Andy Constan on the SpaceX IPO, AI CapEx, and the End of the Buyback Tailwind
The market's $600–$700 billion swing from stock buybacks to new equity issuance is the biggest structural shift in share supply in years — and AI CapEx is the sole reason it's happening.
Excess Returns
Andy Constan on the SpaceX IPO, AI CapEx, and the End of the Buyback Tailwind
The market's $600–$700 billion swing from stock buybacks to new equity issuance is the biggest structural shift in share supply in years — and AI CapEx is the sole reason it's happening.
TL;DR
Andy Constan breaks down three forces reshaping markets: the SpaceX IPO as a bellwether for a new wave of equity issuance, AI CapEx creating a $600–$700 billion shift from share buybacks to net new supply [1] — Andy Constan "$600–$700B shift in share supply: The market has shifted from a ~$600–$700 billion net share reduction annually to effectively zero or net …" 24:00 , and consumer dissaving propping up an otherwise middling economy [2] — Andy Constan "Consumer dissaving for 3–4 years: US consumers have been dissaving — selling assets to fund consumption above income — for roughly 3 to 4 y…" 45:40 . He argues falling oil prices won't kill inflation — core will just rise to replace headline [3] — Andy Constan "Tariffs expire July 24 unless renewed: Current tariffs are set to disappear automatically by July 24th unless actively renewed or replaced …" 58:00 . The single most useful takeaway: the AI investment thesis only works if tokens create disinflationary productivity growth without displacing workers [4] — Andy Constan "AI tokens must either create more output for the same cost — disinflationary growth — or they lower costs by firing people. If workers are …" 30:08 .
Andy Constan's third First Principles episode covers the reopening IPO window (SpaceX as case study), AI CapEx's structural impact on share supply, the consumer economy, inflation dynamics, and what Kevin Warsh's Fed leadership means for markets.
-
Launch announcement for the First Principles podcast with Andy Constan, previewing the episode's topics: SpaceX IPO, AI and the economy, and market structure.
-
Andy explains the fundamental purpose of markets — connecting capital-seekers with capital-providers — and why all IPO participants (issuers, banks, regulators, investors) are aligned toward making deals succeed. [1] — Andy Constan "Markets exist to connect those who need money with those who have it. The IPO is the most critical expression of that purpose — it transfor…" 03:58
-
Andy evaluates the SpaceX IPO: $85B float on a $2T cap, priced at $135, opened at $150, traded into the $170s — the biggest IPO in history and a well-executed deal. [1] — Andy Constan "A good IPO isn't about getting the perfect price — it's about satisfying every party in the room. SpaceX priced at $135, opened at $150, an…" 05:45
-
Mid-episode advertisements for Indeed's sponsored job postings and Canva's design platform.
-
Andy details how buybacks reduce share supply while issuance does the opposite, and quantifies the 2023–24 baseline of ~2% of GDP in net share reduction before the current reversal. [1] — Andy Constan "Net share supply: ~2% of GDP reduction (2023–24): During 2023 and 2024, share repurchases ran at roughly 2% of GDP, representing a net redu…" 21:15
-
The market has swung $600–$700B annually from buybacks to net issuance. Google canceled repurchases and issued $80B; Meta is reducing buybacks; the cause is entirely AI CapEx. [1] — Andy Constan "The market has flipped. For years, buybacks reduced share supply by roughly 2% of GDP annually, acting as a structural tailwind. Now Google…" 19:56
-
The AI investment thesis requires tokens to create disinflationary growth — more output for the same cost. If AI displaces workers instead, aggregate demand collapses and the math breaks. [1] — Andy Constan "AI tokens must either create more output for the same cost — disinflationary growth — or they lower costs by firing people. If workers are …" 30:08
-
Wait — this section actually continues Andy's analysis. Checking timestamps: the Columbia ad appears at the transcript around the 27-minute mark (≈1652s). Correcting: this chapter covers Andy's historical productivity analysis and the messy AI transition argument.
-
The economy is doing reasonably well. CapEx is one pillar, but most Americans — teachers, doctors, construction workers — are untouched by it. No one is being fired by AI yet.
