ends of the range are two emotional extremes: on the day of the high and on the day of the low, the market held very different views of one business. Position tells you where today's price sits between those two extremes. It is a record of revision, not a valuation and not a gain.

When it stops working. Marathon Digital (MARA) is the counter-example: its one-year high and low track bitcoin almost step for step. In that case the position measures the coin price rather than the company's own situation. The ruler is fine. The sample is not.

One more test: when the position changes, ask first whether the business changed. Oracle spent the year moving from database licences to renting AI compute by usage, which replaced the whole rhythm of how it collects money. The first half of that range and the second half are not the same business. Running one range across two businesses still produces a number, but the meaning is discounted.

One AI Chain, Six Times Apart From Top to Bottom

Put five AI-chain companies side by side and the order speaks for itself: Apple (AAPL) at 89.5%, Palantir (PLTR) at 66.2%, Nvidia (NVDA) at 64.6%, Vertiv (VRT) at 42.3% and Oracle (ORCL) at 14.1%. Six times separates the ends.

What each one sells:

Apple (AAPL) | devices, with replacement cycles and service subscriptions, the most predictable cash timing of the five
Palantir (PLTR) | data analysis software for governments and large enterprises, collected in instalments under contract
Nvidia (NVDA) | accelerator chips for cloud operators, one chip for one unit of compute
Vertiv (VRT) | power and cooling inside the data centre rack, since machines need electricity and cold air first
Oracle (ORCL) | moved from database licences to renting AI compute by usage

The pair in the middle is the most instructive. Nvidia at 64.6% and Palantir at 66.2% sit at almost the same position. Yet one sells chips to cloud operators and the other sells software to governments, and the customers, the contract structures and the payment timing have nothing in common. Same position, entirely different situation. That is exactly the point: position is not a quality score. It places each company's own revision on one 0 to 100 axis.

At the other end, Oracle at 14.1% carries the largest revision of the five. The next section takes it on its own.

Company Profile: Oracle (ORCL)

Oracle deserves a section of its own because it has appeared in three consecutive editions in three different roles, and those three together form one continuous story.

In the September 11 edition, it was the subject of the September 10 after-hours results, with cloud infrastructure revenue growth running into triple digits. In the September 14 edition, it was the trigger for a broad AI hardware rally, with HP Enterprise and Dell both moving on the back of its report. In today's edition, it is the only name of six to fall in the peer comparison, down 3.70%, and the lowest position of five in the Academy at 14.1%.

One company went from trigger to most heavily revised inside two weeks.

What it sells. It used to sell database licences on long multi-year contracts, which made revenue highly predictable. It now also rents AI compute by usage, so customers pay for what they consume. Those are two businesses with completely different payment rhythms.

When the money arrives. Signed but undelivered orders reached a record, and that part is not in dispute. But the data centres have to be built first. An order is a commitment rather than cash. Renting out compute requires facilities, power, cooling and chips, all of which are money that goes out first.

The result is this figure: free cash flow for fiscal 2026 turned negative at about −$23.7B.

The orders did not shrink. The cash left first. That is what a position of 14.1% is measuring. The market has not rejected the business. It has revised when the business starts paying back. The orders remain, but the outflow moved earlier and the return moved later, and discounted back those are two different numbers.

Which returns to the test from the previous section: when the position changes, ask first whether the business changed. Oracle's position fell to 14.1% not because the orders weakened, but because it changed businesses, and the market is repricing the new one.

A test you can carry to any name. The one-minute concept closes like this: when you see a position figure, spend thirty seconds on three questions. What does it sell and to whom, when does the money arrive, and can it adapt when demand shifts? Answer those and the position means something. Leave them unanswered and it is only a percentage.

Frequently Asked Questions

Q: Tenable (TENB) rose 16.51% on September 14. Can that be written as the largest gain of the day?

A: No. Zscaler (ZS) closed at +16.52% in the same group, 0.01 points ahead of Tenable, and the two share the front rank. Tenable is the stock in focus for a different reason: it ran 3.05x its average volume against 1.90x for Zscaler, the heaviest in the group.

Q: Why does this day count as a counter-move when HP Enterprise (HPE) in the previous edition did not?

A: The two structures differ. When HPE rose 12.42% on September 11, the computer hardware group averaged +8.29%, so the sector rose together and HPE ran at the front. That is a sector rally. On September 14 all three indices fell, the neighbouring semiconductor ETF fell more than 4%, and the whole cyber line rose double digits. The market and the adjacent sector went down while this group went up, and that is a counter-move.

Q: The group average was only +2.06%, yet five of six names rose double digits. Is that a contradiction?

A: No, and it is the key point. Infrastructure software holds many companies, and cyber security is one line inside it. A group average of 2.06% says the classification did not rise. One line inside it did. Tenable's 14.45 point lead over the group average measures exactly that gap.

Q: Can the two chief executives' safety warnings be read as orders for Tenable?

A: No. Those were public statements, not orders. How much enterprise security budgets actually rise will only appear in next quarter's order numbers. Valuation Temperature scoring 40, the only dimension of five below a passing mark, measures that gap: what was priced that day was an expectation, not realised performance.

Q: Is Trip.com's 3% to 8% figure a market forecast or a reported result?

A: Neither. It is the company's own guidance range, with a midpoint of 5.5%. The last reported quarter grew 17%, after 21% and 16% before that. Guidance is not a result and not the market consensus, so the fourth bar on the chart has to be labelled as a guidance midpoint and cannot be read alongside the first three.

Q: Are Trip.com's revenue figures in dollars?

A: No. They are all on a renminbi basis. Net revenue of RMB 16.2B last quarter, accommodation booking RMB 6.5B, transportation ticketing RMB 6.1B, packaged tours RMB 1.1B, corporate travel management RMB 0.69B and the antitrust fine of RMB 5.18B are all renminbi. Keep the RMB label and do not convert them to dollars.

Q: Does a high 52-week range position mean a stock is expensive?

A: No. It answers one question: where today's price sits within the high and low of the past year. It is neither a gain nor a valuation multiple. Apple at 89.5% means the market barely changed its view over the year; Oracle at 14.1% means the market made a very large revision. Position measures how much the market changed its mind, not how good the company is.

Q: The previous Academy also used 52-week range position. Can the two sets of numbers be compared?

A: No. The earlier sample was six oil companies, framed around whether the barrel was revenue or cost. Today's sample is five AI-chain companies, framed around what the ruler measures. Same metric, entirely different samples, and the two sets must not be quoted against each other.

Disclaimer: This article is compiled and written by the MEXC RealStocks team. The data in this article is based on the closing of the US stock market on September 14, 2026. The content is a compilation of public market information, and individual stocks are publicly discussed targets, which do not represent the recommendation or opinion of MEXC and do not constitute any investment advice. More US stock content: @MEXC | @Alpha_MEXC | @MEXC_Research

Market Opportunity

4 Price(4)

----

USD

4 (4) Live Price Chart

The articles shared on this page are sourced from public platforms and are provided for reference only. They do not represent the position or views of MEXC. All rights belong to MEXC. If you believe any content infringes upon the rights of a third party, please contact [email protected] for prompt removal. MEXC does not guarantee the accuracy, completeness, or timeliness of any content and is not responsible for any actions taken based on the information provided. The content does not constitute financial, legal, or other professional advice, nor should it be interpreted as a recommendation or endorsement by MEXC. For expert insights and in-depth analysis, visit MEXC Learn.