Oil Went Up, Fear Did Not.

This week's five

1. 01

   DELL, view score card · Dell Technologies · Technology / Hardware and AI Infrastructure

   Friday Five Score

   88 / 100

   Move before publication

   +9.3%

   Actionability

   Extended

   Next date

   Sep 10

   Dell published second quarter results after the close on September 1 and raised its full year revenue forecast by 25 billion dollars, to 192 billion, with non-GAAP earnings guidance going to 25.50 dollars from 17.90. It booked a record 60.9 billion dollars of AI server orders and left the quarter with a 95 billion dollar backlog. The stock rose 15.81 percent the next session in a week when the technology sector fell 1.40 percent, so none of this is sector drift. At 20.3 times the company own raised forecast the price is not demanding, and the thing to watch is not the order book but the margin it converts at.

   The catalyst. On September 1, 2026 Dell Technologies reported second quarter fiscal 2027 revenue of 47.0 billion dollars, up 58 percent year over year, and non-GAAP diluted earnings of 7.04 dollars a share, up 203 percent. It booked a record 60.9 billion dollars of AI server orders, recognised 16.4 billion dollars of AI server revenue and exited the quarter with a record 95 billion dollar AI backlog. It raised full year fiscal 2027 revenue guidance to 192.0 billion dollars from 167.0 billion and non-GAAP diluted earnings guidance to 25.50 dollars from 17.90. The stock rose 15.81 percent on September 2 and a further 4.91 percent on September 3. U.S. Securities and Exchange Commission, September 1, 2026

   Why it made the five. It is the largest guidance raise anywhere in the screen this week, and the raise is the whole reason it is here rather than the price move. Dell lifted full year revenue guidance to 192.0 billion dollars from 167.0 billion and non-GAAP earnings to 25.50 dollars from 17.90, which is a 42 percent increase to the earnings forecast in a single quarter. The quarter behind it was 47.0 billion dollars of revenue, up 58 percent, with 60.9 billion dollars of AI server orders taken and a 95 billion dollar backlog left over. My own record says a Guidance Raise catalyst has been the best performing type in this newsletter by a wide margin, and this is the cleanest example of one I have had. The tape agrees for once: the stock rose 15.81 percent on September 2 in a week when the technology sector fell 1.40 percent and software fell 3.05 percent, and it did it on 2.47 times normal volume.

   What to watch next. Sep 10: Oracle reports fiscal first quarter results after the close, the nearest dated read on whether AI infrastructure demand is still accelerating Confirms the thesis: Oracle pointing to more AI infrastructure commitment rather than less, which would say the order book Dell is quoting is an industry condition and not one company sales quarter Breaks it: Oracle signalling that AI infrastructure spending is flattening, which would make a 95 billion dollar backlog a share gain inside a market that has stopped growing

   The case against it. The stock closed at 97.3 percent of its 52 week high after more than quadrupling from that 52 week low, so nobody is early. The specific worry is mix. AI servers are lower margin than the rest of what Dell sells, so a guide that raises revenue by 25 billion dollars while the mix keeps shifting toward those servers can deliver every dollar of that revenue and leave profit roughly where it was. A backlog is also an order book, not a contract that cannot move: 95 billion dollars of it depends on a small number of very large buyers whose capital plans can be revised. And five Form 144 notices of proposed sale were filed by insiders inside the measurement week, after the move. Those are routine under pre arranged plans and prove nothing on their own, but I would rather you heard it here.

   Weekly move (before publication): +9.3%

   Every DELL appearance →Track this pick on the Scoreboard →Size a position in it →

2. 02

   CF, view score card · CF Industries Holdings · Materials / Nitrogen Fertilizer

   Friday Five Score

   84 / 100

   Move before publication

   +9.6%

   Actionability

   Extended

   This is the war trade that is not an oil stock. By CF own account the Middle East usually supplies 25 to 30 percent of globally traded ammonia and 35 to 40 percent of globally traded urea, and the company estimates the conflict has already taken roughly 4.0 to 4.5 million metric tons of urea out of the market. Tuesday strikes tightened it again and the whole nitrogen complex moved together, with Nutrien up 10.5 percent alongside CF. First half earnings were 8.71 dollars a share against 4.20 a year ago, which puts the stock near 7.9 times an annualised run rate. A single digit multiple on doubled earnings is the market saying it does not expect these earnings to last, and it may well be right.

