Buybacks are becoming selective
The aggregate market continues to repurchase shares. At the most capital-intensive AI investors, however, free cash flow and credit quality determine whether that buyer is present in the next downturn.
The Portfolio at a Glance
The following overview summarizes the strategic building blocks, position directions, and allocated assets for this research theme.
Long
Japan as a growing corporate buyer
Rising payouts, governance reform and buybacks make Japan the mirror image of a market in which the corporate side is losing importance.
Long
Less concentration in the US index
The paired position targets the combination of capital intensity and index concentration rather than the entire US equity market.
Short
Credit-dependent AI capacity
This is the clearest test of whether rapidly ageing hardware, financing and contract quality can all hold at once.
Long
Market infrastructure for greater activity
Exchanges, futures markets and liquidity providers earn from hedging and turnover without having to forecast the direction of the next equity decline precisely.
Long
A hedge against fiscal competition
Gold addresses the risk that persistent deficits, higher real yields and currency hedging constrain the valuation of US risk assets.
The analysis
Who absorbs new shares when the largest index companies retire fewer of their own? Index funds and occupational pensions stabilise demand. They do not replace a buyer that removes shares from the market over years.
Five leading hyperscalers were estimated to spend $700–725 billion in 2026, against combined operating cash flow of $600–750 billion.1 The platforms want to invest. Cash becomes scarce after investment, leases and financing.
New shares from listings, compensation and acquisitions therefore need buyers who care about price. Real-yielding Treasuries give those buyers an alternative. The market adjusts through a lower price per dollar of cash flow.
The Argument
Share count is a better measure of the corporate buyer than an authorisation
Aggregate US repurchases remain close to $1 trillion a year. Apple, banks and capital-light companies remain material buyers. For this index thesis, however, the relevant part of the market is the group with the highest valuations and the greatest investment need.
At Alphabet and Meta, buybacks fell to zero in individual quarters while investment rose sharply; Amazon and Oracle bought back little in any case.3 The reliable test is therefore the change in diluted shares outstanding after stock compensation and any equity issuance. A large authorisation without a declining share count is not equivalent market support.
Capital flows accelerate a re-rating
Capital flows determine the speed. Return on capital, free cash flow and credit capacity determine the direction.
Return on capital, free cash flow and credit capacity decide whether valuation falls. Capital flows decide how quickly. Research on inelastic markets shows that small net flows can cause large short-term price moves.6 Arbitrage, rebalancing and new buyers counteract them, so the effect cannot be compounded over a decade.
After a fundamental disappointment, volatility targeting, trend followers, leveraged ETFs and option hedging amplify the move.10 They determine the severity of the decline. A weaker return on AI investment or a tighter credit market would trigger it.
Platforms are becoming infrastructure owners with software margins
vanishingBid.claims.capitalBody
Scale from April to June 2026, in $ billions; Microsoft is fiscal Q4 2026 and Oracle is the fiscal year to May 2026:211
| Company | Period | Investment | Operating cash flow | Free cash flow |
|---|---|---|---|---|
| Microsoft | Apr–Jun 2026 | 41.0 (incl. 5.6 finance leases) | 55.4 | 19.6 |
| Alphabet | Apr–Jun 2026 | 44.9 | 39.1 | −5.8 |
| Meta | Apr–Jun 2026 | 31.1 | 31.9 | 0.8 (after leases) |
| Amazon | Apr–Jun 2026 | 54.2 | – | −8.8 |
| Oracle | Jun 2025–May 2026 | 55.7 | 32.0 | −23.7 |
A tight oligopoly can earn high returns on compute, power and customer access. The figures nevertheless change the platforms’ character: financing and replacement investment now matter more for repurchase capacity than an authorisation.
The thesis fails if investment earns clearly more than its cost of capital after replacement investment and power costs, free cash flow rises, and share counts fall again by 2028 or 2029 without additional debt.
Demography acts later, slowly and through heirs
Demography is not the next market trigger. Required minimum distributions begin at 75 for people born from 1960, so the new cohorts arrive between 2035 and 2039.4 A distribution can be transferred in securities or rolled into an IRA; it is not identical to an equity sale.
The slower channel is the wealth transfer after death. A stepped-up tax basis reduces the incentive to sell appreciated stock before death. Afterwards, heirs may diversify concentrated holdings or empty inherited IRAs within ten years. This can cap later recovery valuations; it does not set a date for a sell-off.
