David Hay — Haymaker Daily: Steroids On Top Of Steroids — the week's fiscal thread resolved into two stimulants propping up growth and profits: 6%-of-GDP deficits (worth ~2% of artificial GDP even at a 0.5 multiplier, and dollar-for-dollar to corporate profits under the Levy-Kalecki Profit Equation) and $700bn+ a year of AI capex whose hyperscaler debt raises are a "reverse crowding out" driving government borrowing costs up; per a Columbia Business School chart relayed by Luke Gromen, no prior efficiency-spending era has been this large a share of GDP — and both doses are unlikely to be maintained (Aug 27 2026)

The week's closing instalment, and the one that turns the fiscal argument into an earnings argument. "In this week's Dailies, we've been examining the feedback loops from the federal government's inability to control its immense spending and the related record peacetime deficits… Yet, there are also substantial near-term benefits that make it hard to kick the habit." Benefit one is growth: "when federal red ink is running at four percent above normal — say, 6% of GDP deficits vs. a more typical 2% — this provides a significant boost to the economy… even assuming only a 0.5 'multiplier effect' this would still amount to around 2% of artificial economic growth." Benefit two is the one investors underweight: "corporate profits also receive a tremendous up-lift from emergency levels of U.S. government largesse. In theory, there is a one-to-one relationship, per the widely accepted Levy-Kalecki Profit Equation. Accordingly, a $1 trillion jump in the federal deficit inflates corporate profits by a similar amount" — qualified honestly by "negative offsets such as elevating trading deficits, as well as raising interest rates as surging government borrowing 'crowds out' the private sector." The second stimulant is AI: "another enormous economic accelerant currently is AI-related spending. This pace, too, is highly unlikely to be maintained longer term. It is also having the bizarre effect of driving up the government's borrowing costs due to the deluge of hyperscaler debt raises to fund their $700 billion plus annual expenditures on the great data center build-out. In other words, this is a reverse 'crowding out'." (Nvidia's Q2 results, "reported yesterday, indicate this AI spending boom will continue to run white hot.") The historical frame comes from outside: "per the following graphic from the Columbia Business School, as relayed by Luke Gromen in his August 14th newsletter, prior eras of excessive spending to finance efficiency/technological breakthroughs have never represented such a large share of GDP. This raises the reasonable question of what happens to the economy when these double doses of amphetamines inevitably fade." The bull case is stated before it is doubted — "the productivity boost from AI will be so mammoth that these outlays will easily pay for themselves. This may well turn out to be the case, but there is a substantial risk that returns on both the federal and AI spending will turn out to be disappointing" — with the fiscal half judged the weaker: "the odds are that much higher excessive government spending will turn out to be a big disappointment… the facts are undeniable that America's economic growth dramatically decelerated over the last 25 years while deficit spending has exploded. The long-term bond market is clearly beginning to sniff out the unsustainability of this situation."

One-line take: the payoff note of the week's fiscal series, and the one with the most direct consequence for equity valuation — because it argues that reported corporate profits are themselves partly a fiscal artifact. The mechanism is named rather than gestured at: the Levy-Kalecki Profit Equation, under which aggregate corporate profit is an identity fed by, among other terms, the government deficit — "in theory, there is a one-to-one relationship… a $1 trillion jump in the federal deficit inflates corporate profits by a similar amount." Hay does not overclaim it: "in practice, it's more complicated," with "negative offsets" from trade deficits and from the higher rates that surging government borrowing produces. But the implication is unavoidable and is the bridge to the house's valuation work — if a deficit running 4% of GDP above normal is inflating both the numerator of the economy (roughly "2% of artificial economic growth" even at a conservative 0.5 multiplier) and the earnings line, then normalised margins are lower than reported margins, which is exactly the claim made about the S&P in the Aug-17 Update ("on a normalized profit-margin basis, the S&P 500 is far more overpriced than it superficially appears"). The second stimulant contributes an unusually clean piece of analysis: AI capex at "$700 billion plus annual expenditures" is not merely large, it is being financed with hyperscaler debt, so private borrowing is now competing with the Treasury for the same pool of savings — "the bizarre effect of driving up the government's borrowing costs… a reverse 'crowding out'." Conventional crowding-out has the state pushing out the private sector; here the causation runs the other way, and it links this note directly to Aug-25's defended long yield and Aug-26's issuance surge — three separate sources of upward pressure on the same rate. The historical check is borrowed rather than asserted: a Columbia Business School graphic relayed by Luke Gromen's August 14th newsletter showing that "prior eras of excessive spending to finance efficiency/technological breakthroughs have never represented such a large share of GDP" — i.e. there is no precedent at this scale, which cuts against both the bulls and the bears and is presented as a question rather than an answer. The verdict is properly asymmetric between the two doses. On AI, Hay concedes the bull case may win — "this may well turn out to be the case." On the fiscal dose he does not: "the odds are that much higher excessive government spending will turn out to be a big disappointment," supported by the 25-year record in which "economic growth dramatically decelerated… while deficit spending has exploded." Closing line, tying the week together: "the long-term bond market is clearly beginning to sniff out the unsustainability of this situation." One name appears: NVDA, in a single parenthesis as evidence that the AI capex dose is still running — no view, target or recommendation is attached, and it is rowed Neutral on that basis.

