The Investment Excitement Ratio
A new framework combining valuation, concentration and capex; and why today looks like past manias.
Last week I covered what many thought was the best idea at the conference.
This week, I think there’s something more important: two experienced investors arguing we may be in an a bubble and backing it with data. And for premium subscribers, my favourite from the conference, a stock which I already owned.
The presentations made by 2 value investors indicated two separate bubbles. I know both presenters – Cole Smead works with his father Bill Smead who was a guest on my podcast a couple of years ago. Jonathan Boyar, who coincidentally took over his father’s research and asset management business, is an old friend.
Both arguments are compelling, but I think they understate one key point: when valuation, concentration and capex all rise together, the risks don’t just add up, they compound.
That’s what makes the current cycle so interesting. Cole Smead explained that they see AI as just another American Capex Cycle mania, likening this to the 1880s railway mania and crash, the dot.com telecoms bubble and the 2010 shale “drill baby drill” mania. What makes this interesting is not the conclusion, but the framework.
First, a word from my podcast sponsor…
As we move through the current earnings season, the narrative around AI is shifting from experimentation to measurable financial impact. AlphaSense’s recent report benchmarks how leading companies are translating AI investments into operating leverage by providing a data-driven view of execution across core operations and AI-native products.
1880s Railway Mania
In all these manias, he identified an emerging technology – for the railroads, steel rails, the air brake, signal systems via telegraph – and real economic applications: for rail, the time it took to travel from east to west declined from 12-14 days on the Oregon Trail, or 6 months or more by boat, to just 7 days from New York to Sacramento on the Transcontinental Railroad.
Rail infrastructure capex was an astonishing 6% of GDP in the 1880s. He put up the names of the current railroads and their bankrupt predecessors on several slides. There were a lot of railroad companies competing, another bubble characteristic.
Norfolk Southern was preceded by Eire Railroad (bust in 1893); Norfolk & Western Railway (1896); Richmond & Danville Railroad (1892); Ohio Southern Railroad (1893); Lake Erie and Western Railroad (1893); and Philadelphia and Reading Railroad (1893).
Railroads raised a lot of debt but were seen as a low risk sector. They delivered real economic benefit but their equity backers lost out:
Railway Share of S&P
Source: Smead Capital Management from Global Financial Data
1990s Telecoms Bubble
Cole then moved on to the dot.com era which is the bubble poster child. The emerging technology was of course the internet, and he put up a slide of the movie You’ve Got Mail, starring Tom Hanks and Meg Ryan. There was a real economic benefit delivered here too, with US productivity in output per hour growing at just under 10% pa in the 1990-95 period, increasing to 13% pa in the 1995-99 period.
In 1996, the telecom sector was deregulated and there was plenty of investor excitement with the sector taking off, although cellular connections only really accelerated after the bubble burst. But telcos raised $1.6tn of equity and $0.6tn of debt in the 1990s. That may not sound much today but remember there wasn’t a single trillion dollar company then and in real terms the numbers are much higher.
1990s Telecom M&A Activity
Source: Smead Capital Management Estimates
With a large gap between supply and demand, prices fell and as a consequence, the sector has been a poor investment and has collapsed from the 2000 peak. Cole estimates that telecom capex peaked at 1.2% of GDP in 2000, much less than the railroads, but at a level he deems critical.
2010s Shale Mania
The shale revolution similarly had an emerging technology – fracking. Investors became excited, with Aubrey McLendon of Chesapeake being the poster child. I recall the Lehman Energy Conference, held in the first week of September in New York. There was standing room only in the vast hall when he spoke. All credit to the Lehman energy analyst – I asked him about Chesapeake and he told me that it wasn’t for me; I didn’t need to look further.
Again there was an explosion of equity and debt issuance:
Shale Fundraising
Source: Smead Capital Management from Dealogic
And of course there was a huge ramp in competition – Cole’s slide had 64 oil company logos. And again leverage was a feature, because of the sheer size of the spend, especially relative to the size of some of the players. Here is the cash burn:
Shale Cash Burn
Source: Deloitte
Smead estimates that US energy infrastructure capex peaked at 0.9% of GDP in 2014.
He then turned to the current cycle and pointed out that Google’s search monopoly, once thought invincible, had seen its market share slide from 98% to 92%:
Search Shares
Source: Smead Capital Management, Various
That’s quite a drop in just over a year, although the obvious question is why it hasn’t been faster and greater. Part may be the Google AI response, but that has been more recent.
Interestingly, I asked a senior Google technical person about this and they laughed and said “search will never go down”. That laugh – I felt stupid.
Cole pointed to studies showing that AI hadn’t had a real effect on productivity as yet, although there is a lot of investor excitement and a huge amount of fund raising. Here is AI as a proportion of VC investment in the US:
AI vs Total VC Investment in US (2024 on Carta)
Source: Carta
And here is global AI investment:
Global AI VC Investment
Source: dealroom.co
Cole went on to highlight the amount of investment by Meta in particular and the total capex across the hyperscalers. He concluded by comparing the capex intensity today to previous cycles:
Capex as % EBITDA
Source: Smead Capital Management
And pointing out that the capex today is equivalent to the previous excesses:
Mania Capex to GDP
Source: Smead Capital Management
Just because the telecoms peak was 1.2% and the shale peak was less, doesn’t necessarily mean that spending above 1% of GDP is a trigger for anything. And the AI investments are global in nature, while capex in the previous manias was more domestic in nature. But it’s clear that this should be a concern for investors, as I have written here before.
