THIS WEEK · 10 PODCASTS · WEEK OF AUGUST 3, 2026
This week we bring you ten podcasts circling the same fault line from ten different angles — an AI capital cycle whose financing is beginning to creak, a Fed that may have to hike into a war, and a bond market no longer willing to be the shock absorber for either. Julien Garran, Ed Zitron and Nobody Special Finance open the edition with the most forensic bear case available: OpenAI as the load-bearing beam under the entire trade, hyperscaler net income mathematically destroyed by depreciation, and misallocated capital running at 70% of GDP. Jim Bianco supplies the mirror image — an economy he thinks is fine, an AI transition he believes will genuinely eat the software stack, and a bond market telling Kevin Warsh he should be thinking about hiking rather than cutting. David Nicoski brings the tape: a genuine rotation out of a technology sector that is 42% of the index, Korea as the leading indicator nobody wants to look at, and a US intervention to buy yen for the first time since 1998. Barry Knapp explains the K-shaped economy as a policy artifact rather than a social accident, then puts money behind it by cutting technology to 25% against a 37% index weight. Luke Gromen argues the Treasury market has been quietly handed to leveraged hedge funds, and that gold rather than the Treasury bond is becoming the world’s backstop asset. Chris Irons expects a sharp deleveraging, a Fed that folds, and gold at $7,500 on the other side of it. John Rubino traces the transmission channel — private credit into AI, big banks into private credit — even as the gold miners quietly become the healthiest sector in the market. Robert Pape supplies the geopolitical backbone with a piece of arithmetic markets have not priced: the drone launch zone that must be suppressed to reopen Hormuz is one and a half times the size of California. Rupert Mitchell turns all of it into positions — long equal weight against the NASDAQ 100, energy equities underwritten by a Chinese collar on crude, and deep value in the UK, Uzbekistan and Turkey. And Garey Aitken closes the edition with the unglamorous discipline that ties the week together: three beaten-down compounders bought near half price, and a framework for telling a broken story from a broken business. Each summary is designed to be immediately actionable — whether you are allocating capital, running a business, or simply trying to understand the forces reshaping the world around you.
THIS WEEK’S LINEUP
EP 1 AI: The Wheels Are Coming Off — Julien Garran, Ed Zitron & Nobody Special Finance — MacroStrategy Partnership; EZPR Founder & “Better Offline” Host; Independent Analyst — Watch Full Video
EP 2 The Fed Is Trapped — Jim Bianco — Bianco Research President — Watch Full Video
EP 3 The US Bails Out Japan — Are US Stocks Next? — David Nicoski — Vermilion Research CIO — Watch Full Video
EP 4 I Cut Tech. Here’s Why. — Barry Knapp — Ironsides Macroeconomics Founder — Watch Full Video
EP 5 Equities Extremely Complacent — De-Lever and Prepare to Buy the Dip — Luke Gromen — Forest for the Trees Founder — Watch Full Video
EP 6 Something Will Break: Sharp Deleveraging, the Fed Folds, Gold to $7,500 — Chris Irons — QTR’s Fringe Finance Author — Watch Full Video
EP 7 The Next Big Short Is Taking Shape — John Rubino — Rubino Substack Author — Watch Full Video
EP 8 You Can’t Bomb Iran Into Surrender — Robert Pape — University of Chicago Professor — Watch Full Video
EP 9 He Called It the Worst Chart Imaginable. Then He Bought It. — Rupert Mitchell — Blind Squirrel Macro Author — Watch Full Video
EP 10 Avoid the Value Traps: Three Beaten-Down Stocks That Can Win Big — Garey Aitken — ClearBridge Investments Head of Canadian Equities — Watch Full Video
Full summaries with actionable insights and investment focus for each podcast follow on the pages below.
