Bi-Weekly Geopolitical Report – The AI Buildout: Profits, Productivity, and the Risk of Overcapacity (10/12/2026)

Investors want to know who will be the ultimate winners from the AI build out. Confluence Associate Market Strategist and Certified Business Economist Thomas Wash joins Phil Adler to discuss lessons from the internet bubble, which might give us some clues.

What is Confluence Podcasts?

Podcasts from Confluence Investment Management LLC, featuring the periodic Confluence of Ideas series, two bi-weekly series: the Asset Allocation Bi-Weekly and the Bi-Weekly Geopolitical Report (new episodes posted on alternating Mondays), and a new monthly Q&A format called the Confluence Mailbag.

Phil Adler: 00:04

Welcome to the Confluence Investment Management Bi-Weekly Geopolitical Report for 10/12/2026. I'm Phil Adler. Investors wanna know who will be the ultimate winners from the AI build out. Confluence Associate Market Strategist and Certified Business Economist Thomas Wash joins us today to discuss lessons from the dotcom bubble, which might give us some clues. Thomas, the potential dangers of AI are all over the news these days.

Phil Adler: 00:32

But at least for now, it appears the AI build out will continue unabated. Anticipating that, your report this week focuses on the possible investment winners and losers. The title is The AI Buildout, Profits, Productivity, and the Risk of Overcapacity. The most positive outlook anticipates companies and savvy investors getting rich as profits eventually soar, while consumers benefit as well from increased convenience and efficiency and a lower cost of living. Now, one of your main findings is that these benefits for AI companies and consumers, divided broadly by profits and productivity, may not be mutually achievable.

Phil Adler: 01:17

So my first question is, why not?

Thomas Wash: 01:20

Well, because those two outcomes tend to pull against each other. When a transformative technology arrives, there's usually a lot of hype and expectation that everyone wins. Firm gets higher profits and consumers get lower prices. In practice, the gains rarely get split that neatly. If a company can use technology in a way its competitors can't, it captures the gains as profit.

Thomas Wash: 01:41

That first mover advantage gives it pricing power and it's essentially how NVIDIA has become so profitable. Its competitors haven't been able to match its chips. Now, if a company can't hold that edge, though the productivity gains are shared, output rises, but there's no pricing power, and the benefit flows to consumers. In my view, how that trade off plays out will be a key factor in determining the winners of the AI build out.

Phil Adler: 02:05

Well, looking back, how did this tug of war play out during the dot com bubble? Broadly speaking, did profits or productivity come out on top?

Thomas Wash: 02:15

Well, I I would say profits won early, but productivity and consumers won in the end. In the early days, there were substantial profits to be made, and many companies could charge high prices and enjoy strong margins. As competition intensified, the advantage gradually shifted toward consumers. New interests eroded one another's profits, which led to a race to the bottom on prices, and in some cases, overcapacities as weaker competitors failed. Telecom companies were amongst the biggest losers, both heavy builders and their suppliers, because they built too much capacity.

Thomas Wash: 02:47

The ultimate winners were the users of that capacity, who could buy cheap bandwidth to build their platforms.

Phil Adler: 02:53

So I'm I'm wondering, what can we glean from the .com era about which companies might be ultimate winners in the current environment?

Thomas Wash: 03:03

Mainly, that building the infrastructure and profiting from it are two different things. In the .com era, many of the builders gained little from their investment. That was especially true of telecom companies and the makers of routers and switches, who were supplying companies that would eventually need that infrastructure. The biggest beneficiaries were the users who built businesses on infrastructure they got for well below what it cost to build.

Phil Adler: 03:27

So the winners of the .com era, were the firms able to step in at a later date and take advantage of the foundation created by the overbuilding telecom companies?

Thomas Wash: 03:39

Largely, yes. Many of today's leading companies such as Netflix, Amazon, and Google built enormous businesses on the fiber laid during that period. That infrastructure became the backbone of streaming, cloud computing, and web hosting, the very services that make these companies today.

