AI Spending Goes Brrr

🌐 The only thing non-artificial about artificial intelligence is the ungodly amount of money companies are spending on this stuff. Find out what it means for the space below.

November 25, 2024




In this issue of the peel:

  • 🌐 The only thing non-artificial about artificial intelligence is the ungodly amount of money companies are spending on this stuff. Find out what it means for the space below.

  • 👕 Gap shares gapped higher on earnings while Super Micro continued to pump for reasons that are borderline too dumb to discuss. Reddit shares plunged on institutional moves, and Palo Alto Networks managed to have a worse quarter than CrowdStrike.

  • đŸ¶ It’s a bad day to be government spending. Elon and Vivek are coming for you, outlining exactly how they plan to un-bloat Uncle Sam below.

Market Snapshot

Banana Bits

Diversify Your Portfolio with UpMarket's Hedge Fund Access

UpMarket is making it easier for accredited investors to access alternative investments. Here's why you should consider adding hedge funds to your portfolio:

  1. Diversification: Access uncorrelated strategies and new asset classes.

  2. Professional Management: Enjoy the potential benefits of professional risk management practices.

  3. Liquidity: Hedge Funds typically have regular redemption windows and can offer more predictable liquidity than other alternatives.

  4. Performance Potential: Aim for high returns, especially during market volatility.

With UpMarket, you can invest in hedge funds and other alternatives seamlessly. Their platform has already helped investors deploy over $150 million into private markets.

Ready to take your portfolio UpMarket?

Macro Monkey Says

AI Is At The Thug Tug

Cinematic masterpieces like the original SpongeBob Movie are rare—one per generation, if we’re lucky.

You probably have every line memorized, but who could forget the classic moment at the Thug Tug, when SpongeBob and Patrick learned: “All bubble-blowing double babies will be beaten senseless by every able-bodied patron in the bar.”

Well, it seems AI was at that bar because there’s not a bubble in sight—business spending is booming.

Let’s get into it.

What Happened?

Midway through last week, analysts at Menlo Ventures came off their latest ayahuasca trip and dropped this banger of a report called “2024: The State of Generative AI in the Enterprise.” And we have to talk about it.

Enterprise operational spending on artificial intelligence products and services skyrocketed 513.33% annually to $13.8bn in 2024 (so far), a more than 6x increase.

Menlo’s report excludes capital expenditures (investments), so that’s why you’re not seeing the impact of the ~$50bn spent by the four big tech companies alone, including Meta, Alphabet, Amazon, and Microsoft.

This explosion in spending is exactly what separates a bubble from a boom. Unlike with blockchain in 2017 and 2021 or Dutch tulips in 1636, companies are putting serious money where their mouth is in AI.

Vertical AI led AI spending growth in 2024, up 1,100% YoY. This segment refers to “AI solutions tailored for specific industries or sectors”, per ChatGPT, who I guess is somewhat of an expert on the subject?

According to Menlo, foundational models became technically advanced enough in late 2023 and early 2024 to allow professionally viable solutions for certain industries to employ AI systems where there otherwise would have been employees.

Unlike you, AI doesn’t ask for time off to get hammered with its friends in Destin, Florida, twice a year. So, it makes sense why we’re seeing this explosion in companies looking to make AI programs to take your job.

Each segment grew an insane amount, with the slowest increase found in spending on Training & Deployment, likely because investments had to be made here earlier in order to support less foundational functions like those above. Still, it grew 283% in 2024.

Along with spending, the one thing digital assets and tulips alike couldn’t produce in their heyday is blowing up in AI, too: actual, valuable use cases.

Of the companies surveyed by Menlo, the use case with the highest adoption rate is Code Generation at 51%.

Next, we see use cases like Support Chatbots, Enterprise Search & Retrieval, and Data Extraction & Transformation. The common denominator here is that each of these use cases focuses on efficiency via cost or time reduction.

Companies are prioritizing ROI over innovation and appear to prefer attaching an AI agent or other kind of support system to human employees rather than replacing them outright
 for now.

That might seem obvious in hindsight, but consider this: If you were to make an investment in an AI firm, which use case would you want to get a piece of in 2025?

It makes sense to track which segment of spending and use case will deliver the most ROI in the shortest timeframe, but which segment or use case will that be? Does meeting summarization create more efficiency than using AI to coach new hires?

These are the questions faced by companies procuring AI systems, but identifying the companies providing those AI systems is much easier to solve:

As the trendsetter of the industry, it’s natural for OpenAI to cede market share after literally creating the industry. Anthropic and Google led market share growth in 2024, but don’t sleep on the “Internal Model” at 3%—in 2023’s report, it wasn’t even mentioned.

Maybe that doubling in market share is what led Amazon to jam another $4bn down Anthropic’s mouth


The Takeaway?

So far, enterprises' preferred segment of AI spending has focused on internal, vertical use cases that deliver maximum short-term ROI.

That’s been done by using foundational LLMs like ChatGPT, Claude Sonnet, Perplexity, Anthropic, etc., and then building internal use cases on top of them.

Following the dollars to find the next highest-growth segment of spending will be challenging for investors, but the one thing that’s clear is that there are no bubble-blowing double-babies within this market.

Honestly, go check out the report. It’s pretty short and provides way more info than I can due to my wordcount limit—plus, I understand these systems about as well as my dog does.

Career Corner

Question

I was wondering if, in interviews for IB internships, the interviewers ask questions about Excel. Do we need to know Excel well and how to build models or DCFs in Excel for interviews?

