In a bizarre twist of fate, tech executives have come to resemble Oliver Twist. Semiconductor procurement officers, dubbed “DRAM Beggars,” are reportedly camping out near major South Korean chipmakers, trying to secure scarce memory supplies.
Memory has become the latest bottleneck in the AI supply chain, causing DRAM (Dynamic Random-Access Memory) prices, along with the profits of the companies that manufacture it, to skyrocket. According to Bloomberg, the world’s two largest suppliers, Samsung and SK Hynix, now make up over 50% of the iShares MSCI South Korea ETF (EWY) and are expected to earn $307 Billion in 2026 compared to $51 Billion in 2025, a 6-fold increase.[1] The windfall profits are so large that workers may take home six- or even seven-figure profit-sharing checks. South Korea’s Presidential Chief of Staff for Policy even went as far as suggesting a national dividend.
This is the latest example of the mad rush to build datacenters rippling through global markets. While it’s clearly a boon for memory producers, it’s hard to imagine that there won’t be any knock-on effects from such a sharp rise in the price of a critical input to the modern economy, not to mention datacenters. During a gold rush, the market may assume that 1 + 1 = 3 but eventually, the sum of the parts needs to equal the whole.
- Denis Rezendes, CFA, Partner and Portfolio Manager
In this edition:
(Not so) Dire Straights?
The Hour-Long Project We Never Would’ve Started
The Wire Fraud Epidemic and How to Protect Your Firm (Podcast)
Herd Behavior: ETF Leaders and Laggards
Points of Interest What the PM team is talking about right now
(Not so) dire straights? We’ve crossed the 80-day mark since the start of the Iran war. While oil prices have rallied, the visibility of the disruption to the American consumer thus far seems to be rather limited: discourse is largely confined to what citizens are paying at the pump. To be sure, this is nothing to scoff at: the additional $40B that Americans have spent on gasoline and diesel since the outbreak has outpaced the (admittedly dubious) $30B figure of direct costs estimated by the Pentagon[2].
There are cracks under the surface. Fertilizers were one of the immediate markets roiled by the closure of Hormuz[3]. With supply from the Gulf shut out, one might’ve anticipated domestic suppliers as a rare beneficiary of the dislocation. However, Tampa-based Mosaic curtailed their phosphate production as input costs (namely sulfur) outpaced the rise in fertilizer prices enough to make marginal production uneconomic[4].
Thus far, American farmers appear more insulated from the spike, having already secured a majority of their needs before the disruption[5]. Thus far, is the operative phrase. As the year progresses, they will inevitably have to secure their future fertilizer needs for 2027 and if both diesel and fertilizer prices are still elevated, something will have to give.
Adding to further uncertainty around agricultural prices is the emergence of El Niño. The climate pattern results in warmer and dryer weather in southeast Asia, India, and Brazil, which historically results comes with higher food and broader commodity inflation[6]. The most recent forecast from the NOAA places a 60% probability of a strong or very strong El Nino emerging by September[7]. A silver lining is that the weather pattern also results in more favorable conditions for American farmers, in aggregate.
The confluence of fragmented supply chains, uncertain weather, and rising input costs do pose a demonstrated risk of food price volatility, and it’s something we will certainly follow in the ensuing months.
- David Hill, Data Science Engineer
The Hour-Long Project We Never Would’ve Started
We may often come off as skeptical about AI in this newsletter. Our skepticism is primarily centered on the AI capital spend and AI company valuations rather than the AI tools themselves, which we have found both useful and exciting. In that spirit, I’m going to walk you through a simple example of how we have used AI tools in our research process.
Our portfolio has had a fair amount of international exposure over the last 9 months, so I wanted to create a monitor of how various countries have been performing of late. I can pull a list of foreign ETF trailing returns with no effort, but what I really wanted was a map so I could easily see regional variations vs. single country variations.
There’s no driving business need for this map and it’s unlikely to generate a dollar of revenue. And so just a few years ago, I would not have bothered to take the week or so of time it would have taken to write and debug the map code myself. With ChatGPT (or Claude or Gemini or your LLM of choice), I no longer needed to consider whether making that map was worth a week of my time because I could now accomplish it with about 30-60 minutes of active time while ChatGPT did the brunt of the work.
I began by double checking with the LLM whether I had an exhaustive list of single-country ETFs. It suggested three countries I hadn’t included, but it turned out these had been de-listed anyway. I ended up with a list of 44 countries that have dedicated US-listed ETFs.
Next, I simply told the LLM I wanted Python code to be able to generate a daily map of trailing 1-month returns for each of the countries in my list relative to the all-world index. I got semi-workable Python code within about 90 seconds. The rest of the minimal time invested was spent debugging and tweaking the output.
If I was doing this myself, any error would have resulted in combing line-by-line through the code, and if nothing obvious jumped out at me, searching for documentation on the map or data retrieval libraries and/or combing through old Stack Overflow posts. Each bug could have taken 1-4 hours to resolve on its own. Instead, I got workable, fixed code back within 30-120 seconds of LLM thinking time after sharing the error messages in the chat.
And with minimal effort, I now have a script I can run every day that generates this map, allowing me to check in on the latest foreign stock market performance as part of my morning routine.