   The catalyst. The United States completed a wave of strikes against Iran on September 1, 2026 and Iran retaliated against United States positions in Jordan, Kuwait, Bahrain, Iraq and the United Arab Emirates, tightening a nitrogen market already constrained by the Strait of Hormuz. CF Industries had reported on August 5, 2026 that second quarter net sales rose 17.6 percent to 2.222 billion dollars and net earnings rose to 727 million dollars, or 4.73 dollars a diluted share, from 386 million and 2.37 dollars a year earlier, and stated that the conflict has reduced Middle East traded nitrogen supply by approximately 4.0 to 4.5 million metric tons of urea and approximately 1 million metric tons of ammonia, from a region that typically accounts for 25 to 30 percent of globally traded ammonia and 35 to 40 percent of globally traded urea. CF rose 3.37 percent on August 31, 4.28 percent on September 1 and 2.71 percent on September 2. U.S. Securities and Exchange Commission, August 5, 2026

   Why it made the five. Because it is the second order version of the trade everyone made this week, and second order is where the crowd is not. The obvious response to crude settling 8.87 percent higher was to buy refiners, and the refining complex duly ran: HF Sinclair rose 9.43 percent, Valero 6.95 percent, CVR Energy 9.61 percent. I passed on all of them, and the reason is in the futures curve rather than in a hunch. Nitrogen fertilizer is made from natural gas, and by CF own published account the Middle East normally supplies 25 to 30 percent of globally traded ammonia and 35 to 40 percent of globally traded urea, with the company estimating roughly 4.0 to 4.5 million metric tons of urea and about 1 million tons of ammonia already removed from the market. The whole nitrogen complex moved together this week rather than CF alone, with Nutrien up 10.5 percent, which is what a genuine supply constraint looks like from the outside. And CF carries a correlation of negative 0.26 to the S&P 500 over the past three months, so it is doing something different from the rest of the index rather than the same thing louder.

   What to watch next. No company event is scheduled before third quarter results in November, so this one is checked daily by the Brent settlement and the Strait of Hormuz shipping picture rather than by a calendar entry Confirms the thesis: Brent holding above 90 dollars a barrel, which keeps the natural gas and shipping constraints that are throttling nitrogen output outside North America Breaks it: Brent settling below 85 dollars on a de-escalation headline, which has already happened once this year and took 6.6 percent out of crude in a week

   The case against it. This is a commodity price, not a company, and I want to be blunt about that. CF published nothing this week. It rose because a war got worse, and it will fall when the war gets better, on a timetable nobody in this newsletter can forecast. It closed at 97.1 percent of its 52 week high on only 1.20 times normal volume, the thinnest participation of the five picks, which is the weakest confirmation in the issue. The single digit multiple is not the bargain it looks like either: 7.9 times an annualised first half run rate is what the market pays for earnings it does not believe will persist, and the market is often right about that. The same de-escalation headline that took 6.6 percent out of crude in a week in early August would take the re-rating with it.

   Weekly move (before publication): +9.6%

   Every CF appearance →Track this pick on the Scoreboard →Size a position in it →

3. 03

   DE, view score card · Deere & Company · Industrials / Agricultural Machinery

   Friday Five Score

   76 / 100

   Move before publication

   +11.5%

   Actionability

   Extended

   Next date

   Sep 11

   Deere rose in a week when industrials fell 2.37 percent, and it is the pick whose trigger I trust least, so I will name it first: the proximate cause was Baird upgrading the stock to Outperform on August 31. What makes it more than a broker note is what happened underneath. China bought 703,000 tons of American soybeans over three business days, November soybeans traded above 13 dollars for the first time, and Deere own chief executive said in August that 2026 marks the bottom of the equipment cycle. Farm cash flow is the input to equipment demand, and it is turning. The price already knows: 38.6 times the midpoint of the earnings the company itself calls a trough.

