Kurz gesagt: The four largest hyperscalers now convert surplus cash into fixed assets faster than those assets earn, making return on capital the single variable that decides this trade.
Capex roughly doubled in a year while the cash engine stalled
Microsoft, Alphabet, Amazon und Meta sind von einer disziplinierten Reinvestition zu einem regelrechten Ausbau übergegangen. Die Gesamtinvestitionen der vier Unternehmen beliefen sich im Jahr 2025 auf fast $410 Milliarden und liegen bei für 2026 auf rund $725 Milliarden angestrebt, a step-up of about 77 per cent that most analysts extend past $1 trillion by 2027. On a trailing four-quarter basis through the first quarter of 2026, the four already $433,9 Milliarden ausgegeben, wobei die vierteljährlichen Ausgaben in Höhe von $129,8 Milliarden im Vergleich zum Vorjahreszeitraum um 80 Prozent gestiegen sind.
This marks a change of business model rather than a change of degree. Roughly half of the 2026 spend funds servers and silicon that depreciate over five to six years; the rest funds data-centre shells and power infrastructure. Companies that spent a decade compounding cash on other people’s balance sheets now build the plant themselves.
| Unternehmen | Investitionen 2025 (ca.) | Prognose für 2026 | Year-over-year growth |
|---|---|---|---|
| Amazon | ~$100B | ~$200B | ~+100% |
| Microsoft | ~$95B | ~$190B | ~+100% |
| Alphabet | ~$85B | $175–185B | ~+110% |
| Meta | ~$70B | $125–145B | ~+80% |
| Kombiniert | ~$410B | ~$725B | ~+77% |
Free cash flow became the swing variable that earnings per share conceals
Earnings per share (EPS) still look strong. The cash flow statement carries the story the income statement absorbs. Amazon reported Freier Cashflow im ersten Quartal von $1,2 Milliarden, ein Rückgang um rund 95 Prozent year over year, as quarterly capex reached $44.2 billion and nearly every incremental dollar of operating cash went back into infrastructure. Meta generated $12.4 billion of free cash flow in the same quarter against $19 billion of capex, and one widely cited model has its Der freie Cashflow wird bis 2027 negativ ausfallen, at a company that produced $43.6 billion of it in 2025.
The mechanical consequence lands on the buyback-and-dividend cushion that supported these multiples. Once capital expenditure climbs above half of operating cash flow, the level Microsoft hat nun die Marke überschritten, the return of capital to shareholders becomes the residual rather than the priority. Markets have begun to reprice both names on that trajectory instead of on their beat-and-raise headlines.
| Unternehmen | Freier Cashflow im 1. Quartal 2026 | Investitionsausgaben im 1. Quartal 2026 | Signal |
|---|---|---|---|
| Amazon | $1.2B (down 95% YoY) | $44,2B | Die Marge von Amazon Web Services (AWS) sank aufgrund von Abschreibungen im Bereich KI auf 37,71 TP3T |
| Meta | $12.4B | $19.0B | Free cash flow modelled negative by 2027, against $43.6B in 2025 |
A depreciation wall lands between 2027 and 2029
The most underappreciated feature of this cycle is timing. Depreciation recognises spending gradually, five to six years for servers and twenty-five to forty for buildings, so today’s income statements capture only about ein Drittel der laufenden Investitionsausgaben als Aufwand. Trailing depreciation and amortisation of roughly $149 billion sits against $433.9 billion of trailing capex. The gap stands deferred rather than saved.
Meta’s infrastructure depreciation already rose from $7.32 billion in 2023 to $13.36 billion in 2025, an 83 per cent increase ahead of the largest spending waves. As the 2026 and 2027 cohorts deploy, that expense flows onto income statements through 2029 and compresses reported margins mechanically, whatever the revenue does. Firms are pre-funding the crossover: Meta, Alphabet and Amazon together raised well over $100 billion of debt across late 2025 and 2026, and sell-side desks project the sector may need $1,5 Billionen an neuen Schulden in den kommenden Jahren.
The revenue gap carries the entire thesis
Every figure above resolves into one question: what is the return on this capital? Artificial intelligence (AI) revenue is real and growing quickly. Amazon Web Services runs near a $150 billion annualised rate, Microsoft’s AI business sits at roughly a $37 billion run rate growing 123 per cent, and Google Cloud grew 63 per cent in the quarter. Investment is scaling about 50 per cent faster than the revenue it is meant to produce, and Sequoia’s analysis puts the jährliche Lücke von rund $600 Milliarden between infrastructure spend and what the AI ecosystem sells.
