Nvidia is no longer just selling shovels in the AI gold rush. It is now financing the gold diggers. And if you’ve spent any time in crypto since 2020, you know exactly how this movie ends.
I didn’t write this for clicks. I wrote it because someone had to. The Bloomberg interview with Janus Henderson’s fund manager last week painted a picture of calm: “Nvidia’s circular financing risks are currently controllable.” Controllable? Let’s unpack that.
Context: The Capital Loop
The AI industry is running on a simple feedback loop. AI application growth -> more compute demand -> tech giants raise capex -> Nvidia sells more GPUs -> Nvidia helps finance those customers so they buy even more GPUs -> customers train models -> hopes of future revenue -> repeat. The fund manager explicitly confirmed that Nvidia is using “financing guarantees and other methods to help promote AI infrastructure construction.”
In crypto, we call this “yield farming.” A protocol issues a token, pays high APY to attract liquidity, the TVL skyrockets, the token price pumps, and everyone feels rich—until the subsidies stop. Then the TVL vanishes, and the APY was never real.
Core: The Numbers That Matter
The key risk flagged in the analysis is blunt: “Whether future revenue growth of AI companies like OpenAI can exceed the growth rate of capital expenditures.” That’s the million-dollar question. But let’s look at the mechanics.
Based on my audit experience from the DeFi yield farming frenzy in 2020, I’ve seen this exact pattern of “capital narrative” before. Back then, Compound and YFI were the darlings. Users borrowed assets to farm governance tokens, and the entire system relied on a continuous inflow of fresh capital. When the inflow slowed, the yields collapsed, and the tokens crashed 90%+.
Now map that onto AI. OpenAI reportedly spends $7 billion annually on compute, but generates only a fraction of that in revenue. The gap is funded by venture capital and, increasingly, by Nvidia’s own financial engineering. Nvidia has the strongest balance sheet in the industry—true. But that strength is now being used as leverage to keep the cycle spinning. The fund manager’s “controllable” is based on Nvidia’s balance sheet today, not on whether the AI revenue machine will ever turn a profit.
Algorithms smell fear, but they respect speed. The speed here is the velocity of capital: Nvidia provides financing, customers order GPUs, Nvidia recognizes revenue now. The fear is what happens when that velocity breaks. In crypto, we learned that exit liquidity is the only thing that matters. When everyone tries to exit at once, the door gets small.
Let’s drill into the hidden layers. The analysis reveals that Nvidia is transferring its client risk from a simple sales exposure to a financial credit exposure. If OpenAI misses a payment, Nvidia doesn’t just lose a sales order—it takes a direct hit on its loan book. That’s a new dimension of risk that the market has not fully priced in. The fund manager says it’s controllable because Nvidia’s cash flow can absorb near-term defaults. But what if the entire AI sector faces a funding winter? Venture capital flows into AI are already showing signs of fatigue. According to Crunchbase, global AI VC funding dropped 15% in Q2 2025 compared to Q1. The cycle is slowing.
Yield is a drug; exit liquidity is the cure. This is my mantra from the DeFi days. Nvidia is essentially acting as the “market maker” for the AI compute hype, providing the liquidity that keeps the narrative alive. But every market maker needs an exit strategy. Nvidia’s exit is its stock price. If the circular financing unravels, the stock will be the first to feel it.
Contrarian: The Blind Spot
Everyone thinks Nvidia’s moat is CUDA or chip performance. The contrarian angle is that the real moat today is financial engineering. AMD and Intel can’t replicate Nvidia’s banking relationships or its willingness to underwrite customer debt. But that also means Nvidia has centralized the risk of the entire AI ecosystem onto its own books.
In crypto, we call that “too big to fail.” We also remember that Terra was too big to fail until it wasn’t. The circular financing creates a feedback loop that can work in both directions. When the narrative is bullish, it supercharges growth. When the narrative turns, it accelerates the crash.
Another blind spot: the fund manager’s statement is self-serving. He likely holds Nvidia shares and has a vested interest in calming the market. His “controllable” is a marketing line, not a risk assessment. In crypto, we know that every bull market ends with someone saying “this time is different.” It never is.
Chaos is just data waiting for a narrative. Right now, the narrative is that Nvidia’s financing is a feature, not a bug. But the data—growing capex, uncertain AI revenue, falling VC flows—is forming a bearish signal. It’s only a matter of time before a new narrative emerges.
Takeaway: What to Watch
The next 6-12 months are critical. Watch OpenAI’s revenue reports. If enterprise subscription growth slows, the entire circular financing model wobbles. Watch Nvidia’s upcoming 10-K for any disclosure of its financing guarantee obligations. If that number comes out large, the market will reprice the stock.
In the crypto world, we’ve seen what happens when liquidity dries up. The floor drops, and the rug gets pulled. Nvidia’s GPU central bank may look robust today, but the accounts are due. And interest is compounding.
We don’t trade markets; we trade narratives. The narrative is shifting. Be ready to exit before the exit liquidity disappears.