Cole then gave a clever summary, introducing his Investment Excitement Ratio. This incorporates valuation, using the Buffett indicator (Russell 2000 market cap to GDP ratio); the concentration (sector vs S&P500) and those capex to GDP data:
Smead Investment Excitement Ratio
Source: Smead Capital Management
They take the Buffett indicator and multiply it by market concentration then divide this by the capex. That formulation almost guarantees today looks extreme.
It’s a clever idea—and I wish I’d thought of it.
But I think the construction actually mis-states the risk.
In reality, these variables reinforce each other. When capex, valuation and concentration all rise together, risk compounds, not offsets. I would multiply them all together to derive an alternative measure:
Alternative Investment Excitement Ratio
Source: Behind the Balance Sheet from Smead Capital Management Data
I think it’s a clever and fun idea, but I think my calculation is a better representation – in investment terms, this is a little worse than the dot.com era – more money being spent and MUCH shorter life assets; it’s much greater than the shale gas mania; but it’s not nearly as extreme, at least not yet, as the railways bubble. But the railroads were very long-lived assets. Data centres, not so much.
Cole went on to look at the aftermaths in share price terms. In the second half of the 19th century, the railroads were the market, but today they are irrelevant, having underperformed over that 100+ years:
Railroads as % of US Market
Source: Smead Capital Management, Birinyi Associates
Similarly, the telecom sector has massively underperformed the S&P 500:
Telecom Total Returns
Source: Smead Capital Management
As have the oil majors:
Oil Major Returns
Source: Smead Capital Management
Cole thinks big tech is yesterday’s story and points out that 3% of the S&P is in energy today. They have 24% of one fund in energy and 43% in another. Industry investment has collapsed and returns are increasing.
Jonathan Boyar had a similar perspective on a bubble today, but he fashioned it as overpaying for quality and likened this period to the Nifty Fiifty (late 60s/early 70s) and the quality bubble (late 90s/early 00s) and asked “Is this deja u all over again?”
Boyar explained that the Nifty Fifty was a group of stocks that were considered so dominant and so reliably profitable that they were one-decision stocks – bought at any price, hold forever, never sell.
By the 1972 peak, the Nifty Fifty traded at 42x earnings or more than twice the S&P500 which traded at 19x. Companies like Polaroid, Avon, McDonald’s and Xerox were seen as impervious to economic cycles and their unique characteristics justified any multiple. Of those four, only McDonalds prospers today – Polaroid went bust; Avon was taken over then resold; and Xerox is probably one of the worst examples of capital allocation in the S&P500 in the last 20 years, as it has consistently bought back stock at ever decreasing prices. I wonder how long it will take to go bust.
73-74 Declines
Source: Boyar Value Group
That 1973-74 selloff highlighted in the table above took just 18 months and the average of the Nifty Fifty fell 60%. It took on average 10 years for the group to recover their nominal price highs. Given the inflation in that decade, this was still a disastrous result in real terms.
Jonathan then turned to the inflation of so-called quality stocks in the dot.com era. Here is a selection of stocks then and their subsequent 5-year return:
Quality Bubble Deflated
Source: Boyar Value Group
He thinks there is a parallel today and I agree. It’s déjà vu all over again. Again. He didn’t pick the tech stocks, but Costco (47x current eps), Walmart (42x), Cintas (38x), GE Aerospace (42x), and Ecolab (32x). I haven’t looked at Ecolab since I met Doug Baker Jr many years ago, but I thought it was a great business and I am not sure that multiple is so scary. GE Aerospace is clearly an expensive stock, but it too is an unusual business with a substantial moat.
I totally agree on the other three stocks – Cintas has significant operational gearing to the incremental customer and must have a long growth runway but it has always been rated more highly than I thought warranted.
I have come across a few stocks like this in the US, which I term “over-loved” – Danaher was similar where I could never get remotely comfortable with the multiple. Nothing wrong with the companies, just that they were extremely popular with investors and I didn’t like the risk-reward. Stupid of me, perhaps, as these stocks mainly did well.
Costco and Walmart puzzle me. I debated this with Christopher Tsai on my podcast because he refused to sell Costco in spite of it being on 55x eps as I recall – he said that it’s always wrong to sell a terrific compounder. I disagree.
Similarly, I wonder to what degree Walmart can grow above GDP on a $750bhn sales base. I understand that it’s recovering in the online area and that incremental revenue from advertising is almost all profit. And that it’s well positioned for trading down.
But I wouldn’t pay 42x for the privilege.
So both Smead and Boyar were bearish on a handful of expensive names in the US for quite different reasons. Next week, I shall cover Alberto Saporta’s views which were much more bearish.
Premium subscribers can read on for a high-quality, capital-light business (which I own) that is being valued as if it were a cyclical and why that disconnect may not last.
Finally, an ad from this week’s sponsor:
Investing in Ukraine: how to gain exposure before global capital returns
Everyone expects a “Marshall Plan II” for Ukraine’s reconstruction – but few investors appreciate how transformative it could be.
Ukraine’s economy currently generates around USD 400bn annually, while reconstruction spending could exceed USD 800bn. In a successful scenario, the economy could grow 3-4x over the next 20 years.
Yet most investors remain on the sidelines due to travel restrictions, banking hurdles, and political uncertainty.
“Investing in Ukraine” explores a simple question: what if the best opportunities emerge before reconstruction – while most investors are still waiting?
Held on 23 April 2026, this online conference brings together investors, analysts, executives, and journalists with direct, on-the-ground experience to examine practical ways to gain exposure today.
A rare chance to understand – and potentially act on – a market the world is watching.
