All Eyes on the Credit Markets

November was a very interesting month as far as financial markets are concerned. The US government shutdown that started October 1^st lasted 43 days and extended into November, ending on the 12^th. This means no official government data for September and October were collected or released to the public. For all practical purposes, the investment community was flying blind for the better part of two months. The only available data were corporate earnings releases and privately sourced economic data estimates. The corporate earnings (except big tech) painted a picture of exhausted consumers that cannot absorb high prices anymore. (We will expand below on what seems to be suspicious reporting from the tech sector). The privately sourced economic data confirmed that the US economy has weakened, the previously resilient labor market is no more, and consumers are hurting. For instance:

• ADP reported that October job losses were around 50,000 (averaging 13,000 per week).
• Challenger, Gray, and Christmas reported 153,000 US October announced layoffs. Examples include Verizon eliminating 15,000 positions, Hewlett Packard 6,000 positions, and Apple unspecified sales positions.
• The University of Michigan Consumer Sentiment Index fell in November to 50.3 below October’s 53.6 making it the second lowest on record. This seems in line with the 6 million US households which are currently delinquent on their utility bills, and the 20% year-over-year increase in foreclosures (which were discouraged by the previous administration by extending the COVID era foreclosures moratorium).
• The Port of Los Angeles reported a drop in traffic of 17.6%, while the Port of Long Beach reported a drop of 20%. This indicates that businesses ordered less inventory as exhausted consumers cannot spend at the same rate as before.
• There were 655 business bankruptcy filings for the first ten months of the year, which indicates that at this pace they will exceed the 2024 filings by 15%. Aside from people losing their jobs and income, business bankruptcies bring loan defaults, and given their size, become the problem of the creditor. Some sizeable ones are Bayside Credit (subprime auto loans), First Brands (which we wrote about in our previous newsletter), Renovo (real estate loans), Sonder (apartment rentals), Paradise Valley (Ritz Carlton developer), United Sites Services (portable toilets), 48forty Solutions (pallets), Zips (carwashes). Private Credit (PC) funds are exposed to billions of dollars in losses from these loan defaults mentioned above..

Given these news releases, it is perhaps not surprising that during November the stock market experienced a small correction. The S&P 500 index declined by 6% while the NASD 100 index declined by 9%. What may be surprising is that the correction was so shallow that by the end of the month the markets had almost fully recovered. On the other hand, it may not be surprising as investors have been accustomed to interpreting bad news as good news, because they precipitate FED interest rate cuts. How can you blame them for buying every dip, when the FED (in its previous actions) has trained them to act like Pavlovian dogs?

What is interesting is that the tech heavy NASD 100 declined by 50% more than the S&P 500. Is it possible that investors are becoming disillusioned with the Artificial Intelligence (AI) trade? We think YES, it is possible, and it’s about time for the stock market to function more like a weighing machine of earnings (like it is supposed to be) and less like the sentiment indicator that it has been for the past three years.

The signs of disillusion are there to see. For instance: after Nvidia (NVDA) reported quarterly earnings on November 19^th with the usual fanfare of bullish statements like “$2 billion beat on revenue” expectations, “off the charts” demand for its processors (chips), and “sold out” of chips, the stock price still declined. This time around analysts actually looked at the financial statements accompanying the quarterly report. Maybe they were alerted by so many voices like ours, objecting to the obscene NVDA projections and proclamations about the future, that they were forced to do a better job analyzing the reporting instead of simply repeating it verbatim to the public.

A cursory review of the quarterly statements revealed a $5 billion increase in inventory (up to $19 billion) from the previous quarter. But how can NVDA be “sold out” of chips if their inventory increased by 36% in three months’ time? Similarly, Accounts Receivable increased by $5.8 billion (to $33 billion) over the same three-month period. Was NVDA’s $2 billion revenue beat even real if Accounts Receivable increased three times over??? Anyone can sell more stuff if they don’t ask to get paid back any time soon. How could this be reconciled with the “off the charts” demand? If that was remotely true, why were there $1.4 trillion “commitments” to buy back the data center services from its customers? The only reasonable explanation is that the Hyperscalers operating these data centers with NVDA chips cannot find enough paying customers to raise the revenue to repay NVDA for the chips they bought. Hence, the only way NVDA can avoid writing off its Accounts Receivable as a loss is by guaranteeing to pay its customers so they can repay NVDA. In other words, NVDA will pay itself!?!? And we’re talking about the most “valuable” company in the world here…

Alas, the analysts are finally paying attention. The AI investment machine is based on “Roundtripping” and accounting shenanigans. The same dollars are recorded as revenue by several companies sequentially, and the total is added to the GDP as output. It is complete nonsense. For instance:

• Microsoft invests in Open AI in order to use Chat GPT. With that money, Open AI then buys Microsoft cloud services to run Chat GPT. In essence, doesn’t Microsoft pay for its own revenue in this situation?
• NVDA invests in Open AI. With that money, Open AI buys Oracle cloud services to run Chat GPT. With that same money, Oracle buys NVDA chips for its data centers. Again, doesn’t NVDA pay for its own revenue in this example? For all practical purposes, since the money originated with NVDA and ended up back with NVDA, the money doesn’t even have to leave NVDA at all for the “transactions” with Open AI, Oracle, and NVDA to each book that money as “revenue”. Notice that these are accounting entries that are not dependent on anything useful to ever be produced.

If you have a hard time wrapping your head around this concept, consider the following hypothetical transaction: ABC Co. hires XYZ Co. to dig a ditch for $50,000. Subsequently, XYZ Co. hires ABC Co. to fill up this ditch for $50,000. Nothing of value was ever produced, yet both ABC and XYZ companies will book each $50,000 in revenue and the GDP will show an additional $100,000 of “output”!!

