After equities erased $800 billion in market value, the bearish AI trade has become easy to explain: automation destroys jobs, spending weakens, and profits get squeezed across white-collar industries. The Kobeissi Letter argues that this reading is too linear. Its view is that AI is repricing cognition, and that process could end in broader abundance rather than collapse.
The argument starts with a challenge to three assumptions embedded in the doomsday narrative: demand is fixed, productivity gains do not create new markets, and economic systems cannot adapt fast enough to absorb the shock. If those assumptions fail, then the outcome changes. The key question is not whether AI hits office work. It already does. The question is what cheaper cognition does to the size of the economy.
Claude announcements have triggered sharp market repricing
The Kobeissi Letter does not dismiss the damage already visible in stocks. It says Anthropic, through Claude, has been hitting sector after sector, wiping out hundreds of billions of dollars in value from Fortune 500 companies. In its telling, this has become a recurring pattern in 2026: Anthropic launches a tool, Claude shows a real step forward in coding or workflow automation, and affected names sell off within hours.
It points to IBM, which saw its worst drop since October 2000 after Anthropic said Claude could simplify COBOL code. Adobe is down 30% year to date, with generative AI seen as compressing the value of creative workflows. Cybersecurity names were also hit. In one example, CrowdStrike sold off almost immediately after the release of “Claude Code Security.” The tool launched at 1 p.m. Eastern on February 20, and CrowdStrike lost $20 billion in market value over the next two trading days.
That reaction, the article says, is not irrational. Markets are trying to price a first-order effect: when AI can replicate parts of a worker’s output, pricing power shifts from the seller to the buyer. Margins come under pressure. That is real.
Commoditized cognition does not automatically mean collapse
Where The Kobeissi Letter breaks from the bearish case is on what comes next. Commoditization, in its view, is not the same as economic failure. Personal computers commoditized computing, the internet commoditized distribution, cloud infrastructure commoditized compute at scale. AI, it argues, is now commoditizing cognition itself.
History matters here. When the cost of producing something falls sharply, demand rarely stays flat. The article notes that lower computing costs did not lead people to consume the same amount of compute at a cheaper price. Consumption expanded by orders of magnitude, and entire industries were built on top of that shift. It also highlights that personal computers are now 99.9% cheaper than they were in 1980.
That is why the piece says investors may be selling the layoff story while missing the larger one: service-price compression. A wide range of knowledge work remains expensive because trained human attention is scarce. If AI lowers the marginal cost of that attention, the price of those services can fall as well.
From “Ghost GDP” to “Abundance GDP”
The article frames the divide as “Ghost GDP” versus “Abundance GDP.” Ghost GDP describes output that looks strong on paper but does little for household living standards. Abundance GDP is different. Output rises while the cost of living falls.
By that standard, the critical test is not whether nominal wages surge. It is whether prices fall faster than incomes. If AI cuts the cost of essential services, real household gains can still increase even if wage growth slows. The Kobeissi Letter argues that this is how productivity can show up in lived experience: not only through higher pay, but through lower prices.
It ties that thesis to the structure of the US economy. Services account for nearly 80% of GDP. If operating costs drop, smaller firms become easier to sustain. If access costs fall, more households can participate in economic activity. The article describes that effect as a kind of invisible tax cut. Firms built on scarce, high-cost cognitive labor may lose pricing power, but a broader set of consumers and businesses could benefit from lower service inflation and stronger real purchasing power.
The piece also notes that productivity has outpaced wage growth over the last 70-plus years. Electricity, the internet, mass manufacturing, and antibiotics were all disruptive, yet each lowered costs and raised living standards over time. Its point is not that AI guarantees the same path. It is that disruption at the firm level does not settle the macroeconomic outcome.
White-collar pressure, SaaS reshaping, and lower-friction commerce
The Kobeissi Letter accepts that white-collar employment pressure is a serious concern, especially because those workers support discretionary spending and housing demand. Still, it argues that AI remains limited in physical dexterity and in fields tied closely to human identity. Skilled trades, hands-on care, advanced manufacturing, and experience-driven work retain structural demand. In many settings, AI looks more like an amplifier than a full replacement.
Its view on SaaS is similar. The model faces pressure, especially as buyers gain leverage and long-tail software products run into structural resistance, but the article says SaaS is a delivery method, not the endpoint of value creation. The next wave of software, in this framing, will be adaptive, agent-driven, and built around outcomes. Margin compression in one layer does not mean the entire digital economy is breaking down.
The same logic is extended to agentic commerce. Lower friction can reduce the ability to extract fees, but it can also increase total transaction volume. The article notes that stablecoin volumes were already surging before AI reached its current stage, for a simple reason: markets prefer efficiency. Better price discovery and lower transaction costs can bring more economic activity onto the rails.
Productivity remains the real scorecard
For The Kobeissi Letter, the ultimate measure is productivity, not the initial equity selloff. If AI delivers sustained gains in healthcare, public administration, logistics, manufacturing, and energy optimization, the long-run result could be lower barriers to entry and wider abundance. Even a steady 1% to 2% increase in annual productivity growth would compound meaningfully over 10 years. The article points to a rise in US labor productivity in the third quarter of 2025, describing it as the strongest pace in two years.
Its conclusion is direct. What Anthropic has triggered is a repricing of workflows and a repricing of cognitive labor. That is a transition. It does not have to be collapse.