-
Real wages lag growth, but consumers keep spending via asset dissaving — selling appreciated stocks, homes, and crypto. This has run for 3–4 years, supported by bubble-level asset prices. [1] — Andy Constan "Real wages aren't keeping up with growth, but consumers keep spending. The reason: extreme wealth effects. Rising stocks, homes, and crypto…" 42:36
-
Falling oil prices lower headline inflation but push core higher as spending shifts. The US has been inflationary for 5–6 years; wars like Ukraine and the Middle East barely move the macro needle. [1] — Andy Constan "Oil falling looks like good inflation news. It isn't. Lower oil frees up consumer spending that flows into other goods, pushing core inflat…" 51:20
-
Falling oil prices lower headline inflation but push core higher as spending shifts. The US has been inflationary for 5–6 years; wars like Ukraine and the Middle East barely move the macro needle. [1] — Andy Constan "Oil falling looks like good inflation news. It isn't. Lower oil frees up consumer spending that flows into other goods, pushing core inflat…" 51:20
-
Andy's summer watchlist: Warsh builds credibility without acting, tariffs quietly expire July 24, AI CapEx flows with no ROI clarity, and equity issuance tests market absorption capacity. [1] — Andy Constan "Four things matter this summer: the new Fed chair builds credibility without acting, tariffs quietly expire on July 24 unless renewed, AI C…" 56:30
-
Andy argues Fed balance sheet mismanagement is a root cause of persistent inflation. Warsh wants reform — lower short rates plus QT — but faces deep institutional resistance inside the Fed. [1] — Andy Constan "The Fed has mismanaged its balance sheet from the beginning, and that's a primary reason inflation persists. Warsh wants to lower short rat…" 59:55
-
Andy thanks the hosts, encourages listeners to subscribe on all platforms, and closes with a 'Go Knicks' sign-off.
- IPO (Initial Public Offering)
- The first time a private company sells shares to the public on a stock exchange, allowing it to raise equity capital from a broad investor base.
- Float
- The portion of a company's total shares that are freely available for public trading; at SpaceX's IPO, the float was approximately $85 billion of a $2 trillion market cap.
- Share repurchase (buyback)
- A company's purchase of its own outstanding shares from the market, reducing total share count and returning capital to shareholders; contrasted here with new equity issuance.
- RSU (Restricted Stock Unit)
- A form of employee compensation granting company shares that vest over time; rising RSU awards increase share supply when employees sell vested units.
- CapEx (Capital Expenditure)
- Money spent by a company on acquiring or upgrading physical assets such as data centers, chips, and cables; used throughout to describe AI infrastructure investment.
- Hyperscalers
- Large cloud computing companies — primarily Amazon, Microsoft, Google, and Meta — that operate massive data centers and are the primary buyers of AI compute infrastructure.
- Disinflationary growth
- Economic expansion in which increased output is produced at the same or lower cost, reducing upward price pressure; the favorable macro scenario for AI productivity investment.
- Dissaving
- Spending more than one earns by drawing down savings or selling assets; Andy Constan uses this to explain how US consumers maintain spending despite flat real wages.
- QT (Quantitative Tightening)
- The Federal Reserve's process of shrinking its balance sheet by allowing bonds to mature without reinvestment, reducing the money supply; contrasted with QE (quantitative easing).
- Wealth effect
- The tendency of consumers to spend more when their asset values (stocks, homes) rise, even without an increase in earned income; cited as a key support for US consumption.
- Reconciliation bill
- A congressional budget procedure allowing certain legislation to pass the Senate with a simple majority; mentioned in the context of potential fiscal stimulus.
- Secondary offering
- A sale of new or existing shares by a public company after its IPO; Andy Constan cites Google's $80 billion secondary as an example of the buyback-to-issuance shift.
- SPAC (Special Purpose Acquisition Company)
- A shell company raised through an IPO with the sole purpose of merging with a private company to take it public; cited as one mechanism of equity issuance.
- Dottage
- A state of senile decay or feeble-mindedness associated with old age; Andy Constan used it humorously to say he will be too old for AI displacement to matter personally.
- Animal spirits
- A term from Keynesian economics describing the confidence and optimism (or pessimism) that drives economic decision-making beyond strict rational analysis; used to explain bubble persistence.
- Underwriter
- An investment bank that manages an IPO process, helping set the price, build the order book, and distribute shares to investors in exchange for a fee.
- Aftermarket
- Trading that occurs in a stock after its initial IPO pricing; a strong aftermarket (trading above issuance price) is considered a sign of a successful deal.
