Silver banana goes to…

In this issue of the peel:

• 🏪 Consumer spending got off to a weak start in January, but it’s tough to compete with the month that boasts holidays like Toyotathon. Find out why markets didn’t about the spending dip care below.

• 🃏 DraftKings would have you believe that gambling is always a win for everyone involved, while Airbnb rose on announcing “new businesses or somethin’.” Meanwhile, Applied Materials got too big for its britches, and investors in Coinbase are dipping out on the eve of cr*pto regulation.

• 🧑‍⚖️ The first AI court ruling has been made—and I almost guarantee you haven’t heard about it. Soak it all up while humans are still the ones making the rulings.

Banana Bits

• Europe is tweaking over conversations between U.S. President Trump and Russian President Putin.

• And as a result, they’re finally thinking about how to defend themselves.

• Chinese self-driving firm WeRide booms over 83% after Nvidia disclosed an ownership stake.

• Speaking of which, here are the rest of the AI winners, according to Nvidia’s latest full disclosure.

• The DOGE effect is already impacting public companies.

• But DOGE’s power may be on the verge of becoming greatly reduced.

• Mind Medicine had a great trip on Friday, thanks to RFK’s confirmation as Secretary of Health and Human Services.

The Daily Poll

Previous Poll:

What do you think about Elon Musk’s $97.4bn buyout offer for OpenAI?

Genius move – Musk is playing 4D chess: 20% // Another chaotic Musk moment: 39.5% // Just a PR stunt, won’t happen: 40.5%

Tilt the Odds in Your Favor

In the competitive world of high finance, every advantage counts. Our exclusive curriculum, designed by industry experts, sharpens your skills and knowledge, making you a top candidate.

Enjoy personalized coaching, targeted internship opportunities, and a robust network of finance professionals with WSO Academy. 

» Apply Now «

Macro Monkey Says

Retail Sales

Lexus was absolutely right—that really was a December to Remember. They just didn’t tell us it was gonna be followed by a January we’d love to forget.

The Holidays give us a great excuse to go out and spend/waste money on others in order to make ourselves feel good. January is a great opportunity to make yourself feel even better—by returning those gifts.

Maybe returns count as negative revenue? That’s the only explanation I can think of for the plummet in retail sales last month.

Let’s get into it.

The Numbers

Clearly, no one at the Commerce Department was anyone’s Valentine this year as the agency tried to completely kill the vibe on Friday, reporting the largest monthly drop in retail sales in almost 2-years.

Total retail and restaurant sales declined 0.9% in January. Luckily, markets largely shrugged off the news.

Equities broadly rose in the morning session following the report’s 8:30am release but consolidated in the afternoon session. The S&P 500 finished essentially flat (down 0.01%) and the Nasdaq up 0.41%. Not that anyone cares, but the Dow fell 0.37%.

Treasury yields sank as January’s weakness in spending growth gives JPow and the FOMC data in support of further cuts, increasing the probability that the path of rates goes lower for sooner, as opposed to higher for longer.

This was especially encouraging, considering Friday was a brief example of the rare instance in which markets are capable of separating the signal from the noise in real-time.

Yes, this was the worst monthly growth figure since March 2023.

However, annual growth remained strong, and January still earned the Bronze Medal in all-time Retail Spending, posting the 3rd highest dollar amount spent in a single month in our history, behind only December and November.

Speaking of those months prior, the good news in the report was found in December’s upward revision. Previously, the Commerce Department reported a 0.4% monthly increase, but on Friday, we found out December actually posted a 0.7% increase.

So, taking a longer-term view, there’s not much reason for concern… unless you happen to be a care salesman.

It’s hard to follow the Holiday szn with Americans busy celebrating Lexus’ December to Remember, Happy Honda Days, and, of course, the big dawg of them all, Toyotathon. So, it makes sense why auto sales fell 3% monthly, even with seasonal variations.

After fueling up on vehicles in December, it only made sense to celebrate by taking those bad boys to bars and restaurants all throughout January. Sales at “Food Service and Drinking Places” were up 5.4% from last year and 0.88% for the month.

By the way, in case you need more evidence that no one at the Commerce Department has a Valentine, just look at how they call restaurants “Food Service and Drinking places.” That only happens if you’ve never been invited to one.

Anyway, clearly, this shows the real growth in January was in the great American pastime of drunk driving. At least fun is on the rise, even if retail sales weren’t.

The Takeaway?