This is the amazing AI investment machine that has investors super excited, the stock market currently at record high levels, and on paper the US economy appearing to be out of a recession. However, excluding the AI spending, the US GDP growth rate for the first three quarters of the year is 0.1%. And if you adjust for inflation, the growth rate would turn negative (hence, a recession).

The amazing AI investment machine works until analysts and investors start questioning the numbers, and they have started to do so. The Financial Times finally posted a story about Open AI needing to raise $207 billion in the next five years in order to continue losing money (HSBC estimate). Analysts are now questioning how Open AI can make $1.4 trillion in “commitments” to buy data center services from hyperscalers, when it has annual revenue of only $13 billion, while losing $12 billion during the third quarter alone?

It is always interesting to see how financial bubbles evolve. A few weeks ago, no one wanted to see the abysmal losses such AI investments would inflict on naïve investors. This month however, they are eager to observe that NVDA is pouring funds into money-losing companies (X-AI, Open AI, Coreweave, Nebius) to boost its own revenue and coined the term “circular finance” to describe the act of “round-tripping” like it is a new development. They also noticed that another hyperscaler, MSCI, reported a 15% decline in revenues and 61% decline in earnings. Then, C3AI put itself up for sale after disappointing performance and CEO exit. In our opinion, none of this says “off the charts” demand for the sector.

Speaking of abysmal losses, while the stock market is still dancing at the party, oblivious to the 655 bankruptcy casualties we listed above, the credit market is absorbing losses from loan defaults. Private Credit funds in particular are at the forefront, as they typically lend to companies that commercial banks would not lend to, either because of poor credit quality or large loan size. These Private Credit funds raise money from wealthy investors, foundations, endowments, and other institutions under the premise that they can generate superior returns by investing (or lending) to “opportunities” not publicly available. Of course, higher returns are promised for taking higher risks but since they are “Private” funds they do not have to mark-to-market their loans, so investors never get to see the actual losses, until they try to get their funds out. Then of course it is too late as the Private Credit funds stop redemptions not to show further losses, a practice called “gating” which traps investors’ funds. Fierce Financial Group does not invest in Private Equity or Private Credit funds because they are not transparent, they are not liquid, and common sense tells us that the “opportunities” of higher potential returns are simply junk holdings and loans to high-risk entities. Nevertheless, we pay close attention to their losses because their excessive risk taking has the potential to freeze all credit and possibly blow up the credit markets. Recently, there has been a flurry of losses taken by the Private Credit and Private Equity sector that is very worrisome. For instance:

• Blackrock had to write down (take a complete loss) the $150 million loan to Renovo, which just a few weeks ago had reported in its books at full value at 100 cents on the dollar! Didn’t they see the bankruptcy filing coming? If they didn’t see that one, how many more haven’t they seen?
• Apolo and KKR reported the same debt from a borrowing firm (Medallia) at widely different valuations. Apolo at $0.77 on the dollar, while KKR at $0.91 on the dollar. In other words, in the world of Private Credit, no one knows the true condition of the borrowers, not even their lenders.
• Fortress, Ares Capital, and Blackstone must absorb the $1.4 billion loan default from the United Sites Services bankruptcy.
• Atlantic Street Capital must absorb at least $654 million loss from the ZIP Carwash bankruptcy.
• UBS is closing two of its O’Connor hedge funds due to credit losses and high redemption requests.
• Blue Owl tried to merge two of its credit funds but received great backlash from their investors as the fund that would remain has 22% loss of capital.

There are too many such losses to list here. The point is however, that when loan defaults start, lenders become less willing to lend, and institutional investors abruptly pull their funds out of these Private Credit funds exacerbating the credit crunch.

The credit market is waking up to the abysmal losses lying ahead due to the irresponsible bubble market euphoria lending. And lenders are trying to hedge their exposures by purchasing insurance for loan losses. This insurance comes in the form of Credit Default Swaps (CDS) that became famous during the Great Financial Crisis of 2008. The cost to insure loans from the tech sector has recently jumped substantially. For example, the cost of a CDS on Coreweave bonds last week was 7.4%. How much yield must be offered to a lender if the lender incurs a 7.4% cost to insure the loan? Yields move higher and those higher yields raise the odds that the loans/bonds will end up in default.

Defaults create a vicious credit tightening cycle which can quickly bring markets to their knees. Banks can stop lending or even pull existing credit lines until they have better visibility. That leads to more defaults and more tightening. Investors are currently confident that a couple of FED rate cuts will do the trick, and liquidity will flow again into the system. We are not so sure. We believe a rate cut or two are just Band-Aids. Given the size of the Private Credit assets and that banks have already refused to lend to each other (REPO market is drained), the FED has to do more to inject liquidity into the system. A simple way will be to eliminate the Supplemental Liquidity Reserve Ratio from commercial banks, which could potentially release $2.6 trillion of commercial bank lending.

Investors should pay attention to FED announcements over the next couple of weeks, looking for signs that the FED is aware of the situation and is willing to take action to inject liquidity into the system. The next FED meeting is on December 9^th but the FED could announce something even sooner (like a new emergency Reverse REPO facility etc). Closer to the Holidays and toward the end of the year, markets tend to float higher (Santa rally). In our opinion, this year a Santa rally is conditional on the Fed providing ample liquidity to the markets. We prefer to avoid dangerously overvalued AI related stocks and bonds. We also strongly suggest avoiding Private Credit, Private Equity, and commercial bank stocks. We like quality companies at a good valuation, and some cash on the sidelines for taking advantage of potentially better asset prices in the future…

Happy Holidays!

Stay Fierce!