- Core vs. headline inflation
- Headline inflation measures all consumer prices including food and energy; core inflation excludes those volatile categories. Andy Constan argues falling oil moves headline down while pushing core up.
Chapter 1 · 00:00
Intro and key themes
Launch announcement for the First Principles podcast with Andy Constan, previewing the episode's topics: SpaceX IPO, AI and the economy, and market structure.
Markets exist to connect those who need money with those who have it. The IPO is the most critical expression of that purpose — it transforms private companies into entities that can raise equity, reward employees, and fund growth at scale.
Chapter 2 · 04:18
How Andy reads the SpaceX IPO
Andy explains the fundamental purpose of markets — connecting capital-seekers with capital-providers — and why all IPO participants (issuers, banks, regulators, investors) are aligned toward making deals succeed. [1] — Andy Constan "Markets exist to connect those who need money with those who have it. The IPO is the most critical expression of that purpose — it transfor…" 03:58
A good IPO isn't about getting the perfect price — it's about satisfying every party in the room. SpaceX priced at $135, opened at $150, and ran into the $170s. The deal worked because SpaceX was aggressive about setting its own price, and every constituency walked away satisfied.
Chapter 3 · 08:27
Why underwriters and regulators want IPOs to work
Andy evaluates the SpaceX IPO: $85B float on a $2T cap, priced at $135, opened at $150, traded into the $170s — the biggest IPO in history and a well-executed deal. [1] — Andy Constan "A good IPO isn't about getting the perfect price — it's about satisfying every party in the room. SpaceX priced at $135, opened at $150, an…" 05:45
SpaceX's IPO surpassed Saudi Aramco to become the largest IPO in history, pricing at $135, opening at $150, and trading into the $170s.
Conventional wisdom says issuers hate leaving money on the table. Wrong. When you're selling just 4% of a $2 trillion company, a strong aftermarket pop signals quality to every future investor, employee, and M&A counterparty. The cost of underpricing 4% is tiny vs. the reputational return.
SpaceX issued approximately $85 billion of stock, representing roughly 4% of its $2 trillion market cap at IPO.
Chapter 4 · 17:05
Why issuers may want IPOs to trade higher
Mid-episode advertisements for Indeed's sponsored job postings and Canva's design platform.
The SpaceX IPO created over 4,000 paper millionaires among employees and contractors, with roughly 400 worth over $100 million on paper.
Chapter 5 · 18:44
From stock buybacks to new equity supply
Andy details how buybacks reduce share supply while issuance does the opposite, and quantifies the 2023–24 baseline of ~2% of GDP in net share reduction before the current reversal. [1] — Andy Constan "Net share supply: ~2% of GDP reduction (2023–24): During 2023 and 2024, share repurchases ran at roughly 2% of GDP, representing a net redu…" 21:15
The market has flipped. For years, buybacks reduced share supply by roughly 2% of GDP annually, acting as a structural tailwind. Now Google has canceled buybacks and issued $80B in stock. Meta and Amazon are following. The net swing is $600–$700 billion — entirely because AI needs to buy chips.
During 2023 and 2024, share repurchases ran at roughly 2% of GDP, representing a net reduction in share supply of about 2% of GDP.
Google canceled its share repurchase program and then issued $80 billion of common stock, exemplifying the broader buyback-to-issuance shift driven by AI CapEx.
The market has shifted from a ~$600–$700 billion net share reduction annually to effectively zero or net issuance, driven by AI CapEx funding needs.
One word: CapEx. Every market anomaly — declining buybacks, surging corporate bonds, rising equity issuance, growing RSU awards — has the same root cause. Hyperscalers are burning through free cash flow, shrinking balance sheet cash, and tapping every capital source to buy chips and build data centers.
Chapter 7 · 26:42
The 600 to 700 billion dollar shift in share supply
The AI investment thesis requires tokens to create disinflationary growth — more output for the same cost. If AI displaces workers instead, aggregate demand collapses and the math breaks. [1] — Andy Constan "AI tokens must either create more output for the same cost — disinflationary growth — or they lower costs by firing people. If workers are …" 30:08
AI-related capital expenditure is running at approximately $1 trillion per year and growing, forcing hyperscalers to raise capital through bonds, equity, and reduced buybacks.