Answer

Unlikely, you'll get a question about Excel specifically. I would focus more on understanding the concepts behind modeling, DCFs, LBOs, etc.

You may get a case study, in which case Excel work and speed will matter, but unlikely in a face-to-face interview.

Hear Mentor, WSO Academy

What's Ripe

Gap (GAP) 12.84%

  • Kanye West said, “Look at my check, wasn’t no scratch” when rapping about his time working at Gap, but bro should’ve just bought shares. Q3 was an all-time.

  • Gap beat sales and earnings estimates for last quarter, causing investors to get extra hyped as the rest of the consumer discretionary reels (*cough*, Target).

  • Total sales grew 2% and same-store sales 1%. The firm’s higher-end brands, like Banana Republic and Athleta, lead sales growth, up 2% and 4%, respectively, while Athleta’s comp sales grew 5% YoY.

  • The apparel retailer also raised guidance across sales, margins, and operating income for FY’2024.

Super Micro (SMCI) 11.62%

  • Like Ben Graham said, meat riding is way more important than fundamentals. And that’s exactly what Super Micro looks to do as shares ride Nvidia’s coattails.

  • Shares gained >84% in the last 6 trading sessions as of Friday. Part of the bump was driven by getting namedropped on Nvidia’s earnings report as a “reliable partner” of the AI king.

  • However, the even better news from last week was that Super Micro pursued a bunch more band-aid-style strategies to avoid getting delisted from the Nasdaq. We’ll see if that holds up.

What's Rotten

Reddit  (RDDT) 7.18%

  • Reddit’s just a chill guy trying to host discussion forums and a ton of porn while institutional investors are killing the vibe with share-backed borrowing shenanigans.

  • On Friday, investors learned that Advance Magazine Publishers, the parent company of Conde Nast, is using its 7.8mn Reddit shares as collateral to get $1.2bn in liquidity.

  • Shared tanked as traders took this as a sign of declining confidence in Reddit. Still, AMP purchased derivative contracts to maintain exposure to Reddit’s price moves.  

Palo Alto Networks (PANW) 3.61%

  • Kendrick came out of Compton to drop a surprise album on Friday, while Palo Alto Networks came out of another California city to drop a surprisingly bad earnings report.

  • The cybersecurity firm floundered despite beating sales and earnings estimates, with revenue up 15% YoY because of a 14% decline in bookings.

  • Analysts took the slowdown in billings, which reflects future orders, as a sign that the company’s goal of being the one-stop-shop for all things B2B cyber is not working on its planned schedule.

Thought Banana

The Game Plan

They might not have chainsaws like Javier Milei or the ambition of Milton Friedman, but the two heads of our soon-to-come Department of Government Efficiency just laid out exactly how they plan to save America.

Let’s dive in.

What Happened?

Late last week, Elon Musk and Vivek Ramaswamy wrote an op-ed in the Wall Street Journal outlining their plan to Make America Efficient Again.

Obviously, step one was to have two people share the role of one leader of this commission, but we’ll give ‘em a pass since neither are getting paid.

For starters, it’s clear that Vivek—a graduate of Yale Law School—did most of the writing. You’d need at least a pre-law degree to understand what the hell they’re talking about, but let’s go over the highlights.

The leaders of DOGE announced a three-pronged plan on how to reduce government spending, targeting 1) Regulatory Rescissions, 2) Administrative Reductions, and 3) Cost Savings.

The regulatory rescission strategy hinges on recent Supreme Court decisions, including West Virginia v. EPA and Loper Bright v Raimondo, the latter half we wrote about on July 2nd.

Basically, these Supreme Court decisions overturned existing legal precedents that allowed agencies to write their own regulations beyond the authority explicitly granted them by Congress.

So, now that courts no longer have to defer to agency interpretations, which effectively gave agencies the ability to create their own regulation, Musk and Ramaswamy said this “suggests that a plethora of current federal regulations exceed the authority Congress has granted under the law.”

This leads to the second prong of DOGE’s plan, administrative reductions. Basically, the idea here is that as agency regulations are removed by courts—and because agencies can no longer make their own regulations—a lot of people can get fired.

Finally, the third prong focused on cost savings is set on targeting the ~$500bn in current federal spending that was not authorized by Congress. 

The examples cited include a $535mn annual payment to the Corporation for Public Broadcasting, $1.5bn awarded to international organizations, and more.

The Takeaway?

It’s definitely worth the ~10mins to read this if you’re an American citizen, and especially if you’re a government official to figure out how you can keep sucking from the cash cornucopia that is the federal government.

It’s unclear whether DOGE’s efforts will be enough to end the death-debt spiral the U.S. is currently in without cutting entitlement spending, but any step in this direction is a welcome one to those concerned about government spending.

The goal is to shut down DOGE by July 4th, 2026—America’s 250th Birthday—so expect a lot of noise over this commission until then.

The Big Question: Will DOGE get the job done? What happens if they can’t?

Banana Brain Teaser

Previous

From 2000 to 2003, the number of employees at a certain company increased by a factor of 1/4. From 2003 to 2006, the number of employees at this company decreased by a factor of 1/3. If there were 100 employees at the company in 2006, how many employees were there at the company in 2000?

Answer: 120

Today

A collection of 16 coins, each with a face value of either 10 cents or 25 cents, has a total face value of $2.35. How many of the coins have a face value of 25 cents?

Send your guesses to [email protected]

❝

The intelligent investor is a realist who sells to optimists and buys from pessimists.

Benjamin Graham

How Would You Rate Today's Peel?


Happy Investing,
David, Vyom, Ankit & Patrick