Props to Mr. Market for not absolutely tweaking out like he usually does. This is a textbook case of separating signal from noise—one “bad” month doesn’t mean consumer spending is trending in the wrong direction.

And we can see this in the data. Earlier this month, we talked about how Bank of America reported strong spending growth among its customers in January.

Plus, January’s 0.9% decline was 1) dragged lower by weakness in auto sales and 2) pulled higher by strength in restaurant spending, not something you’d expect to see if consumers are pulling back on spending.

Just make sure you do your part in February. We only have 28-days, so make sure to spend as much as physically possible—the economy’s counting on it.

Career Corner

Question

If net debt increases, the EV goes up, and the EQV goes down. (Net debt = debt - cash). However, If debt increases, EV doesn’t change, and EQV doesn’t change. Is this correct?

Answer

For the second point, yes, as long as the new debt sits as cash on the BS, the net debt remains the same.

Your first point is not such an easy statement.

If your point is that you take on more debt and use that money to invest in the business, the EV goes up, but equity value (from a book perspective) stays the same because you haven’t done anything to affect equity value (EV increases, but debt increases as well and no change in cash since you spent the funds, so equity value flat).

Head Mentor, WSO Academy

Check out WSO Academy

What's Ripe

DraftKings (DKNG) 15.2%

• There’s a lot of talk about OpenAI becoming a “public benefit company,” and personally, I can’t stand it. What an insult to real, charitable companies that exist solely to benefit the public, like DraftKings.

• Gambling is one of those rare things in society where literally everybody wins. After getting fleeced all NFL season, shares still managed to boom despite missed earnings.

• DraftKings reported a loss of $0.17/sh on $1.39bn in revenue, hitting the under vs analyst estimates for a loss of $0.10/sh on $1.4bn. Monthly unique “players” grew 36% YoY.

• Shares rose on strong user growth and a slight upgrade to Fy’25 revenue guidance, expecting a base case of $6.3bn–$6.6bn vs prior estimates for $6.2bn–$6.6bn.

Airbnb (ABNB) 14.5%

• Sure, beating earnings is nice. But to get your best trading day ever, you gotta go a little further. Airbnb’s innovative strategy was to hype up a “thing” that might happen. Exciting stuff.

• The vacation rental platform reported EPS of $0.73/sh on $2.48bn in revenue vs estimates for $0.58/sh on $2.42bn. Revenue grew 12% annually as a gross booking value of $17.6bn beat estimates, too.

• However, shares really popped on an announcement of plans to launch “one or a couple businesses” with the goal of becoming “... kind of similar to Amazon—to be one place to go for all your travel and living needs.”

• Don’t worry about the conspicuous lack of detail and randomly thrown-out guesstimate that these businesses could earn $1bn in revenue. They said it might be like Amazon! 

What's Rotten

Applied Materials (AMAT) 8.2%

• The El Chapo of the semiconductor manufacturing industry, Applied Materials, dove on a weaker-than-expected outlook despite strong Q4 earnings. Blame DeepSeek and tariffs.

• The company reported EPS of $2.38/sh on $7.17bn in revenue vs estimates for $2.28/sh on $7.15bn. Unfortunately, markets don’t give a f*ck what you did, just what you’ll do.

• Applied Materials expects a $400mn revenue headwind from tariffs in 2025, adjusting their revenue guidance to $7.1bn in Q2 while the market was looking for $7.21bn.

Coinbase Global (COIN) 7.9%

• This is why you should always break the law. Coinbase followed all the rules even when there weren’t any, and now that the Trump Administration wants to make the U.S. the “cr*pto capital of the world,” they’re first on the chopping block.

• Not last quarter, however. The digital asset brokerage killed it in Q4, reporting an explosion in net revenue, up 88% QoQ and 143% YoY. FY’24 sales grew 115% from FY’23.

• Total revenue of $2.27bn beat estimates for $1.84bn while EPS of $4.68/sh made a joke of Street estimates looking for $2.11/sh. Unfortunately, markets are forward-looking.

• The adoption of a digital asset regulatory framework could hurt Coinbase as other exchanges will know how to run without breaking the law, potentially lowering barriers to entry.

Thought Banana

F-AI-r Use

Anyone who listens to the All-In Podcast, of which I imagine there’s a lot of crossover here, must’ve been just as utterly shocked as I was to realize on this last episode that Jason Calacanis was actually right about something.

A broken clock is right twice a day, I guess. But let’s see what this guy—who Nassim Taleb has called a “psychotic ignoramus”—was yapping about.

What Happened?