NVIDIA, despite generating substantial revenue from chip sales, issued $28 billion in corporate bonds to help fund its CapEx ambitions.
AI tokens must either create more output for the same cost — disinflationary growth — or they lower costs by firing people. If workers are fired, who buys the output? The macro math only works if AI grows the pie without destroying the purchasing power of the workers who consume it.
The AI investment thesis only works macroeconomically if tokens produce more output for the same cost (disinflationary growth); if they merely replace workers, aggregate demand collapses.
Every productivity revolution — coal, railroads, assembly lines, the internet — follows the same arc: an initial boom as the tool deploys, then a plateau once it's fully implemented. China urbanized its farms and tripled output, then growth stopped. AI will likely follow the same script.
Chapter 8 · 35:29
Can AI create disinflationary productivity growth?
Wait — this section actually continues Andy's analysis. Checking timestamps: the Columbia ad appears at the transcript around the 27-minute mark (≈1652s). Correcting: this chapter covers Andy's historical productivity analysis and the messy AI transition argument.
Fear of technology ending jobs has happened every time a productivity revolution hit. It never did — because humans are wired to improve their situation. The 30-year-old displaced by AI has time to retool. The 15-year-old is already doing it. The transition will be painful, but it won't be final.
Chapter 10 · 41:30
Is AI CapEx holding up the economy?
Real wages lag growth, but consumers keep spending via asset dissaving — selling appreciated stocks, homes, and crypto. This has run for 3–4 years, supported by bubble-level asset prices. [1] — Andy Constan "Real wages aren't keeping up with growth, but consumers keep spending. The reason: extreme wealth effects. Rising stocks, homes, and crypto…" 42:36
Personal consumption represents two-thirds to three-quarters of US GDP, making consumer behavior the single most important driver of economic health.
Real wages aren't keeping up with growth, but consumers keep spending. The reason: extreme wealth effects. Rising stocks, homes, and crypto let Americans sell assets to consume above their income. It's been working for 3–4 years — but dissaving has a hard mathematical limit.
US consumers have been dissaving — selling assets to fund consumption above income — for roughly 3 to 4 years, supported by wealth effects from rising asset prices.
Chapter 11 · 50:52
Oil prices, war, and inflation
Falling oil prices lower headline inflation but push core higher as spending shifts. The US has been inflationary for 5–6 years; wars like Ukraine and the Middle East barely move the macro needle. [1] — Andy Constan "Oil falling looks like good inflation news. It isn't. Lower oil frees up consumer spending that flows into other goods, pushing core inflat…" 51:20
Oil falling looks like good inflation news. It isn't. Lower oil frees up consumer spending that flows into other goods, pushing core inflation back up. The US has been in an inflationary environment for 5–6 years and central banks show no real appetite to end it.
Chapter 12 · 52:00
Jalen Brunson, incentives, and long-term value
Falling oil prices lower headline inflation but push core higher as spending shifts. The US has been inflationary for 5–6 years; wars like Ukraine and the Middle East barely move the macro needle. [1] — Andy Constan "Oil falling looks like good inflation news. It isn't. Lower oil frees up consumer spending that flows into other goods, pushing core inflat…" 51:20
Falling oil prices lower headline inflation but shift consumer spending to other goods, pushing core inflation back up — the overall inflationary pressure does not disappear.
Chapter 13 · 54:00
Fed policy, tariffs, and what matters this summer
Andy's summer watchlist: Warsh builds credibility without acting, tariffs quietly expire July 24, AI CapEx flows with no ROI clarity, and equity issuance tests market absorption capacity. [1] — Andy Constan "Four things matter this summer: the new Fed chair builds credibility without acting, tariffs quietly expire on July 24 unless renewed, AI C…" 56:30
Andy Constan argued the US has been in an inflationary environment for 5 to 6 years with no sign central banks are willing to do anything meaningful about it.
Jalen Brunson signed a 4-year deal instead of 5, taking $113 million less, enabling the Knicks to build a championship-winning roster.
Four things matter this summer: the new Fed chair builds credibility without acting, tariffs quietly expire on July 24 unless renewed, AI CapEx keeps flowing with no return-on-investment clarity in 2026, and equity issuance continues to test how much supply the market can absorb.