   The catalyst. On August 31, 2026 Baird upgraded Deere to Outperform, citing North American row crop demand and mid 2027 corn futures moving above farmer breakeven. Over the same stretch China bought a further 136,000 tons of United States soybeans ahead of the September 1 open, taking three business days of buying to 703,000 tons for 2026 and 2027 delivery, and November 2026 soybeans traded above 13 dollars for the first time. On August 20, 2026 Deere reported third quarter net sales and revenues of 12.608 billion dollars, net income of 1.379 billion dollars and diluted earnings of 5.10 dollars a share, improved full year net income guidance to 4.75 to 5.00 billion dollars, and stated that it continues to believe 2026 will mark the bottom of the current agricultural equipment cycle. Deere rose 3.90 percent on August 31, 3.23 percent on September 1 and 3.30 percent on September 2. U.S. Securities and Exchange Commission, August 20, 2026

   Why it made the five. It is the third leg of the same chain and the one furthest from the headline, which is why it interests me. War raises natural gas, gas raises nitrogen, nitrogen raises the cost of planting, and grain prices decide whether a farmer can absorb it. This week the grain side finally moved in the farmer favour rather than against: China bought 703,000 tons of American soybeans over three business days, November soybeans traded above 13 dollars for the first time and set a contract high for a fifth consecutive session, and December corn set a contract high for a third. Farm cash flow is the input to equipment demand, and Deere own chief executive said on August 20 that the company continues to believe 2026 marks the bottom of the current agricultural equipment cycle. The stock rose 11.52 percent in a week when industrials fell 2.37 percent, so it is not sector drift.

   What to watch next. Sep 11: The USDA World Agricultural Supply and Demand Estimates update Confirms the thesis: A corn yield estimate held or lowered, which keeps grain prices where farmer cash flow improves and equipment orders follow Breaks it: A raised yield estimate that sends corn back toward its August range and removes the cash flow argument the upgrade rests on

   The case against it. The proximate trigger was a broker upgrade, and I am not going to dress that up as anything else. Baird moved Deere to Outperform on August 31 and the stock rose 3.90 percent that day. Broker upgrades are the weakest catalyst type this newsletter tracks. The valuation is the harder problem: at 694.41 dollars against full year net income guidance of 4.75 to 5.00 billion dollars on 270.7 million diluted shares, the stock trades at 38.6 times the midpoint of earnings the company itself calls a trough. That is a coherent price only if 2027 earnings are materially higher, and the recovery being argued for is a 2027 event that has not happened. The stock closed at 98.4 percent of its 52 week high, the most extended of the five, and tariffs remain a stated headwind in Deere own commentary. If the September 11 USDA report raises the corn yield estimate, the cash flow argument weakens in a single morning.

   Weekly move (before publication): +11.5%

   Every DE appearance →Track this pick on the Scoreboard →Size a position in it →

4. 04

   GTLB, view score card · GitLab · Technology / Software

   Friday Five Score

   82 / 100

   Move before publication

   +10.0%

   Actionability

   Fresh Breakout

   Software as a group fell 3.05 percent this week and GitLab rose 10.04 percent, which is the kind of divergence worth understanding. The company reported on September 1 with revenue of 286.3 million dollars, up 21 percent, a net retention rate of 117 percent, and 1,571 customers spending more than 100,000 dollars a year, up 17 percent. It raised the full year revenue forecast by about 1.4 percent and the operating profit forecast by about 9 percent. Note the asymmetry: the price moved further than the revenue guide did, so what is being repriced here is the margin trajectory, not this year sales.

   The catalyst. On September 1, 2026 GitLab reported second quarter fiscal 2027 revenue of 286.3 million dollars for the quarter ended July 31, up 21 percent year over year, with a non-GAAP operating margin of 15 percent, a dollar based net retention rate of 117 percent and 1,571 customers with more than 100,000 dollars of annual recurring revenue, up 17 percent. It raised full year fiscal 2027 revenue guidance to 1,129 to 1,133 million dollars from 1,112 to 1,118 million, and non-GAAP operating income guidance to 148 to 152 million dollars from 135 to 141 million. The stock rose 9.98 percent on September 2. U.S. Securities and Exchange Commission, September 1, 2026

   Why it made the five. Because it went the other way from its own group, and that is usually where the information is. Software as a whole fell 3.05 percent this week, giving back most of the pop it took on August 27, and GitLab rose 10.04 percent through it on 2.52 times normal volume. The quarter is the reason: revenue of 286.3 million dollars, up 21 percent and roughly five points ahead of the company own guidance, a dollar based net retention rate of 117 percent, and 1,571 customers spending more than 100,000 dollars a year, up 17 percent. The raise is verified against GitLab own prior published guidance rather than against a consensus estimate: full year revenue went to a range of 1,129 to 1,133 million dollars from 1,112 to 1,118 million, and full year operating profit to 148 to 152 million from 135 to 141 million.