1. Stocks & names mentioned

Ticker	Name	Research	View	What he said	At
NVDA	NVIDIA	QT · SA · STK · FA	Neutral	Passing mention — cited as the read-out on AI capex, not as a stock call. The entire reference is one parenthesis: "(Nvidia's second-quarter results reported yesterday indicate this AI spending boom will continue to run white hot.)" Its role in the argument is as a coincident indicator of the second stimulant — the note's case is that "AI-related spending" is "another enormous economic accelerant" whose "$700 billion plus annual expenditures on the great data center build-out" are being funded by "a deluge of hyperscaler debt raises," producing "a reverse 'crowding out'" that drives up the government's own borrowing costs. Nvidia's quarter is the evidence that the spending has not yet slowed. The surrounding view of the capex itself is two-sided but ultimately cautionary — "this pace, too, is highly unlikely to be maintained longer term," against the bullish rebuttal that "the productivity boost from AI will be so mammoth that these outlays will easily pay for themselves," which "may well turn out to be the case," though "there is a substantial risk that returns on both the federal and AI spending will turn out to be disappointing." No valuation, target, rating or recommendation is given on the shares.	read ↗

This is a macro Daily: NVDA is the only name in the post, and it appears in a single parenthesis as the datapoint confirming AI capex is still running hot — no stance is taken on the shares, so it is rowed Neutral as a passing mention. Not rowed — deliberately: the hyperscalers are referenced only as a category of borrower with no company named; Columbia Business School and Luke Gromen's Forest for the Trees newsletter are the chart's source and relay; the Levy-Kalecki Profit Equation is an economic identity, not a security. Research: QT Qualtrim · SA Seeking Alpha · STK Stock Analysis.

2. Talking points

The week's thread, and why the habit is hard to break

"In this week's Dailies, we've been examining the feedback loops from the federal government's inability to control its immense spending and the related record peacetime deficits."
The consensus is granted: "it's reasonable to conclude most Americans realize this is a totally unsustainable situation."
The twist that makes the note: "yet, there are also substantial near-term benefits that make it hard to kick the habit" — the addiction framing is deliberate and returns at the end as "double doses of amphetamines."

Dose one, part A — deficits as artificial GDP

The comparison is to a normal deficit, not to zero: "when federal red ink is running at four percent above normal — say, 6% of GDP deficits vs. a more typical 2% — this provides a significant boost to the economy."
The magnitude is deliberately conservative: "how much is the subject of intense debate. Yet even assuming only a 0.5 'multiplier effect' this would still amount to around 2% of artificial economic growth."
Note what that implies against yesterday's scorecard: with GDP averaging "roughly 2%," the entire measured growth rate is the size of the estimated fiscal contribution.

Dose one, part B — the Levy-Kalecki channel into corporate profits

"Corporate profits also receive a tremendous up-lift from emergency levels of U.S. government largesse."
The mechanism is named: "in theory, there is a one-to-one relationship, per the widely accepted Levy-Kalecki Profit Equation. Accordingly, a $1 trillion jump in the federal deficit inflates corporate profits by a similar amount."
The qualification is given immediately and honestly: "in practice, it's more complicated… there are negative offsets such as elevating trading deficits, as well as raising interest rates as surging government borrowing 'crowds out' the private sector."
The consequence for valuation is the reason this matters: profits inflated by a deficit are not a repeatable margin, which is the analytical basis for the archive's normalised-margin claim in the Aug-17 Update.

Dose two — AI capex, and the size of it

"Another enormous economic accelerant currently is AI-related spending."
Sustainability is doubted in the same breath as the scale is acknowledged: "this pace, too, is highly unlikely to be maintained longer term."
The number: "$700 billion plus annual expenditures on the great data center build-out."
The coincident read-out: "(Nvidia's second-quarter results reported yesterday indicate this AI spending boom will continue to run white hot.)"

Reverse crowding out — the private sector pushing up the government's cost of money

"It is also having the bizarre effect of driving up the government's borrowing costs due to the deluge of hyperscaler debt raises to fund their build-out. In other words, this is a reverse 'crowding out'."
The conventional version runs the other way — government borrowing displacing private investment — and Hay has just invoked it as one of the Levy-Kalecki offsets, so the inversion is deliberate.
Read with the rest of the week, three forces now push on the same long yield: the deficit itself, the 8.4% issuance surge of Aug-26, and now hyperscaler debt competing for the same savings — against a Treasury trying to hold that yield down (Aug-25).