Current tariffs are set to disappear automatically by July 24th unless actively renewed or replaced by the administration.
The Fed has mismanaged its balance sheet from the beginning, and that's a primary reason inflation persists. Warsh wants to lower short rates while running the balance sheet off — Bessant and Miran agree. But institutional resistance inside the Fed is fierce, and any change will take years of small tweaks.
No indexed bits in this chapter.
Show stoppers
The market has flipped. For years, buybacks reduced share supply by roughly 2% of GDP annually, acting as a structural tailwind. Now Google has canceled buybacks and issued $80B in stock. Meta and Amazon are following. The net swing is $600–$700 billion — entirely because AI needs to buy chips.
AI tokens must either create more output for the same cost — disinflationary growth — or they lower costs by firing people. If workers are fired, who buys the output? The macro math only works if AI grows the pie without destroying the purchasing power of the workers who consume it.
Real wages aren't keeping up with growth, but consumers keep spending. The reason: extreme wealth effects. Rising stocks, homes, and crypto let Americans sell assets to consume above their income. It's been working for 3–4 years — but dissaving has a hard mathematical limit.
Snapshots ()
Key Quotes ()
This episode
Claims & Sources
Factual claims made this episode, and whether a source was named.
The SpaceX IPO was the largest IPO in history, surpassing Saudi Aramco.
SpaceX's IPO priced at $135, opened at $150, and traded into the $170s.
SpaceX issued approximately $85 billion in stock, representing roughly 4% of its $2 trillion market cap.
The SpaceX IPO created over 4,000 millionaires on paper, with approximately 400 worth over $100 million.
During 2023 and 2024, net share repurchases were roughly 2% of GDP, representing a net reduction in share supply.
The shift from net share reduction to net equity issuance represents a $600–$700 billion annual swing in share supply.
Google canceled its share repurchase program and subsequently issued $80 billion of common stock.
NVIDIA issued $28 billion in corporate bonds despite generating substantial revenue from chip sales.
AI-related capital expenditure is running at approximately $1 trillion per year and growing.
Personal consumption accounts for two-thirds to three-quarters of US GDP.
Current tariffs are scheduled to expire on July 24 and will automatically disappear unless actively renewed.
Falling oil prices increase spending on other goods, pushing core inflation up even as headline inflation falls.
The US has been in a persistently inflationary environment for 5 to 6 years with no meaningful central bank response.
Jalen Brunson took $113 million less by signing a 4-year instead of 5-year contract, enabling the Knicks to sign talent and win the championship.
This episode
Cast
-
Discussed as the incoming Fed chair who Andy Constan expects to build credibility without dramatic action, while hoping he will reform balance sheet policy.
-
Used as an example of long-term incentive alignment: took $113 million less to sign a shorter contract, enabling the Knicks to build a championship roster.
-
Central case study throughout the episode as the largest IPO in history, priced at $135, opening at $150 and trading into the $170s.
-
Discussed extensively regarding balance sheet management failures, the incoming chair Kevin Warsh, and the institutional resistance to QT policy changes.
-
Track
Discussed as a company increasing both share repurchases and RSU awards simultaneously, while also issuing $28 billion in corporate bonds to fund CapEx.
-
Track
Cited as the clearest example of the buyback-to-issuance shift: canceled share repurchases and then issued $80 billion of common stock.
-
Referenced as the NBA championship-winning team enabled by Jalen Brunson's salary sacrifice, used as an analogy for long-term incentive alignment.
-
Track
Mentioned as a hyperscaler decreasing or eliminating share repurchases as CapEx demands grow, with equity issuance expected to follow.
-
Track
Cited as a hyperscaler that does not repurchase shares and is already issuing corporate bonds, with equity issuance expected soon.
-
Track
Noted as still repurchasing shares but expected to eventually shift to equity issuance as CapEx demands grow.
-
Referenced as the previous record-holder for largest IPO, now surpassed by SpaceX.
-
Mentioned as one of the large AI companies expected to IPO following SpaceX, contributing to the new equity supply pipeline.
-
Mentioned alongside Anthropic as a major upcoming IPO that will add to the new equity supply entering the market.
-
Andy Constan's Substack newsletter and website where his market research and analysis is published.
-
Referenced as approaching all-time highs at the time of recording, cited as evidence of the asset price wealth effect supporting consumer spending.
Stats