The demand-side evidence offers limited reassurance so far. MIT’s Project NANDA found that five per cent of enterprise generative-AI pilots produced a measurable profit-and-loss impact, against $30 to $40 billion of corporate spending. Capital intensity across the group now sits at 45 to 57 per cent of revenue, the range of a regulated utility rather than an asset-light platform, while sector enterprise-value-to-EBITDA (earnings before interest, taxes, depreciation and amortisation) multiples near 25 times approach the extremes last seen at the 2000 telecom peak. That comparison describes the risk embedded in today’s price.
Was dies von den Kapitalverteilern verlangt
The bull case still holds. Compute demand is genuine, the leaders are the credible builders, and whoever owns the winning capacity may earn extraordinary returns. The shape of the trade has changed: an investor buying these names in 2026 underwrites a capital-allocation decision where they once bought a cash-compounding machine. The asset-light model that justified premium multiples is giving way, quarter by quarter, to a capital-intensive one, and the market has started to price the buildout ahead of the boom.
Allegory Capital invests where capital intensity, returns on capital and the energy transition intersect, so this shift sits at the centre of how the firm reads the current cycle. This piece opens the argument. Allegory Capital bittet Fondsmanager und Analysten aus dem Technologiesektor um offizielle Stellungnahmen sowie Hintergrundinformationen zu drei Fragen:
1. Ab wann wird die Investitionsintensität zu einem echten Problem, und wie ließe sich die Rentabilität nachweisen? Earnings are growing while capex absorbs a rising share of operating cash flow. At what point does that shift from a feature to a problem for you? What specific evidence would convince you that the hundreds of billions committed to AI infrastructure are earning an adequate return on capital, and which metrics do you actually watch: incremental return on invested capital (ROIC), capex-to-revenue, free-cash-flow conversion, cloud backlog, or unit economics per token?
2. Is revenue and earnings growth enough if free cash flow falls sharply? Wenn Aktienrückkäufe und Dividenden zurückgehen, weil Barmittel in den Ausbau der Produktionsanlagen fließen, rechtfertigt das nominale Gewinnwachstum den Konjunkturzyklus dann für Sie noch? Und unterschätzen Anleger, wie grundlegend die KI die „Big Tech“-Unternehmen von einem vermögensarmen, cash-generierenden Modell in ein kapitalintensives Modell verwandelt, das ein anderes Bewertungsverhältnis verdient?
3. Wie lange darf die Zeitspanne zwischen der Ausgabe und dem Erhalt eines Dollars maximal betragen? What is the outer bound you will tolerate between capital deployed on AI infrastructure and the incremental revenue and cash flow needed to justify it, one year, three, or five, before the thesis itself comes into question rather than the timing?
Die Antworten fließen in ein Briefing von Allegory Capital zum KI-Kapitalzyklus ein. Die Beiträge können auf Wunsch namentlich genannt oder anonym veröffentlicht werden.
Referenzen
- Tom’s Hardware. Die Investitionsausgaben von Google, Microsoft, Meta und Amazon werden 2026 voraussichtlich $725 Milliarden erreichen, was einem Anstieg von 77% gegenüber dem Vorjahr entspricht. https://www.tomshardware.com/tech-industry/big-tech/big-techs-ai-spending-plans-reach-725-billion
- Silizium-Analysten. Investitionen der Hyperscaler in KI im Jahr 2026: $434B in den letzten vier Quartalen, Rückstand bei Abschreibungen, Schuldenwelle. https://siliconanalysts.com/analysis/hyperscaler-ai-capex-depreciation-wall-2026
- Globaler Rechenzentrums-Hub. Amazon und Metas $335B im 1. Quartal: Der Wendepunkt beim freien Cashflow. https://www.globaldatacenterhub.com/p/amazon-and-metas-335b-q1-the-free
- ValueAdd VC. Vergleich der Investitionsausgaben von KI-Hyperscalern. https://valueaddvc.com/blog/ai-hyperscaler-capex-compared-why-microsoft-google-meta-and-amazon-are-all-spending-at-once
- Forbes (Jason Kirsch). Die Lücke zwischen AI-Investitionen und Umsatz vergrößert sich. https://www.forbes.com/sites/jasonkirsch/2026/06/02/the-ai-capex-to-revenue-gap-is-widening—and-markets-are-starting-to-notice/