The historical check — no precedent at this share of GDP

Sourced twice over, which is worth noting: "per the following graphic from the Columbia Business School, as relayed by Luke Gromen in his August 14th newsletter."
The finding: "prior eras of excessive spending to finance efficiency/technological breakthroughs have never represented such a large share of GDP."
The question it raises is left open rather than answered: "what happens to the economy when these double doses of amphetamines inevitably fade?"

The bull case, stated before it is doubted

"A bullish rebuttal to these concerns is that the productivity boost from AI will be so mammoth that these outlays will easily pay for themselves."
It is genuinely conceded: "this may well turn out to be the case."
But the risk is symmetric across both doses: "there is a substantial risk that returns on both the federal and AI spending will turn out to be disappointing."

Why the fiscal dose is judged the weaker of the two

The asymmetry is explicit: "of course, the odds are that much higher excessive government spending will turn out to be a big disappointment."
The evidence is a 25-year natural experiment: "the facts are undeniable that America's economic growth dramatically decelerated over the last 25 years while deficit spending has exploded."
The closing line ties the week's three Dailies together: "the long-term bond market is clearly beginning to sniff out the unsustainability of this situation."

Housekeeping

One exhibit referenced in the original and omitted from the saved text: the Columbia Business School chart of efficiency/technology-driven spending eras as a share of GDP, credited as "Secondary Source: Luke Gromen, Forest for the Trees."
NVDA is the only security named, in a single parenthesis, with no view attached; the hyperscalers are referenced as a category with no company named.
Signed "The Haymaker Team."

3. In plain English

This note answers a question the rest of the week's writing left hanging: if the government's finances are as bad as Hay says, why does the economy feel fine and why are corporate profits so strong? His answer is that two large, temporary stimulants are propping both up — and neither can run forever.

The first is the deficit itself. The government is spending about 6% of the economy more than it collects, against a more normal 2%. That extra 4% doesn't vanish — it is paid to contractors, benefit recipients, employees and suppliers, who spend it again. Economists argue endlessly about how much extra activity each dollar creates, so Hay deliberately uses a low-ball estimate: even at fifty cents of activity per dollar, the excess deficit is generating roughly 2% of extra growth. Set that against yesterday's note, where measured growth has been running about 2%, and the arithmetic is uncomfortable: something close to all of the growth may be borrowed.

The second effect is the one investors tend to miss, and it is the reason this note matters for share prices. There is a long-established accounting relationship in economics called the Levy-Kalecki profit equation. Its logic is simple once stated: company profits in aggregate come from what other sectors spend into the company sector. When the government spends far more than it taxes, that surplus lands, eventually, on corporate income statements. In theory the relationship is one-for-one — a trillion dollars of extra deficit shows up as roughly a trillion dollars of extra corporate profit. Hay is careful to say reality is messier: money spent on imports leaks abroad, and heavy government borrowing pushes interest rates up, which costs companies money.

Still, the conclusion follows. If a chunk of reported corporate profit exists because the government is running an emergency-sized deficit, then the profit margins the stock market is being valued on are flattered by something temporary. That is the analytical backbone of Haymaker's standing claim that the S&P is more expensive than its headline multiple suggests: the price may look reasonable, but only against earnings that are partly fiscal.

The second stimulant is the AI build-out — more than $700 billion a year on data centres. This one comes with a twist Hay calls "reverse crowding out." Normally, economists worry that when a government borrows heavily it soaks up the available savings and pushes interest rates up, making it harder for companies to borrow — the government crowds out the private sector. Right now the causation is running backwards: the big technology companies are issuing so much debt to fund data centres that they are helping push borrowing costs up for the government. Nvidia's latest quarter, reported the day before this note, shows the spending is still accelerating.

Put the week together and three separate forces are pressing the same long-term interest rate higher — the deficit, the surge in new government bond issuance, and now the hyperscalers' borrowing — while the Treasury is trying to hold that rate down. That is the fight described in Tuesday's and Wednesday's notes, seen from the supply side.

The historical check comes from a Columbia Business School chart that Hay found in Luke Gromen's newsletter: never before has this kind of spending — money poured into a technological or efficiency breakthrough — been such a large share of the economy. There is no precedent at this scale, which is honest but double-edged: it means neither the optimists nor the pessimists have a comparable episode to point at.

Hay then splits his verdict rather than dismissing both doses together. On AI, he grants that the bulls may be right — the productivity gains could genuinely be large enough to pay for the spending. On government spending he is much less generous, and his evidence is a 25-year record in which deficits exploded while economic growth slowed markedly. Spending that much and growing that little is not an argument for spending more.

The closing thought is where the week lands: the bond market is starting to work all of this out. That is why the arguments about growth, profits and deficits keep resolving into the same conclusion about the long-term interest rate — and why what happens to it matters more than almost anything else in a portfolio.

Macro viewpoint with a single passing security reference (NVDA, cited as a read-out on AI capex — no stance taken). Summary derived from the paid Haymaker Daily (text in transcript.txt) for personal study. The chart referenced in the original (Columbia Business School, via Luke Gromen's Forest for the Trees) is omitted. Not investment advice. © Haymaker / David Hay for source material.