🟢⚠️ OPENAI GOES INTO “TOO BIG TO FAIL” MODE
OpenAI's massive investment (~$1.4B for 30 GW by the end of 2025) ends the bubble / non-bubble debate. AI is entering a phase of systemic dependence.
WHAT CHANGES:
– We move out of the narrative, financial viability becomes central
– AI actors become highly interdependent
– OpenAI is established as an assumed systemic risk
THE HEART OF THE PROBLEM:
– Projected revenues are very high (~$155B in ARR by 2030)
– But a structurally deficit model
– Negative cash flow, accumulated deficit ~ $971B (2025–2030)
– For every $1 earned = $3 spent
– Permanent dependence on external financing
CHANNELS OF CONTAGION:
– Increase in the cost of capital → compression of AI multiples
– Potential domino effect on NVIDIA and Microsoft
– Regulatory pressure related to public aid
– Physical constraints (energy, water)
– Monetization still not scalable
THE AI DILEMMA:
– Success → macroeconomic tensions
– Failure → massive devaluations
– In any case, systemic risk
THE ROTATION IS HERE:
– Non-profitable pure play AIs penalized
– Profitable and essential components (energy, memory, utilities) favored
– AI exposure occurs through profitable infrastructures
CONCLUSION:
– OpenAI becomes a central point of fragility
– High contagion risk in the event of a shock
– Public intervention likely
– Stability preserved, returns under pressure