EP 1 - AI: The Wheels Are Coming Off
Julien Garran, Ed Zitron & Nobody Special Finance — MacroStrategy Partnership; EZPR Founder & CEO, “Better Offline” Host; Independent Analyst
Julien Garran, Ed Zitron and Nobody Special Finance, hosted by George Noble, deliver the week’s most complete structural bear case on the AI complex — and, unusually, they build it from three independent directions that converge on the same conclusion. Zitron works the accounting and the flow of funds, showing that a single private company underwrites the revenue growth of most of the listed AI trade. Garran works the depreciation schedule and the capital-allocation math, arriving at a number that dwarfs both the dot-com peak and the eve of the financial crisis. Nobody Special Finance works the corporate structures and the analyst community that refused to examine them. The shared conclusion is that this is not a valuation problem to be corrected but a financing problem to be discovered.
Actionable Bullet Points
OpenAI Is the Load-Bearing Beam — Remove It and the Floors Above Collapse: Zitron’s central claim is that one unprofitable private company is the demand base for the entire listed complex. In calendar 2025 OpenAI spent $17.2 billion with Microsoft Azure, which he calculates as 69% of Microsoft’s year-over-year growth for that period; strip it out and the residual growth was roughly 8%, barely ahead of inflation (3:15). The same customer justifies effectively all of Oracle’s capex, a large share of CoreWeave’s IPO story, and by proxy a meaningful slice of the neocloud revenue base (3:35). Contract exposure is enormous and disclosed piecemeal — roughly $138 billion at Amazon, $250–280 billion of remaining performance obligations at Microsoft, and $300 billion-plus at Oracle (56:56). Map your holdings by OpenAI dependency, not by sector label.
The Funding Chain Runs on Venture Capital, and the IPO Window Is the Whole Question: Zitron argues OpenAI has announced roughly $122 billion of capital but has likely received something closer to $60–70 billion, with tranches from SoftBank and Nvidia either partial or unconfirmed (6:13). Microsoft only keeps booking that Azure revenue for as long as OpenAI keeps being fed venture money, which makes the listing the load-bearing event rather than a milestone. His flat prediction: if OpenAI cannot go public within roughly eight months it is finished, and SoftBank is materially threatened alongside it, having already sold down Nvidia and most of its T-Mobile stake and approached the ceiling on its ARM margin loans (55:13). Garran adds the tell that the credit side has already moved — banks that were dancing near the door three months ago now appear to be walking out of it, with SoftBank unable to secure a $10 billion bridge against its OpenAI stake (42:02). Treat funding access, not earnings, as the timing signal.
The Depreciation Schedule Destroys Hyperscaler Earnings — and They Cannot Build Their Way Out: Garran ran a deliberately conservative schedule against roughly $1.5 trillion already spent and a projected $1 trillion a year over the next six years, using the companies’ own assumptions including a six-year chip life he personally believes is too generous given the thermal load (10:51). On that basis, hyperscaler net income falls 98% by 2033 — earnings entirely consumed by the capex that was supposed to create them (12:20). To break even instead, the industry would have to recreate roughly twenty killer applications — another search, another Office, another AWS, another marketplace — inside six or seven years (12:32). Zitron notes the accounting response has already begun: Microsoft extended the useful life of its data center buildings from 15 to 25 years, which suppresses depreciation and inflates operating income (9:45). Watch useful-life assumptions and capitalization policy as the earliest evidence of strain.
Productivity Gains Are Real at the Task Level and Absent at the Commercial Level: Garran cites recent National Bureau of Economic Research work showing coding output rising roughly 280% with AI tooling, projects completed up only about 80%, and applications actually shipped up around 30% (15:16). Downloads and usage, meanwhile, are flat — supply up 30%, demand unchanged (17:48). The point is not that the tools are useless but that speed of code production was never where enterprise value came from; Garran uses Microsoft as the example, arguing the value came from the IBM distribution lock-up and network effects, not from anyone typing faster. On the model side he flags OpenAI’s share falling from roughly 90% in May 2023 to 49.7% in May 2026 while Chinese models undercut inference costs to something like 5% of OpenAI’s (23:56). Distinguish user delight from commercial moat; only the latter services debt.
