Die Pensionszusage an Gesellschafter-Geschäftsführer im Fokus der Finanzverwaltung.
Kaum eine Gestaltung ist so effektiv und tückisch. Wirksamkeit, § 6a EStG, Fremdvergleich, Finanzierbarkeit und aktuelle BFH-Rechtsprechung entscheiden.
Ein strukturierter Überblick zu den zentralen Grundsätzen und typischen Fehlern der Praxis: https://t.co/si9kzeHafw
⚡️Big Tech just entered the dilution phase of the AI war
For years, Meta was one of the cleanest cash machines on Earth. Monster margins.
Buybacks. Cash pile. Ads engine. Social graph. Distribution. Operating leverage.
The whole story was that Big Tech generated so much cash that it could fund the future internally and still return capital to shareholders.
Now Meta is looking at the AI frontier and saying: internal cash may not be enough.
That is insane in the old framework.
In the new framework, it makes sense.
AI is forcing the largest companies in history to behave like wartime infrastructure states. The question is no longer “can Meta afford AI?” The question is “can Meta afford to lose the AI platform layer?” The answer is no. Losing that layer means Meta risks becoming downstream of someone else’s model, someone else’s agent, someone else’s interface, someone else’s operating system for consumer attention.
Zuckerberg sees the threat clearly. Social media was the last major platform layer Meta truly owned. Mobile wounded them because Apple controlled distribution. AI agents could wound them again if users, advertisers, creators, and businesses start routing through OpenAI, Google, xAI, Apple, or some agentic interface that weakens Meta’s control over attention.
That is why Meta will spend almost whatever it takes.
The raise is a confession that the next platform war is physically expensive. This is not code alone. This is data centers, GPUs, networking, power, land, cooling, model training, inference, video generation, ads optimization, recommendation systems, AI companions, wearables, creator tools, enterprise messaging, and agentic commerce. It is an empire-scale buildout.
The brutal part: AI is converting Big Tech from shareholder-return machines into capital-hungry infrastructure empires.
That changes the whole equity compact.
The old promise was: “We print cash and buy back stock.”
The new promise is: “Let us dilute you now so we do not get displaced later.”
That is a very different shareholder relationship.
Meta’s case is especially sharp because the market still remembers Reality Labs. Zuckerberg has already asked investors to trust a huge future-oriented capital burn once. That made sense strategically, but the timing and monetization were rough. Now he is effectively asking the market for trust again, only this time the threat is more immediate and the prize is bigger.
The market will have to decide whether Meta’s AI spend is defensive oxygen or another giant future-story hole.
Deep down, the answer is probably both.
Meta genuinely needs to build. Underbuilding would be strategically suicidal. But the cost is going to be uglier than investors want to admit. The AI race is becoming so expensive that even the winners may look financially worse before they look dominant. Higher capex, more dilution, more depreciation, more power commitments, more infrastructure complexity, more margin pressure.
That is why this matters.
This is the end of “AI is margin-light software magic.”
AI at frontier scale is industrial. It eats capital, electricity, chips, land, talent, and time. The companies closest to the frontier are now revealing the truth through their financing choices.
Google raised.
Meta is considering raising.
Others will follow.
The floodgates are open.
The winners will be the companies that turn dilution into control.
The losers will be the companies that dilute merely to stay alive in the race.
⚡️Big Tech just entered the dilution phase of the AI war
For years, Meta was one of the cleanest cash machines on Earth. Monster margins.
Buybacks. Cash pile. Ads engine. Social graph. Distribution. Operating leverage.
The whole story was that Big Tech generated so much cash that it could fund the future internally and still return capital to shareholders.
Now Meta is looking at the AI frontier and saying: internal cash may not be enough.
That is insane in the old framework.
In the new framework, it makes sense.
AI is forcing the largest companies in history to behave like wartime infrastructure states. The question is no longer “can Meta afford AI?” The question is “can Meta afford to lose the AI platform layer?” The answer is no. Losing that layer means Meta risks becoming downstream of someone else’s model, someone else’s agent, someone else’s interface, someone else’s operating system for consumer attention.
Zuckerberg sees the threat clearly. Social media was the last major platform layer Meta truly owned. Mobile wounded them because Apple controlled distribution. AI agents could wound them again if users, advertisers, creators, and businesses start routing through OpenAI, Google, xAI, Apple, or some agentic interface that weakens Meta’s control over attention.
That is why Meta will spend almost whatever it takes.
The raise is a confession that the next platform war is physically expensive. This is not code alone. This is data centers, GPUs, networking, power, land, cooling, model training, inference, video generation, ads optimization, recommendation systems, AI companions, wearables, creator tools, enterprise messaging, and agentic commerce. It is an empire-scale buildout.
The brutal part: AI is converting Big Tech from shareholder-return machines into capital-hungry infrastructure empires.
That changes the whole equity compact.
The old promise was: “We print cash and buy back stock.”
The new promise is: “Let us dilute you now so we do not get displaced later.”
That is a very different shareholder relationship.
Meta’s case is especially sharp because the market still remembers Reality Labs. Zuckerberg has already asked investors to trust a huge future-oriented capital burn once. That made sense strategically, but the timing and monetization were rough. Now he is effectively asking the market for trust again, only this time the threat is more immediate and the prize is bigger.
The market will have to decide whether Meta’s AI spend is defensive oxygen or another giant future-story hole.
Deep down, the answer is probably both.
Meta genuinely needs to build. Underbuilding would be strategically suicidal. But the cost is going to be uglier than investors want to admit. The AI race is becoming so expensive that even the winners may look financially worse before they look dominant. Higher capex, more dilution, more depreciation, more power commitments, more infrastructure complexity, more margin pressure.
That is why this matters.
This is the end of “AI is margin-light software magic.”
AI at frontier scale is industrial. It eats capital, electricity, chips, land, talent, and time. The companies closest to the frontier are now revealing the truth through their financing choices.
Google raised.
Meta is considering raising.
Others will follow.
The floodgates are open.
The winners will be the companies that turn dilution into control.
The losers will be the companies that dilute merely to stay alive in the race.