   What to watch next. No company event is scheduled before third quarter results in December, which is unusually far out and is worth knowing before anyone treats this as a short horizon idea Confirms the thesis: The software group as a whole steadying, since a name that outperforms a falling group is a different proposition from one that outperforms a rising one Breaks it: Operating margin progress stalling at the next report, because the profit line is what the price just moved on

   The case against it. Read those two raises together and the picture is less comfortable. The revenue guide went up about 1.4 percent and the profit guide about 9 percent, while the stock went up 10.04 percent in a week. The price moved further than the business did, which means what re-rated is the margin trajectory rather than this year sales, and a margin trajectory is a promise about future quarters rather than a fact about this one. It is also the thinnest name in this issue at about 185 million dollars of average daily dollar volume, and share based compensation is material at this company and dilutes over time. There is no scheduled company event before the next report in December, which is an unusually long stretch with nothing to check the thesis against. And GitLab has been the subject of recurring takeover speculation, which can put a floor under a stock right up until it does not.

   Weekly move (before publication): +10.0%

   Every GTLB appearance →Track this pick on the Scoreboard →Size a position in it →

5. 05

   BBY, view score card · Best Buy · Consumer Discretionary / Specialty Retail

   Friday Five Score

   80 / 100

   Move before publication

   +4.7%

   Actionability

   Watch Pullback

   Next date

   Sep 11

   Best Buy raised all four of its full year guidance lines on August 27, including comparable sales going to a range of 1.9 to 3.0 percent from a range of negative 1.0 to positive 1.0 percent. It fell 4.44 percent that day and kept falling for two more sessions. Then it rose 3.09 percent and 5.29 percent as the market changed its mind. This is the same pattern I wrote about in August when Walmart raised every line and fell 9 percent, except this time the market came back for it. At 12.9 times the midpoint of its own raised earnings forecast, the price is not assuming the raise repeats. It is only assuming the raise holds.

   The catalyst. On August 27, 2026 Best Buy reported second quarter fiscal 2027 revenue of 9.779 billion dollars with comparable sales up 4.1 percent and adjusted diluted earnings of 1.47 dollars a share, up 15 percent. It raised all four full year fiscal 2027 guidance lines: revenue to 42.3 to 42.8 billion dollars from 41.2 to 42.1 billion, comparable sales to 1.9 to 3.0 percent from negative 1.0 to positive 1.0 percent, the adjusted operating income rate to 4.4 to 4.5 percent from 4.3 to 4.4 percent, and adjusted diluted earnings to 6.70 to 6.90 dollars from 6.30 to 6.60. The stock fell 4.44 percent that day and a further 1.34 percent and 2.78 percent, then rose 3.09 percent on September 1 and 5.29 percent on September 2. U.S. Securities and Exchange Commission, August 27, 2026

   Why it made the five. It is the smallest gain in the issue and the one I find most instructive. Best Buy raised all four of its full year guidance lines on August 27 and the market sold it for three straight sessions, down 4.44 percent, then 1.34 percent, then 2.78 percent. Then it changed its mind and bought it back, up 3.09 percent and 5.29 percent. I wrote in Issue No. 09 that the market had stopped paying for good news, using Walmart falling 9 percent on a raise as the example. This is the same setup with the opposite ending, two weeks later, and watching a market reverse its own verdict on the same kind of news is worth more than another name that only went up. The raise itself was not marginal: comparable sales came in at 4.1 percent against a full year outlook that had been negative 1.0 to positive 1.0 percent, and the company moved that outlook to 1.9 to 3.0 percent.

   What to watch next. Sep 11: The August Consumer Price Index Confirms the thesis: Goods inflation staying contained, which is what a raised electronics guide needs in order to survive a crude move of this size Breaks it: A hot goods print, since the reported quarter already leaned on a tariff refund and the same policy that helped it can reverse

   The case against it. Part of the reported quarter came from a tariff refund rather than from selling more televisions, and the same policy that produced the refund can reverse. The forward guide is modest in absolute terms: 1.9 to 3.0 percent comparable sales growth is a recovery from a decline, not an expansion, and consumer electronics is a category that has spent years being deflationary and discounted. The stock closed at 95.9 percent of its 52 week high, so the market has now paid for the raise it initially refused, which removes the specific thing that made this interesting. And the September 4 employment report and the September 11 inflation print both land inside the next week: this is the pick most exposed to the household side of the same oil move that is helping CF Industries.

   Weekly move (before publication): +4.7%

   Every BBY appearance →Track this pick on the Scoreboard →Size a position in it →