Valuation: earnings against cash flow
Earnings-based ratios understate the problem. Investment leaves the cash account immediately, while it enters earnings gradually through depreciation. This analysis therefore values the companies on free cash flow and share count, both from reported figures. Microsoft, the largest of the five:12
| Measure | Microsoft |
|---|---|
| Market capitalisation, 15 October 2026 | $4.02 trillion |
| Free cash flow, fiscal year to June 2026 | $67.0 billion |
| P/E on reported earnings | 30 |
| Price to operating cash flow | 22 |
| Price to free cash flow | 60 |
| Change in share count in the fiscal year | −0.1% |
Investors pay twice as much for a dollar of Microsoft free cash flow as for a dollar of earnings. At Alphabet and Amazon, the retreat of the corporate buyer already appears in share count: repurchases no longer offset stock compensation. Free cash flow fell at Meta and Amazon in 2025.12
From April to June 2026, free cash flow was already negative at Alphabet and Amazon, and at Oracle for the full fiscal year to May 2026.
| Company | Free cash flow, latest fiscal year ($bn) | Prior year ($bn) | Share count, four quarters to June 2026 |
|---|---|---|---|
| Microsoft (fiscal year to June 2026) | 67.0 | 71.6 | −0.1% |
| Alphabet (2025) | 73.3 | 72.8 | +0.9% |
| Meta (2025) | 46.1 | 54.1 | −0.2% |
| Amazon (2025) | 7.7 | 32.9 | +0.9% |
The gap between earnings and cash-flow valuation closes in one of two ways: free cash flow catches up because investment earns high returns, or the share price adjusts. The base case combines both. Platforms continue to grow, but the market pays less for each dollar of cash flow and buybacks remain thin.
In a credit or utilisation shock, the share price falls faster than cash flow can catch up. In the oligopoly case, cash flow catches up and buybacks return.
Foreign investors are already hedging dollar positions more heavily without selling US equities.7 A weaker dollar can therefore precede the valuation decline.
The Investment Selection
The selection does not require one large market decline. It combines a relative allocation to Japan and equal weighting with businesses that can benefit from greater trading, hedging and liquidity provision, plus a tightly defined negative exposure at the credit-dependent end of the AI build-out.
The portfolio expresses a relative shift: Japan and equal-weighted US equities against the concentrated index; exchanges and market makers that earn from trading and hedging; and a limited short at the credit-financed end of the AI build-out.
| Building block | Position | Direction | Role in the thesis |
|---|---|---|---|
| Japan buys its own shares | NEXT FUNDS TOPIX ETF (1306.T) | Long | Rising buybacks and payouts at a lower starting valuation |
| Less index concentration | S&P 500 Equal Weight (RSP) versus Nasdaq-100 (QQQ) | Long RSP, short QQQ | Equal weighting wins if the largest platforms retire fewer shares |
| Credit-financed AI capacity | CoreWeave (CRWV) | Short | Utilisation, GPU residual value, refinancing and customer concentration must all hold |
| Market infrastructure | Japan Exchange Group (8697), Cboe (CBOE), CME Group (CME), Flow Traders (FLOW) | Long | Earn from trading, hedging and liquidity regardless of market direction |
| Fiscal risk | iShares Physical Gold ETC (SGLN) | Long | Hedge against persistent deficits and a weaker dollar role |
The structural position and Leeway ratings answer different questions. The thesis evaluates a longer-term driver; ratings assess business quality, current fundamentals and historical valuation. A conflict is information: a high Cycle Rating on a negative position can flag a potential value trap, while a weak rating on a long position can flag a catalyst not yet visible in current fundamentals.
The Business-Rating scores the quality of the business model, independent of the share price. The Market-Fit-Rating tests eighteen fundamental figures for how well the company currently fits the market; a negative reading implies expected negative performance. The Cycle-Rating places the valuation in the stock’s own history: a higher figure means the shares are historically cheaper. The Leeway-Score combines the three in equal parts. How the ratings are calculated
The recommendations since publication
Not enough trading sessions have elapsed since publication to compute a consistent performance series.
Japan as a growing corporate buyer
LongRising payouts, governance reform and buybacks make Japan the mirror image of a market in which the corporate side is losing importance.
NEXT FUNDS TOPIX ETF
Long
Role in thesis: Core allocation to a market with rising buybacks and payouts.
Investment case: Japanese governance reform and a lower starting valuation create a different corporate counterparty to sellers from the capital-intensive US technology sector.
What would invalidate it: The case weakens if cross-shareholding unwinds release more supply than buybacks absorb, or if the yen abruptly damages exporter earnings.
Position and what to watch
Position: Use as a multi-year Japan allocation, or against a beta-adjusted US index short; hedge the yen only partly.
What to watch: Monitor repurchase announcements, payout ratios, domestic and foreign NISA flows, and Bank of Japan policy.
Principal risk: A stronger yen, tighter Bank of Japan policy or a setback in governance could impair the relative case.
