On November 19, 2025, Nvidia delivered what looked like another coronation. Quarterly revenue soared past $57 billion, data center sales alone exceeded $51 billion, and guidance for the coming quarter pointed to yet another record. Jensen Huang spoke of insane demand, half-trillion-dollar backlogs, and a multi-trillion-dollar build-out of AI infrastructure stretching to the end of the decade. For a few hours, the market rewarded him with a 5% spike.
Then the algorithms read the footnotes. Within eighteen hours the gain vanished. By the close of November 22, Nvidia had surrendered nearly everything it briefly won, and the broader technology sector bled with it. What the machines saw and what most humans are only now beginning to grasp is that the most celebrated growth story of our era may rest on a foundation far shakier than anyone dared admit.
A single viral essay, published under the headline “The $610 Billion AI Ponzi Scheme Just Collapsed,” crystallized the fear. Its author, independent analyst Shanaka Anslem Perera, argued that Nvidia’s headline numbers are being artificially propped up by a closed loop of credit, commitments, and circular funding among a handful of mega-cap players. The piece spread like wildfire across X, Threads, and private trading channels, racking up tens of millions of impressions in days.
This is the most exhaustive, clear-eyed examination of that thesis yet produced.
The Balance Sheet Red Flags, Plainly Stated
Accounts receivable now stand at $33.4 billion, up 45% in a single quarter and nearly 90% in a year. That is not trivial. It equals almost 60% of the revenue Nvidia recorded in the same period. The average payment term has stretched from 46 days to 53 days. In a company growing this fast, a one-week delay translates into billions of dollars of cash that have not yet arrived and may never arrive in full. Two customers, widely understood to be Microsoft and Meta, account for a third of the entire receivable balance.
Inventory is also climbing faster than sales can clear it. Days of inventory outstanding rose from 106 to 119 in one quarter. Nvidia insists this is deliberate buildup ahead of the Blackwell ramp, but the trend is unmistakable: chips are leaving factories faster than they are being installed and paid for.
Cash flow, the ultimate reality check, tells a similarly uncomfortable story. Nvidia generated $23.8 billion in operating cash over the past quarter, a colossal figure until set beside non-GAAP profit of roughly $19.3 billion in some earlier interpretations (or $16.6 billion under strict GAAP). In healthy semiconductor companies such as TSMC and AMD, cash conversion routinely exceeds 95% of reported profit. Nvidia’s ratio, once adjusted for working-capital swings, has dipped into the mid-70s in recent quarters. That gap is the space where doubt now lives.
The Great Circular Funding Machine
The most disturbing charge is that much of Nvidia’s demand is not organic but manufactured through a daisy chain of -party transactions.
Nvidia invests $2 billion in xAI. xAI raises $12.5 billion in debt and immediately spends the bulk of it on Nvidia GPUs. Microsoft pours tens of billions into OpenAI. OpenAI commits $250 billion to Microsoft Azure over the coming six years. Microsoft turns around and orders tens of billions more Nvidia silicon to power that same cloud. Oracle hands OpenAI $300 billion in future cloud credits starting in 2027, credits that will be consumed on servers stuffed with Nvidia chips.
The same dollars travel in an endless circle, counted as fresh revenue at every stop. The chips are shipped, the invoices are booked, but the actual cash often remains a promise rather than a fact. This is not fraud in the legal sense, but it is financial alchemy, and alchemy has a tendency to turn back into lead when confidence falters.
OpenAI itself is the clearest exhibit. Revenue is growing fast, yet the company is on pace to lose more than $16 billion in 2025 alone. Its current $300 billion private valuation implies it must one day generate cumulative profits measured in the trillions. MIT research published this year found that 95% of enterprise generative-AI projects fail to deliver positive return on investment. Arithmetic and history are not on the same side here.
The Smart Money Quietly Heads for the Exits
Peter Thiel’s family office disclosed a complete sale of its Nvidia stake in the third quarter, walking away from a position that had been worth well over $100 million at the peak. SoftBank unloaded $5.8 billion of shares in October and November to redeploy capital into OpenAI and other “upper-layer” bets. Michael Burry’s latest 13F reveals more than a billion dollars in put options betting on a sharp decline in Nvidia by early 2026.
Bitcoin, the speculative thermometer most closely cor with AI euphoria, has already fallen 25% from its October high near $111,000 and now trades in the low $83,000s. Dozens of AI startups have used Bitcoin as loan collateral; a further 40% drop in Nvidia could trigger forced sales that push the cryptocurrency below $50,000 and create a second-order shock across the entire risk-asset complex.
Ponzi Scheme or Second Dot-Com? The Crucial Distinction
A true Ponzi scheme pays early investors with money from later ones and collapses the moment recruitment slows. Nvidia is not that. It ships real silicon that performs real computations for real (if currently unprofitable) customers. The company retains a fortress balance sheet, negligible debt, and a near-monopoly on the hardware that powers the frontier of AI.
