GROWTH WITHOUT YOU (Guest: Paul Krake)

Paul Krake argues that AI and technology, not traditional macro policy, are the primary drivers of long-term economic change. He builds this case on a T. Rowe Price study showing US corporate margins rose from roughly 1 percent in 1975 to nearly 15 percent at the time of writing, with only 130 basis points of that expansion attributable to lower interest rates and lower taxes. Everything beyond that 130 basis points came from innovation including container ship standardization, globalization, enterprise software, cloud computing, and mobile. He extends this to argue that economic volatility was already declining before QE, that Japan has printed money for 35 years without generating inflation, and that the US has not had an organic business cycle recession excluding Covid in 17 years. He attributes the reduction in boom-bust cycles to technology rather than better policy management, while acknowledging the counterfactual cannot be tested.

Krake draws on Robert Gordon of Northwestern to argue that the 1870 to 1940 period was the greatest era of innovation the world has ever seen, with electricity, combustion engines, and indoor plumbing more transformative than the internet or computerization. Indoor plumbing alone raised life expectancies by approximately 10 years over roughly 40 years. He stresses that productivity and innovation must not be conflated, citing Robert Solow's observation that computerization is everywhere except in the productivity statistics. Encyclopedia Britannica was a 250 million dollar business that went to zero when replaced by Wikipedia, which contributes effectively nothing to GDP, illustrating how innovation can destroy measured economic output even while improving welfare.

Krake identifies human cognition as the last remaining layer of economic friction, and AI as cognition as a service that replaces expensive human cognition with cheap artificial cognition. He predicts US unemployment will reach 10 percent by 2035 driven by AI displacement of knowledge workers, who represent approximately 45 percent of the US workforce or roughly 75 million people. Goldman Sachs estimates AI is currently eliminating approximately 20,000 jobs per month. Krake predicts unemployment will follow a nonlinear path, reaching roughly 5.5 percent by end of 2027 before accelerating to losses of 80,000 jobs per month as adoption broadens. He argues no replacement industry exists today comparable to the manufacturing boom of 1910 to 1920, which itself was driven by immigration rather than workers leaving farms. The US labor pool is at best flat going forward due to poor demographics and restricted immigration, limiting AI-driven growth.

On markets, Krake predicts a five-year equity boom driven not by the Magnificent Seven but by everything else, as the narrative shifts from AI providers to AI beneficiaries. He estimates companies in US small and mid cap, European mid cap, and emerging markets could add 25 to 75 basis points of margin per year for five to seven years, potentially making many non-tech companies two, three, or four baggers without significant revenue growth. He favors Germany, France, Korea, Singapore, and Brazil. He is specifically bullish on enterprise software vendors including SAP, Oracle, and Salesforce because AI will be embedded into their existing platforms and switching costs are prohibitive, with one CTO estimating a beta testing period of roughly two and a half years to deploy Salesforce. He flags Oracle as the most likely candidate to announce stress from approaching negative free cash flow, which he identifies as his biggest downside surprise for the year.

Krake warns that AI investment is the most broadly distributed investment narrative he has ever seen, embedded in indices and leveraged ETFs, and that an earnings miss could produce a decline of 20 percent or greater. Earnings expectations were not walked down last quarter as is typical, meaning the market now requires very good results just to stay afloat. IBM lost approximately 25 percent of its market value in a single day after a rare pre-announcement of bad results. One-year implied S&P correlation was trading around 12 at the time of recording, which Krake described as signaling extreme complacency. Goldman Sachs estimates CTA sell triggers for the S&P were closing in on 7,400 but had not been triggered at the time of recording.

Krake's outlier prediction is that Chinese LLMs get banned in the United States, drawing on the precedents of DJI drones, ZTE, Huawei, and the anticipated eventual ban on BYD. He considers such a ban market-positive because it would remove the narrative that hyperscalers are overspending relative to cheap Chinese alternatives. On fiscal structure, he notes that approximately 88 percent of US tax revenue comes from labor payments while corporations represent only about 9 percent of the tax base, and that the roughly 7.5 percent employer-side FICA payroll tax actively incentivizes replacing workers with software. He proposes taxing corporations at the same effective rate as individuals and argues a globally coordinated corporate tax agreement is necessary because every country will face the same AI-driven disruption.

Crude oil large speculator positioning fell to the 19th percentile after the late June washout, leaving significant room for speculators to rebuild long positions. Krake argues Trump tweets attempting to talk down oil prices are removing the natural long speculator cushion, making a supply-driven spike more dangerous. He expects the Strategic Petroleum Reserve will eventually need to be refilled, creating a surprise to the upside. Gold was described as outright distributive with every rally met with supply, approximately 1,600 dollars off its high, with the 50 percent retracement on the monthly chart sitting at 3,600.