Your knowledge edge
Understand the real drivers of our portfolio
Look behind the scenes: see not only what we buy, but why. Get the complete view of our strategy.
- Every position revealed: Access each holding in our portfolio.
- Understand the investment thesis: See the exact role each security plays in our return.
- See the connections: Analyse our long and short strategies in direct context.
The Sequence of Evidence
| Window | What happens | What it means for the portfolio |
|---|---|---|
| Late 2026 to 2027 | Hyperscalers disclose investment budgets, buybacks, share counts and new lease, debt or joint-venture commitments. | The difference between gross buybacks and declining share count shows whether capital-intensive platforms remain reliable buyers. |
| 2027 to 2028 | Listings, lock-up expiries and data-centre funding meet the first serious test of AI utilisation and credit spreads. | A credit event or weaker rental prices would be a plausible fundamental trigger. Systematic strategies could amplify the ensuing move. |
| 2028 to 2030 | The first major replacement cycle for 2024–27 investment reveals depreciation, replacement needs and cash returns on the added capital base. | This is the clear test between a durable AI-oligopoly return and a lower multiple for more capital-intensive platforms. |
| 2035 to 2050 | New distribution cohorts, inherited IRAs and heir diversification increase demand for income and broader asset allocation. | Demography acts as a slow restraint on recovery multiples, not as a short-term selling trigger. |
Three Possible Paths
| Path | Weight | What happens |
|---|---|---|
| AI becomes a tight oligopoly | 25% | Usage and pricing support returns clearly above the cost of capital. Hyperscaler free cash flow and share repurchases recover by 2028 or 2030, leverage remains manageable and the US index retains high multiples. |
| Returns on capital normalise | 50% | AI remains useful, but surplus is shared with customers, suppliers, labour and the state. Repurchases at capital-intensive platforms stay thin, new supply becomes more relevant and the US multiple drifts towards 16 to 19 over time. |
| Credit and supply meet weaker utilisation | 25% | Rental pricing or utilisation disappoints, data-centre credit becomes more expensive and new shares arrive into a weaker market. Concentrated indices may then fall much further than the broad market before multiples settle at 13 to 15. |
Counter-arguments and Risks
The strongest counter-case is that AI spending is high because leading platforms have broad moats. Distribution, data, proprietary applications and access to power constrain competition. Then weak free cash flow is temporary: revenue rises, hardware is used more efficiently and buybacks return.
Global savings demand may also remain stronger than assumed. Japan’s expanded NISA programme has bought substantial foreign securities, not just domestic assets.8 Auto-enrolment, child accounts and growing retirement systems can offset demographic pressure for a long time. The test therefore remains the free cash flow and share count of the largest companies.
Unresolved Market Factors
How strongly capital flows affect long-run prices remains empirically open. The short-run effect is well documented. No formula converts slowly falling inflows into a ten-year equity decline because markets adjust through price, new buyers and issuance.
Financing also remains open. Insurers and private lenders can sustain the AI build-out for longer than bond markets alone suggest.9 That shifts risk into the financing chain. If utilisation falls, depreciation and losses arise there while power and site infrastructure remain usable.
What Investors Should Watch
- The change in diluted shares outstanding at Microsoft, Alphabet, Amazon, Meta and Oracle after stock compensation and any equity issuance.
- Hyperscaler capital expenditure against operating cash flow, including leases, joint ventures, purchase commitments and bond issuance.
- US net equity supply after buybacks, stock compensation, initial public offerings, follow-ons and stock-financed acquisitions.
- Credit spreads in AI infrastructure, data centres and highly levered cloud providers.
- GPU rental pricing and utilisation, plus useful-life assumptions in depreciation disclosures.
- The gap between the ten-year US real yield and the S&P 500 earnings yield.
- Foreign dollar hedge ratios, especially among European, Japanese, Korean and Taiwanese investors.
- The frequency of unusual ETF premiums or discounts and the level of implied volatility.
What Would Falsify the Thesis
The lower-US-multiple view needs revision if these conditions occur together and persist.
- AI investment earns returns clearly above the cost of capital after replacement capital expenditure and power costs, without rising leverage.
- Large hyperscalers resume material diluted-share reduction by 2028 or 2029 from free cash flow.
- New equity issuance is absorbed at high valuations while global savings continue to flow disproportionately into capitalisation-weighted US indices.
