If you’ve spent any amount of time on Twitter, you’ve likely come across the hypothesis that AI will bring an extinction level cataclysm for white collar workers. Instead of having junior employees bothering upper-level management with their; lack of experience, contemporary pop culture references, and a strange fixation on maxxing, senior management will replace these nuisances with a swarm of AI agents. These agents will accomplish all the grunt work previously done by younger employees, freeing up senior management time for boomer pursuits–like scrolling Reddit. The result will be some combination of endless layoffs, social malaise, and massive youth unemployment. Played forward a few years, there’s a fear that joining the middle-class will be unattainable for the youth. Instead, they’ll join the “Permanent Underclass.”
You can either agree with that view or disagree. Maybe AI lives up to its hopes, and maybe it’s a big flop. My view is somewhere in-between, though I expect plenty of disruption to the workplace, as corporates gain efficiencies with AI, offset by whatever new careers are created by the advent of AI. What matters to the theme I’m about to discuss, is what the youth think about their own prospects—which are admittedly quite dim if they remain on the traditional path of a four-year college degree that trains them to be an office drone.
Since my last posting on datacenters, a few hundred billion has been spent on additional capex, and monthly revenues are now clocking in at a few billion—most of which is round-tripped between a handful of players. I remain convinced that the economics of AI datacenters are deeply negative. I am also convinced that the hyperscalers will continue to build them anyway. For that matter, I have witnessed AI continue to advance at an impressive rate. While there are frustrating episodes of hallucination and error, AI has become indispensable to my investment process, and I think that many industries will forever be impacted by AI, mainly for the better.
As a guy who avoids tech, I have missed much of this trade. Yet, I have no regrets. As an inflection investor, I was rather bullish on electricity near the bottom of the cycle in 2021—I actually caught the inflection, only to have Private Equity undercut my upside. However, I’m not one to chase. I never would have bought semis on crazy pricing extrapolations, nor would I have believed that power equipment producers could ever moon-shot as they have—both are highly cyclical industries that are obviously over-earning as they cycle through short-term bottlenecks. I prefer industries that have real and sustainable macro tailwinds, not short-term cyclical ones; where my hopes will be crushed on new supply.
With that in mind, I’d like to talk about a macro trend that seems to have real tailwinds. One that was already accelerating before AI became prominent, but may become a major beneficiary of AI adoption. Interestingly, unlike many of the other popular AI beneficiaries, I think this one trades at a single-digit adjusted cashflow multiple looking out a few years.
There has been much debate about how the employment market will be impacted by AI. On one side, there is the technologist view that tens of millions of jobs will be disintermediated by AI, offset with the futurist view that every productivity advancement throughout human history has simply created more jobs—though very different jobs. I’ve listened to this debate, and can see both arguments. However, I think that the only certainty is that AI will create disruption for workers, and it will happen on a massive scale. Fortunately, I don’t think I have to guess correctly on the outcome of AI upon the job market, I just need to put myself in the direction of travel—which seems to be disruption. Though I do tend to think that AI, autonomous vehicles, and robotics are ultimately going to eliminate many millions of jobs, improve profits for corporates, and strand unskilled workers. It’s Feudalism, applied to the workplace.
So, what’s the play?? I think that technical colleges will see an explosion in new student starts. I don’t know if these are former tech employees, or office workers who got AI’d, or high school graduates who have come to realize that by the time they graduate college in 2030, they’ll likely have $200,000 in college debt and zero job prospects. All I need to believe is that student populations at technical colleges will explode, and the businesses have a whole lot of operating leverage to this fact.
When I turn to social media, I frequently see young people bemoan their college debt, and the hopelessness of being in the “Permanent Underclass” as they chose wrong in their career paths. As a nation, we’ve over-educated office workers, while under-educating in trades. For many years now, more people have been retiring than getting trained in most trades, and given our highly bureaucratic economy, in many of these fields, you cannot simply walk on and do the job—instead, you need all sorts of certifications and specialized training. This all takes time, and has dramatically tightened the labor market, driving wages much higher.
Recently, the explosion in datacenter construction has further tightened this employee market, especially as datacenters are able to outbid most other industries for talent. I’m a big believer in anecdotes, and over the past few months, I’ve heard crazy stories of companies doing insane things to hold onto people—things that smack of desperation. They also show the incredible pricing power of skilled employees in specialized trades; from turbine and engine maintenance, to welding and electrical engineering.
The earning power is rapidly shifting to those who know how to do things outside of an office cubicle, and have the certifications needed to step onto the field. The free market is sorting this all out, and pushing young people into trades. I think this will only accelerate as younger people work with AI and realize that anything involving a computer interface will likely get disintermediated. The world of the future is not going to be driven by 4-year college programs—it will be driven by doing something useful. Interestingly, the bottleneck here is that there are really only two national-scale and publicly traded technical colleges; Lincoln Technical Institute (LINC – USA) and Universal Technical Institute (UTI – USA).
There are a few ways these companies win financially, but most of it is driven by increases in enrollment on somewhat fixed cost structures. From there, the companies likely see lower marketing spend per student—especially if students are organically approaching them. Finally, I think there’s a good deal of pricing power in these programs—especially as the students don’t really pay for their tuition; and this is actually the most interesting part of my thesis. Many corporates are so desperate for workers, that they’ll subsidize education, they’ll sign agreements where students can work off their debt in the first few years of employment, they’ll literally guarantee student debts. Corporates have become desperate for trained workers, and they don’t want the cost of education to be a hindrance. Go to the homepages of LINC and UTI, and look at all the large corporate partnerships. More importantly, the pay scales are quite attractive for first year workers. This is a macro trend that’s already inflecting as the shortage of workers accelerates—AI disruption will only accelerate it further as we need to re-skill millions of future employees.
How do I think about valuations?? I have my internal models where I think these are both trading at mid-single-digit multiples on 2030 earnings, before expenditures on new campuses. Given the extreme operating leverage in these businesses, it’s hard to be precise, especially as a few thousand incremental students will materially change the math, and low current utilization of campuses, when compared to theoretical limits could supercharge earnings well beyond my own models.
Fortunately, I don’t feel like I need to be precise in terms of my financial model here. I see where the trend is going and AI will forever change the nature of work. We need to re-skill people, and I think that the growth in new student applications could be explosive. Besides, given all the uncertainty regarding the path of AI development, where else do you find AI beneficiaries with this much clarity in their future trajectory?? There’s a reason I avoid tech—there are usually easier ways to win on the same trend. Turning to multiples, where else do you find non-cyclical AI beneficiaries that trade so cheaply?? Will these two get AI beneficiary multiples?? If not, why not??
I’ve written this piece as part of a crowd-sourcing project to uncover other, similar ideas. We have large positions in LINC and UTI. I believe in the massive macro tailwinds at play. I know there are other situations like this—situations where I don’t have to guess correctly about components of chip design or understand the newest iteration of an LLM. Rather, I can use my skills to analyze a prosaic industry with big tailwinds, a lot of operating leverage and a big fat moat. I want to think outside the box on AI beneficiaries. My ask to all of you reading this, is to reach out if you have another of these trends that the market has seemingly missed. I want to envision the world of 2030. Who wins?? What are you buying to play AI??
Thanks in advance…
Email me at info@pracap.com
Disclosure: Funds that I control have written puts, and are long shares of LINC and UTI. The Fund’s gross and net performance for all relevant periods is available in our investor letters and tear sheets at pracap.com.
Time To Buy Stuff…