Speculative credit in the technology sector gives way, many investors liquidate positions

The most significant event of recent weeks has been the sharp weakening of speculative credit in the technology sector. Bonds across all companies linked to the AI sector have suffered a broad-based contagion after the collapse of Oracle, CoreWeave and Softbank bonds. These events have been joined by spectacular collapses in recent IPOs of technology-related stocks in Korea and Australia. Doubts about the sustainability of “circular finance” - a euphemism used to avoid the term “Ponzi scheme” - invented to support investments in the Datacentre sector, have begun to spread among lenders and credit underwriters. The credit market, unlike equity, does not look to the short term and lenders assess the sustainability of the business over the medium to long term. When such investors liquidate positions, as happened already in 2000 and 2007, the bull market in equity runs into crisis within a few months.

The bonds of the leading companies in the tech sector (Amazon, Google, Meta, etc.) have seen spreads widen. The question that now needs asking is: if the bonds of the strongest companies are widening, what is happening to the speculative credit financing AI capex held by banks, private credit and shadow banking?

After years of public and monetary support, aimed at propping up ad infinitum a colossal speculative bubble that could never have given way, here we are with the stimuli designed to sustain the system’s confidence and appetite for risk no longer sufficient to protect the financial bubble. So, while we are still providing fiscal stimulus of 7.5% of US GDP and an expansionary monetary backdrop, backed by USD 500bn of QE per year, the speculative credit that supports the driver of the economy (AI and Datacentres) is nevertheless giving way, despite Wall Street posting stratospheric, rising profits. The problem is that such profits are partly dependent on the euphemistic “circular finance” and partly due to discretionary revaluations applied to stakes in unlisted companies. Without these accounting tricks, profits would be at marginal growth of zero or even negative. In fact, despite the rise in profits shown in the second quarter of 2026, US stock markets have not materially advanced.

These accounting tricks serve to mask a structurally unsustainable financial position of the investments under way and usually coincide with the peak of the financial cycle. Over the past two years, the great “advertising machine” run by Wall Street’s leading investment banks has served to make the unbelievable seem believable. We have all known for some time now that the balance sheets of the main companies in the technology sector, thanks to a perfectly legal accounting rule, show only revenue and keep costs and debt off balance sheet. We all know perfectly well that if we were to consolidate these off-balance-sheet costs and debts, the market valuations of listed companies would explode to unthinkable levels and the income statement would already be in crisis. The cash flow shown in the accounts, in fact, would already be negative, and the spreads on the bonds financing debt to support capex on AI and Datacentres would already be at junk levels.

Wall Street bankers know full well that between 2027 and 2028, the costs and debts not shown in the accounts will have to be consolidated and that the financial structure of AI and Datacentre capex will suffer a clear downgrading. The accounting rule that allows you not to book costs and debts in fact provides that, when Datacentres become operational, such costs and debts must be booked. It is rather curious that the US financial system can use such a mechanism to “hide” the real costs of an investment, allowing the market to sustain a narrative, mainly created for the retail investor, that is already false from the outset. It is exactly the same mechanism that made it possible, between 2004 and 2007, to extend uncollectable real-estate credit and turn it into “subprime Triple A” securitisations, which then went into default a few years later.

Already during the second half of 2025, Wall Street investment banks began unloading the risks linked to such credit through synthetic derivatives. In addition, private credit funds shifted part of the credit risk into the retail life policies of insurance companies held by private equity. Coincidentally, in these very days the credit linked to those insurance companies has also begun to show widening spreads. History is repeating itself, once again, but I believe that this time the level of institutional complicity is markedly higher than during the 2008 crisis, when everyone accused the rating agencies and the audit firms of turning a blind eye to the truthfulness of the accounts and the credibility of Triple A ratings on subprime mortgages. Despite all the promises of regulation after 2008, nothing concrete has been done and the financial system is more opaque and deregulated than before.

But if a financial crisis breaks out, what kind of dynamics could we expect today compared with 2008? Compared with the past, today we have more speculative credit in the system (45% vs 28% relative to US GDP), more concentration of risk because of investment in ETFs (over USD 10 Tr of passive assets) and more leverage because of many leveraged ETFs. Add to this the current leverage on hedge funds (12 times NAV for Relative Value and “low volatility” strategies), not to mention liquidity risk in the financial market. We also have excess day-to-day speculation through derivatives, and 50% of options expire in 24 hours; we have illiquid assets held by Private Credit and Private Equity officially quantified at USD 11 Tr (equal to 25% of nominal US GDP); and finally we have the buyers of last resort (pension funds and insurance companies) stuffed with illiquid assets that reduce room for manoeuvre and available liquidity.

An example of what can happen in the event of a reversal in risk appetite is Korea and the Kospi index, characterised by high concentration, leverage and illiquidity. After hitting a peak of 9400 points in July, following a correction triggered by the liquidation of speculative positions in semiconductors this summer, the index lost 50% in the space of a month, received support from the government but is still at 5500, after a recovery attempt sparked by government intervention. Numerous positions are still stranded and margin calls have caused losses of the entire capital and more for hundreds of thousands of investors.