- Andrew Rice, Partner and Portfolio Manager
Advisor Turntable Podcast
Episode 5:
The Wire Fraud Epidemic and How to Protect Your Firm
In this episode
Wire fraud is no longer an obvious, easy-to-spot scam. The reality today is much more sophisticated.
In this episode of the Advisor Turntable Podcast, Brendan Ryan talks with Michael Brice, founder of BW Cyber, about the growing epidemic of wire fraud targeting financial firms, advisors, and even individual investors. In addition to working with advisors and asset managers through BW Cyber, Michael works with the FBI and other government agencies on cybersecurity investigations and has seen it all. He shares real-world examples ranging from compromised email threads to AI-generated voice cloning and deepfake video calls being used to manipulate wiring instructions.
One of the more reassuring takeaways from the conversation: most wire fraud is not really a technology problem — it’s a human one. And in many cases, relatively simple verification procedures and operational controls can dramatically reduce risk.
As AI continues making digital communication harder to trust, this is becoming an increasingly important operational issue for advisors and firms to understand. The cost of improving these processes is often insignificant compared to the financial and reputational damage one successful scam can cause.
Advisor Turntable is short-form series featuring experts and advisors who have moved beyond outdated industry tracks and adopted the modern tools, revenue models, and workflows today’s clients expect.
Herd Behavior
ETF Leaders and Laggards
1-Month Trailing Returns as of 5/18/2026
Top performers:
Semiconductors (SOXX) +19.28%
Technology (QTEC) +16.28%
Cybersecurity (HACK) +16.14%
Bottom performers:
Indonesia (EIDO) -15.66%
Gold Miners (GDX) -13.16%
Nuclear Energy (NLR) -12.63%
Keep reading below to learn the factors and themes we believe are driving these returns.
Source: EOD Historical Data. These are the three highest- and lowest-returning ETFs over the stated period drawn from the curated 230-ETF investment universe used in our Decathlon strategies. We use this universe because we believe it captures all investable areas of the global markets while excluding derivatives, single-company ETFs, overlapping exposures, and ETFs with very low AUM. We believe this makes it a better lens for highlighting prevailing market themes than looking at all available ETFs on the market. Our strategies may or may not own one or more of these ETFs. ETF performance is net of the ETFs internal expense ratio and includes reinvestment of dividends. Past performance is no guarantee of future results.
Themes
This month’s top and bottom performing ETFs continue to highlight the biggest themes driving the market—Datacenters and commodities. The best performing ETFs were universally focused on AI capex. The bellwether of the AI trade, SOXX was the top performer as it continued a historic run over the last two months on the back of further Datacenter capex and continued shortages in memory chips driving sky high price increases within that group. Nasdaq Technology, Cybersecurity, South Korea (memory makers SK Hynix and Samsung are more than 50% of the ETF), and Solar round out the top 5. There continues to be a loose tie in between clean energy and data centers due to their voracious energy appetites.
The worst performing ETFs had far less unifying theme, some represent volatile metals which fell rapidly in just the last few days, while ironically, others represent economies uniquely hurt by higher oil prices such as Indonesia, the worst performer on the month. Gold miners and the Uranium and Nuclear sector found themselves as the next worst, round-tripping gains from just a few weeks earlier and highlighting how narrative-driven the market has become. Rounding out the 5 worst performing ETFs, Brazil had some recent politics driven volatility and Homebuilders continue to face headwinds from a lack of mortgage rate relief.
- Brendan Ryan, CFA, Partner, Portfolio Manager
ICYMI
In case you missed it, here is the latest from Our Thinking:
We have a 13-year track record of helping investors seek consistent returns and lower volatility.
Our global, multi-asset strategies are designed to:
This post is for informational purposes only and does not constitute advice, a recommendation, or an offer to sell or solicit any security or financial product. Inherent in any investment is the risk of loss.
The information presented in this report is based on data obtained from third party sources. Although it is believed to be accurate, no representation or warranty is made as to its accuracy or completeness. The charts and infographics contained in this blog are typically based on data obtained from third parties and are believed to be accurate. The commentary included is the opinion of the author and subject to change at any time.
Certain information contained herein constitutes “forward-looking statements,” which can be identified by the use of forward-looking terminology such as “may,” “will,” “should,” “expect,” “anticipate,” “project,” “estimate,” “intend,” “continue,” or “believe,” or the negatives thereof or other variations thereon or comparable terminology. Due to various risks and uncertainties, actual events, results or actual performance may differ materially from those reflected or contemplated in such forward-looking statements. Nothing contained herein may be relied upon as a guarantee, promise, assurance or a representation as to the future.
The views and opinions expressed throughout this presentation are those of the presenter as of date the podcast is episode is published. The opinions or outlooks may change over time with changing market conditions or other relevant variables.
Algorithmic Investment Models (AIM) is an asset manager registered with the SEC and provides ETF portfolios to financial advisor and institutions.
Algorithmic Investment Models, LLC (AIM), 125 Newbury Street, 4th Floor, Boston, MA 02116, United States, 844-401-7699
%20no%20logo.png?upscale=true&width=1200&upscale=true&name=Un-Herd%20Banner%20(1200%20x%20400%20px)%20no%20logo.png)
.png?upscale=true&width=60&upscale=true&name=iconfinder_linkedin_1279057%20(2).png){kind=small}