What it does resemble is the late-stage dot-com boom of 1999: a transformative technology surrounded by astronomical valuations, circular capital flows, and a collective willingness to suspend disbelief about near-term profitability. In that earlier episode the shovel makers (Cisco, Sun, EMC) were the last to fall, but fall they eventually did when capital expenditure budgets were slashed.
Nvidia today is the undisputed shovel maker of the AI era. Its gross margins exceed 75%, its market share in high-end training GPUs sits above 90%, and its order backlog remains enormous. Those facts are genuine. The open question is whether the customers who placed those orders can ever turn their models into businesses that pay the electric bill, let alone repay the billions they have borrowed or been gifted.
Valuation in the Crosshairs
Consensus analyst targets still cluster between $240 and $250 per share, implying 30-40% upside from today’s price near $180. Morningstar recently argued that even after the pullback Nvidia remains undervalued relative to a realistic five-year growth path. Bears, citing credit risk and potential inventory writedowns, place fair value as low as $110-$140. The extreme collapse scenario that yields $71 requires not just a recession but a near-total evaporation of enterprise AI spending, an outcome that feels improbable given sovereign wealth funds, defense budgets, and scientific computing demand that exist independent of Silicon Valley startups.
The Path Forward: Turbulence, Then Clarity
The next ninety days will tell us a great deal. February’s fourth-quarter report will reveal how many of those receivables have aged past 60 or 90 days. March may bring the first credit-rating reviews. April could force restatements if allowances for doubtful accounts prove too optimistic.
None of this guarantees catastrophe. The AI build-out may simply shift from manic to merely aggressive, from hundreds of billions per year to a still-immense hundreds of billions sustained over a longer horizon. Power grids, chip fabrication capacity, and actual enterprise ROI will impose their own hard limits, and the market will price the new reality long before every data center is fully utilized.
What is certain is that the era of unquestioned exuberance has ended. The tourists who chased triple-digit returns on hype alone are being separated from the builders who intend to ship useful intelligence for decades. Nvidia will not disappear, but its multiple is unlikely to reflate until the circular funding games give way to old-fashioned cash profits earned from paying customers.The silicon empire has developed deep cracks. Whether it merely settles onto a stronger foundation or suffers a more violent reconfiguration will define wealth creation, technological progress, and economic power for the remainder of the decade. One way or another, the reckoning has begun.
Lessons Emerging
The Nvidia saga offers stark reminders for all market participants. Corporates must prioritize sustainable innovation over hype-driven expansions, ensuring that AI investments yield tangible ROI rather than relying on circular financing or speculative valuations that could evaporate overnight. Bourses and regulators should enhance transparency around -party transactions and credit extensions in high-growth sectors to prevent systemic vulnerabilities, perhaps mandating stress tests for AI-dependent listings. Banks, as lenders to this ecosystem, need to tighten risk assessments on collateral like Bitcoin or unproven tech assets, bracing for potential defaults in a downturn while diversifying away from overexposure to Silicon Valley darlings. For common investors, the volatility underscores the perils of chasing momentum without due diligence diversify portfolios beyond tech megacaps, set stop-losses amid bubble fears, and favor long-term fundamentals over short-term euphoria, as evidenced by JPMorgan’s warnings of an impending AI correction and the broader economic angst rippling through credit markets.
India, with its burgeoning tech sector and IT services giants like TCS and Infosys deeply intertwined with global AI supply chains, stands at a crossroads. The potential burst of the AI bubble could trigger a domino effect on Indian IT stocks, which have ridden the wave of offshore AI development but face revenue squeezes if Western clients slash capex amid profitability doubts as experts note, a 20-30% drop in US tech spending might shave 10-15% off Indian IT earnings in 2026. Policymakers should accelerate domestic AI infrastructure builds, leveraging initiatives like the India AI Mission to reduce dependency on foreign hyperscalers and foster homegrown startups resilient to global shocks. Investors in India must hedge against rupee volatility tied to Nasdaq corrections, prioritizing diversified mutual funds over direct US tech exposure, while the NSE and BSE could introduce AI-specific indices with built-in volatility buffers to educate and protect retail participants from the kind of rout seen in Asian markets this week.
Overall Assessment and the Likely Future
In sum, the AI frenzy epitomized by Nvidia’s rollercoaster is not a outright Ponzi but a classic overextension, where trillion-dollar promises collide with prosaic realities like ROI shortfalls and power constraints yet the technology’s transformative potential endures, with sovereign funds and enterprises likely sustaining a moderated build-out. Looking ahead, expect a 2026 shakeout: corrections in AI stocks by 20-40% as unprofitable ventures fold, but a rebound by 2027-2028 as mature applications in healthcare, finance, and automation deliver verifiable value, potentially adding $15-20 trillion to global GDP by 2030 if bubbles give way to balanced growth. The future hinges on adaptive regulation, ethical scaling, and inclusive innovation, turning today’s tremors into tomorrow’s stable foundation for an intelligence-augmented world.
(Major General Dr. Dilawar Singh, IAV, is a distinguished strategist having held senior positions in technology, defence, and corporate governance. He serves on global boards and advises on leadership, emerging technologies, and strategic affairs, with a focus on aligning India’s interests in the evolving global technological order.)