If we were to suffer a reversal of the trend after 14 years of bull market, sustained by trying to sell the theory that risk does not exist thanks to the “virtual” put provided by policymakers, very few would be able to sell. The proof of this scenario is confirmed by the legislative measures implemented in recent years on fund redemption mechanisms and by the special rules introduced by the US government to block capital flowing out of the United States in the event of a national emergency. In any case, market liquidity would not allow long positions to be exited and it is highly likely that policymakers would have to introduce a period of “bank holiday”, that is, a formal closure of financial markets.

The problem of market illiquidity would cause a sharp credit crunch in the economy and would prompt the Fed to trigger a colossal QE, markedly larger than that of Covid and 2008. The Shadow Banking System, unfortunately the holder and provider to date of 60% of credit to the economy, would nonetheless remain trapped in further loss-making positions, already significant and not booked. This situation would cause a total freeze in securitisations, which are the basis of the system’s money multiplier. So the Fed’s QE would do little if this mechanism were to disappear.

In 2008 the Fed removed from the system precisely the subprime MBSs that were blocking securitisation from restarting. Today it would have to buy the securitisations of speculative credit from Private Credit, Private Equity, Insurance companies and Banks, so the problem is not confined only to the banking system as in 2008 and is therefore more difficult to resolve. In any case, the resulting defaults would be a much larger percentage than those recorded in the past and the credit crunch would persist until the system had cleared itself of excess financial leverage, which after more than 15 years of QE is decidedly high and would produce a Japanisation scenario.

The monetary rescue intervention might not have immediate effectiveness as in the past, and we could therefore see the same situation still under way in Korea, where investors still need to sell in order to close leveraged positions. Another problem would be the dollar’s holding up, given that the fiscal stimulus currently already under way (7.5% of GDP) would need to be expanded further to 15%-20%, triggering further QE to support the fiscal intervention and thus causing a flight from the Dollar and the crisis of the Dollar Standard.

Gold would open up another pronounced phase of gains together with government bonds of core countries (Germany, US, Japan and China). As the Dollar weakens, the fly to quality would concentrate on JPY and JGBs, CHF, Bunds and Gold. The end of this bull market would clearly put at risk the West’s world leadership and the free movement of global capital, since everyone would try to keep at home the capital needed to support their economies. Deficit countries would be the most penalised (the US at the top of the list), while surplus countries (Japan, China and Germany) would see a repatriation of capital driven both by risk off in the US market and in the Dollar, and by more restrictive rules on capital exports and foreign investment.

The priority would be to support domestic investment and globalisation would suffer yet another serious downsizing. The debt crisis would produce deflation that monetary policies would not be able to resolve easily, both because of the structural crisis of the Shadow Banking System, because of deleveraging needs, and because of the loss of capital and savings in the private sector, which would have repercussions on debt-fuelled consumption, currently equal to 20% of total US consumption. It should be borne in mind that the recovery will depend on how long the system takes to complete the deleveraging. In 2008 deleveraging was quite rapid because the Fed took over the stranded positions of the banking system; today the timing could be much longer because we have much larger and much more widespread positions to liquidate (Banks, Pension Funds, Insurance companies and the Shadow Banking System, in particular Private Credit and Private Equity).

The end of the bull market will also bring an end to financial deregulation and the sector will enter a regulatory environment similar to that of the years following the 1929 crisis. The financial sector will be held responsible before public opinion for the system’s crisis. Even today the main US banks spend millions of USD on the security of their executives and their families, despite living through the most widespread phase of financial exuberance in history. This highlights the very poor reputation enjoyed by the American financial system.

Despite the recent falls in US and EU stock markets, market participants remain positive about the recovery (also because there is no alternative scenario to tell in advance). To support the analysis they point to rising profits, the “year-end rally” and the “January effect”. Apart from the first positive factor, which remains to be verified, it is clear that the other two rely on expectations backed by historical statistics that take it for granted they can recur. The superficiality of such arguments is unconvincing and seems more to reflect the attempt to grant a little more time to the “terminal patient” that is our economic and financial system.

The emerging social instability, the failure of the middle class to perceive that the economy is doing as well as the stock market shows, the need to keep pursuing endless fiscal and monetary stimulus, the doubt that the macro data are true, and wars that no one wants but everyone pays for, expose the system to structural instability in the medium to long term.

What emerges from the growth data, if one wants to go deeper, is a GDP that grows only thanks to stimulus and is no longer able to produce growth autonomously. US GDP generated by the private sector, even with the current AI capex included, currently shows negative growth of 2.7%. This worrying figure tells us that public spending, financed by the Fed’s current QE of USD 500bn a year, is the only factor preventing the economy from falling into recession. Unfortunately, we also know that public debt, like private debt, cannot rise forever. This is the context into which we have poured the largest long position in history on financial markets that are concentrated, illiquid and dangerously ever less transparent; the rest is just narrative.