When we think of AI lawsuits, not many come to mind. Sure, there’s a handful out there, and maybe you're plugged in enough to be aware of the New York Times’ ongoing lawsuit against OpenAI. But…

The most important legal decision related to AI yet just happened right under our noses. No one—outside of super nerds who have read more legal textbooks than bar menus—was talking about the Thomson Reuters case against Ross Intelligence.

Storytime: Founded in 2014, Ross Intelligence was a startup building an AI research tool for law/legal cases. In May 2020, Thomson Reuters filed a lawsuit against Ross, alleging unauthorized use of content from its own legal database by the AI startup.

By January of 2021, Ross had to close up shop, citing litigation costs as the primary factor.

And over 4-years later, on February 11th, 2025, the case finally came to an end with U.S. District Judge Stephanos Bibas (sick name) granting a partial summary in favor of Thomson Reuters.

Who Cares?

I know it’s been kinda underwhelming so far, but trust me, this was huge.

The reason Thomson Reuters sued was because Ross Intelligence was using facets of a legal database owned by Thomson Reuters, called Westlaw, without permission from Thomson Reuters in order to train their AI models.

Specifically, Ross was relying on Westlaw’s headnotes and Key Numbering System.

Ross argued that although their models did use Westlaw’s database to train, this action was covered under the Fair Use doctrine.

Talking about Fair Use to a copyright lawyer is like asking a Roman historian why the Empire fell—everyone’s got an opinion, and they all only seem partially right.

Basically, Fair Use is a protection that allows the use of copyrighted materials without the copyright owner’s permission in some cases. Generally, this applies to works of criticism, commentary, education, parody, and news reporting.

Fair Use is what allows your favorite commentary YouTube channel not to get hit with a cease and desist immediately after upload.

Anyway, this is to say that Ross Intelligence’s use of Westlaw to train its AI model constituted a novel application of the Fair Use doctrine, where LLMs use copyrighted works to train on and base their answers.

Ross argued that because its models were not directly plagiarizing from Westlaw and instead were using the information within Westlaw to power its responses, this was covered under the Fair Use doctrine.

However, courts disagreed on the basis of two primary factors:

• Transformation: A key aspect of Fair Use is that the work has to be transformative. Because Ross relied on the headnotes and Key Numbering System directly from Westlaw, the threshold for transformation was not met.

• Competition: Someone related, a key aspect of works being considered “transformative enough” is that the work is altered so that the reproduced work does not compete directly with the original. In this case, courts found that Ross was, in fact, directly competing with Westlaw while using Westlaw’s database to do so.

This is huge because what Ross did to Westlaw is broadly what LLMs like ChatGPT and others do to the open internet. The difference is that Westlaw is a proprietary legal database, while ChatGPT searches open, publicly available information.

The ChatGPTs of the world also do not reproduce content directly but summarize and interpret things, giving original responses that surpass the Fair Use doctrine.

This gets tricky with sites like Getty Images, CNBC.com, Reddit posts, or similar content. 

For example, The NYT’s content is behind a paywall and is clearly proprietary, but what about Getty’s publicly available photos or CNBC’s articles that are absolutely proprietary but also publicly available?

Public availability does not necessarily imply the works belong to the public domain. This is the legal gray area copyrights are looking at without salivating mouths and soon-to-be-filled wallets.

The Takeaway?

We’re gonna need a minute to figure this one out.

Because LLMs present an entirely novel application of the Fair Use doctrine, there’s little/no precedent for deciding what makes a work covered or not by Fair Use. Applying the threshold to LLMs is totally new.

That’s what the legal system needs to figure out. For example, if someone asks ChatGPT about January’s Retail Sales report and it starts by hyping up Toyotothon, is it plagiarizing today’s Peel? Where’s my $97.4bn buyout offer?

This is gonna be interesting.

The Big Question: How should Fair Use apply to LLMs? 

Banana Brain Teaser

Previous

The product of two negative numbers is 160. If the lesser of the two numbers is 4 less than twice the greater, what is the greater number?

Answer: -8

Today

Carol purchased one basket of fruit consisting of 4 apples and 2 oranges and another basket of fruit consisting of 3 apples and 5 oranges. Caril is to select one piece of fruit at random from each of the two baskets. What is the probability that one of the two pieces of fruit selected will be an apple and the other will be an orange?

Send your guesses to [email protected]

How Would You Rate Today's Peel?

Happy Investing,
David, Vyom, Ankit & Patrick

ADVERTISE // WSO ALPHA // ACADEMY // COURSES // LEGAL