No Half-Measures
The Credit Strategist - March 2026
The following is an excerpt from the March 2026 issue of The Credit Strategist that discusses, among other issues, the attack on Iran, Private Equity and Private Credit, the CBO report on the federal deficit, and other topics. The full issue is available for paid subscribers. Please consider subscribing today.
Even before the attack on Iran, stocks were under pressure. The Iran military operation is likely to be inflationary by increasing energy prices and raising the federal deficit by increasing military spending significantly. That lowers the odds of the Fed lowering rates, odds that were already pretty low. War is among the most destructive economic forces in the world so the longer the Iran action goes, the worse it will be for stocks. If successful, however, the attack on Iran should be beneficial for the global economy and positive for markets. Investors should look through the attack. Traders will do what traders do.
Software stocks have been leading stocks down while the FAANGs are losing their buzz as well (their huge profits consumed by epic AI spending). AI-leader Anthropic inflicted three stages of grief on software stocks in February. On February 3rd, it announced a legal plug-in for its large language model (LLM) Claude that sent Thomson-Reuters (TRI), LegalZoom.com, Inc. (LZ), and RELX PLC (RELX) down 10% or more. Then on February 20th, Anthropic announced Claude Code Security, sending CrowdStrike Holdings Inc. (CRWD), Cloudflare Inc (NET) and Okta Inc. (OKTA) down sharply. And then on February 27th, the company announced that Claude had the capability of modernizing Cobalt databases (used widely for banking, payroll, ATMs and other government services and run 95% on IBM machines) which sent IBM”s stock to its largest drop in 25 years. Despite its disagreement with termination by the Department of War, Anthropic appears to be taking the lead in important parts of AI competition while sowing fear in the software industry. As discussed below, this is causing serious problems in the corporate credit markets because of massive lending to the software industry in recent years.
Markets were also shaken by a negative reaction to another quarter of blockbuster earnings from NVDA. NVDA has dropped after each of its last four earnings reports, each one better than the other, suggesting it may be hitting a ceiling in terms of investor expectations. But there may be a deeper reason, concern that the company won’t be able to maintain its results once the AI building boom slows. OpenAI lowered its future spending forecast from $1.4 trillion to $600 billion shortly before raising $110 billion from AMZN ($50 billion), NVDA ($30 billion) and Softbank ($30 billion) at a $730 billion valuation (down from a previously projected $850 billion valuation). This was another circular deal that is not what it seems (other than it seems absurd to pour that much money into a company losing tens of billions of dollars a year) as AMZN is paying only $15 billion up front with the rest of the money contingent on reaching certain milestones. META and other hyperscalers are also spreading their AI spending beyond NVDA which may also be tempering enthusiasm for the stock. CoreWeave (CRWV) reported another horrendous quarter, driving its stock down sharply, and investors didn’t believe a single word spoken by its CEO in a long interview on CNBC claiming business is great (he doth protest too much). AI needs to start showing results to impatient investors to prevent further stock market declines. Headlines like Block Inc. (XYZ) cutting half its workforce and firing 4,000 may boost the stocks of companies ostensibly saving money but the other side of those announcements are cuts in software and other products that will hurt the economy and markets. AI is not a zero sum game. There is dark side to any efficiency gains it produces that has to be factored into future forecasts.