Phil Adler: 03:55

Well, there's worry that right now that firms investing in AI data centers and pursuing AI dominance are overbuilding and overspending, similar to what happened in the nineties. But maybe history doesn't repeat itself exactly. Are there any key differences between today's environment and the era of the telecom and dot com bubble, might affect today's competition between profits and productivity?

Thomas Wash: 04:21

Yes. And I would say the main difference is the rise of China. Unlike in the dot com era, much of the infrastructure needed today doesn't have to be built in The US to be effective. Chinese companies can compete with The US from anywhere cloud services are available. That has made it easier for Chinese AI lab to develop and compete since they no longer depend on US infrastructure.

Thomas Wash: 04:40

It has also shifted the productivity trade off. Chinese AI models are significantly cheaper than their US counterparts, which has benefited users.

Phil Adler: 04:48

And one other factor you mentioned in your report was the federal government and federal government policies today. Exactly how are federal government policies boosting prospects for AI firms?

Thomas Wash: 05:02

I would say mainly by treating AI as a national security priority. You know, the White House has made AI central to its national security agenda, which has given these companies far greater support in building the necessary infrastructure. Beyond tax incentives, the administration has also used trade pressure as part of a broader strategy to secure an advantage over foreign competitors.

Phil Adler: 05:22

And carrying this forward, how might that hurt consumers?

Thomas Wash: 05:26

The government's involvement suggests that it may intervene to make sure these companies keep their advantage, at least domestically. That could mean regulation that limits consumers' ability to use Chinese competitors, which are typically cheaper and more flexible. Measures like that could hurt consumers by forcing them to pay higher prices to access AI models.

Phil Adler: 05:45

Thomas, demand for AI right now does not seem to be keeping pace overall with expansion, at least at least not at this point. Should we be concerned, or does demand normally take a while to catch up?

Thomas Wash: 05:59

I think I'd frame it a little bit differently. Demand for AI itself is clearly strong and growing. Where we're falling short, I think, is demand for the more sophisticated ability, which hasn't kept pace with the capabilities themselves. Most people still use these models in fairly basic ways, such as answering questions or explaining how to do things. Meanwhile, the models are increasingly built for complex tasks involving multi step, second order reasoning.

Thomas Wash: 06:22

Many see that kind of reasoning as a step toward artificial general intelligence where models can think more like humans.

Phil Adler: 06:29

Thomas, you mentioned that the US government is treating AI as a national security priority. So does this mean that the federal government is seeking to ensure success by US AI firms as a way to stay ahead of China militarily?

Thomas Wash: 06:45

I think The US views AI as the future of warfare, so it's willing to do whatever takes to keep its advantage over rivals, including China.

Phil Adler: 06:53

Is there a side between profits and productivity which appears to be closer to reality right now? Corporate profits or the elevated productivity that can decrease inflation and benefit consumers?

Thomas Wash: 07:06

Right now, I'd say the picture is mixed. Some hardware companies are earning high profits, but that's largely due to massive AI infrastructure build out. That build out has also driven up prices elsewhere, including electricity and chips, which has hurt consumers. Among the AI labs, however, rising competition from China has forced domestic providers such as Anthropic and OpenAI to lower their prices. That said, it's too early in this tug of war to really decide who is winning.

Phil Adler: 07:31

Well, history tells us that among companies which do succeed during times like these, times that experience huge buildouts, success is not universal. The rising tide may not lift all boats. Which sectors or groups of companies may have the most AI upside right now long term?

Thomas Wash: 07:51

We think the biggest long term beneficiaries are likely to be sectors that haven't integrated AI into their operations. So, you know, the health care, the retail trades, construction, they could all benefit from these models. And we think those areas are where investors should look for upside.

Phil Adler: 08:09

Thank you, Thomas. The title of this week's report is the Productivity, and the Risk of Overcapacity, and you can find a link to the written report dated 10/12/2026 on the Confluence webpage, confluenceinvestment.com. Our discussion today is based upon sources and data believed to be accurate and reliable. Opinions and forward looking statements expressed are subject to change without notice. This information does not constitute a solicitation or an offer to buy or sell any security.

Phil Adler: 08:38

Our audio engineer is Dane Stole. I'm Phil Adler.