Sources
- Reuters, 22 July 2026: AI investment and free cash flow at large technology companies
- CNBC, 28 August 2026: AI spending and cash flow at large technology companies
- Goldman Sachs, cited by TheStreet: buybacks and AI investment at large technology companies
- US Securities and Exchange Commission: SECURE 2.0 overview
- Bank for International Settlements: Passive investing and securities markets
- Gabaix, X. and Koijen, R.: In Search of the Origins of Financial Fluctuations: The Inelastic Markets Hypothesis, 2021
- Bank for International Settlements: Dollar movements and currency hedging, 2025
- Daiwa Institute of Research: Japan’s NISA and securities flows
- European Central Bank, Financial Stability Review special feature on private credit, May 2026
- International Monetary Fund, Global Financial Stability Report, April 2026
- Reuters, 29 July 2026: Microsoft fiscal Q4 2026 results and finance-lease detail
- Leeway API: market capitalisation, valuation multiples and fundamentals for Microsoft, Alphabet, Meta and Amazon, 15 October 2026
How this analysis is produced
The assessment is produced in several steps. Independent model families answer the same question separately and then attack the results. What you read here has survived several rounds.
- Two independent first theses. The same opening question goes to several model families that cannot see one another. Disagreements are kept, not averaged away.
- Dated evidence. Every claim that depends on facts is broken into individual search questions and answered with dated, sourced web research. Question, answer, sources and timestamp are logged and remain traceable.
- Adversarial review. Several review roles attack the thesis from different angles: one hunts for the strongest refutation, one for the awkward edge cases, one tests whether a path from thesis to share price exists at all, one checks the timeline for contradictions. Each role raises its own questions, which are again answered with evidence.
- Merge, then the next round. The surviving theses are merged into one and attacked again. The counter-position and the unresolved tension on this page come out of that step. They were not bolted on afterwards to look balanced.
- Back to the start. The process runs again until there is a clear result and a list of tradable companies with structural advantages.
Any analysis can be wrong. That is why the falsification criteria and the counter-position sit on the same page as the thesis, not in the small print.
The numbers shown against individual companies do not come from this process. The Business-Rating scores business-model quality, the Market-Fit-Rating eighteen fundamental figures against the current market, the Cycle-Rating the valuation against the stock’s own history. They are documented under the Leeway scores.
Investor Questions
Does the thesis claim passive funds will cause the next sell-off?
Passive funds move prices through the creation and redemption of units. In redemptions, however, they sell in proportion to market weights. In this thesis, a disappointment in platform returns on capital and credit capacity would be the trigger.5
Are buybacks not still close to one trillion dollars per year?
They are. The thesis does not claim that aggregate US repurchases have ceased. It asks whether the largest and most highly valued AI investors can still reduce their share count when capital expenditure, leases and financing grow faster than free cash flow.
Why is demography not an immediate selling reason?
A required retirement-account distribution is neither necessarily an equity sale nor necessarily an exit from capital markets. Rollovers, in-kind transfers, continuing contributions and tax rules spread the effect. Later, inheritances and inherited IRA accounts may nevertheless increase the demand for liquidity and diversification.
Evaluate the selected companies with the three Leeway ratings
The Business Rating assesses business model quality, the Market-Fit Rating evaluates fundamentals in the current market regime, and the Cycle Rating contextualises valuation within the company’s history. Evaluate each company with Leeway’s general equity analysis.
Company Valuation and Fundamental Analysis
The data is recalculated on a weekly basis and depends on the current market value of the company and the balance sheet figures of the annual financial statements. The market value changes continuously with price changes, the balance sheets are created annually and change the valuation massively. The time of the annual financial statements and the metrics used can be viewed under "Metrics". Further information on how the analyses work can be found as tooltips directly on the analyses as well as in our explanations.
General
The information is intended to support you in your independent decision-making in implementing investment strategies and serves purely informational purposes. Past performance is not a reliable indication of future performance. No warranty can be given for the completeness, correctness and accuracy of the listed content. The information does not constitute specific investment recommendations. We neither know you nor your financial situation and do not provide investment advice. Only licensed investment advisors with knowledge of your personal circumstances may do this. PWP Leeway UG (limited liability) is not an investment advisor and does not collect any personal data for the purpose of investment optimization. PWP Leeway UG (limited liability) is a provider of investment recommendations and investment strategy recommendations. As such, it is registered with and supervised by the Federal Financial Supervisory Authority (BaFin). The legal notices on the use of the website and the General Terms and Conditions of PWP Leeway UG (limited liability) also apply.
All price data are closing prices of the respective stock exchanges. Price information and master data are provided by an external service provider. Furthermore, public trading data, such as provided by Finra.org, is used to analyze market behavior. No warranty can be given for the completeness, correctness and accuracy of the listed content.
Disclosure of interests: To the best of our knowledge, Leeway and the persons responsible for creating or publishing the respective content do not hold a net long or net short position exceeding 0.5% of the issued share capital of any issuer discussed in that content at the time of publication. There are no other interests or business relationships that could conflict with objectivity, unless expressly stated otherwise in the respective content.