Misallocated Capital Is 70% of GDP — and the Real-Terms Capex Cut Has Already Started: Using the Wicksell spread framework, Garran now puts misallocated US capital at roughly 70% of GDP, four times the level on the eve of the global financial crisis and 24 times the dot-com eve, raised from a prior estimate on the back of this year’s fiscal and monetary stimulus (51:11). The macro arithmetic is unforgiving: the data center buildout adds roughly 1.5 percentage points to GDP growth and the Nvidia wealth effect another 1.5, so merely standing still removes about three points of nominal GDP and a reversal removes three to six (48:43). Nobody Special Finance points out the cut is already happening in real terms — neither Meta nor Microsoft raised the top end of capex, Microsoft shifted some capex into operating expenses, and against rising memory, steel, copper and GPU prices flat nominal spending is a unit-volume cut (44:26). Zitron expects visible capex reductions within roughly three months, with one company moving and the rest following (46:13). Track real capex, not headline capex.
Investment Focus
This is the week’s definitive short case, and its distinguishing feature is that every leg is falsifiable on a published number rather than a valuation opinion. The investment template: (1) treat OpenAI’s ability to access public markets as the single binary that governs the entire complex, and the S-1 disclosures as the moment private accounting meets public scrutiny; (2) follow Garran in moving the short exposure away from the software layer and toward the chip owners, chip suppliers, data centers and Nvidia itself, on the logic that announcing you are renting chips out rather than using them now earns a stock a bid (58:24); (3) avoid the neoclouds entirely — Nobody Special Finance frames them as piles of debt sitting next to piles of depreciating assets, where the assets depreciate away before the debt is satisfied (1:00:28); (4) position for the reflationary aftermath in resources and emerging markets, which Garran argues relative investors should already be starting, on the 2000–2001 template where the short leg paid for the wait; (5) monitor the concrete tells the panel names — bank and private credit appetite, useful-life and capitalization changes, real-terms capex, and the rate backdrop, where Nobody Special Finance notes the 30-year yield at a 19-year high is kryptonite for a debt-financed buildout, and that an analyst was willing to downgrade Caterpillar on a data center moratorium while nobody will touch Nvidia (1:04:45).
EP 2 - The Fed Is Trapped
Jim Bianco — Bianco Research President
Jim Bianco, speaking with Anthony Fattouh on the What the Finance podcast, is the deliberate counterweight to the opening episode and the most important one to read against it. He is constructive on the US economy, genuinely bullish on what AI does to the software stack, and bearish on bonds precisely because of both. His framing is that the consensus is still marking this economy against a pre-2020 template that no longer exists — and that the bond market, not the equity market, has already worked this out. The result is a call almost nobody else on the buy side is making out loud: the market is pricing a rising probability that the Fed’s next move is a hike, and Bianco thinks it should be.
Actionable Bullet Points
Stop Watching Labor Demand and Start Watching Labor Supply: Bianco’s reframing of the labor market is the analytical core of his optimism. With fertility below replacement since 2007 and immigration effectively halted, the US is no longer generating new working-age population, which means the economy needs only around 35,000 jobs a month, plus or minus twenty or thirty thousand (3:44). Against that bar, June’s 57,000 print was adequate rather than alarming, and the trailing six-month average of 92,000 is comfortably above what the economy requires (23:38). Historical comparisons are meaningless because they were drawn from a period of much higher population growth. Do not read soft payroll prints as recession evidence in a zero-population-growth economy.
His AI Bull Case Is a Software Substitution Case, Not a Productivity Slogan: Bianco separates generative AI, which most of Wall Street uses as a better search box, from agentic AI, which does work on your machine — and he argues the second is where the revenue comes from (8:52). The funding mechanism is substitution: the average corporate computer costs more per year in software than the machine itself, spread across the operating system, office suite, CRM, security and specialist tools, with the human acting as the integration layer between systems that do not talk to each other (10:34). If a prompt replaces that stack, the money already being spent gets redirected rather than newly found — which is also, he notes, why software equities are cratering. He points to Larry Page’s reported willingness to see Google go bankrupt rather than lose the AI race as a statement that AI would otherwise obsolete Gmail, Maps and Sheets (12:10). Underwrite AI revenue as a transfer from incumbent software budgets, not as incremental spend.
The Risks He Names Are Risks of Success, Not of Failure: Bianco is explicit that he is watching for what could break the thesis, and the candidate he takes most seriously is free Chinese open-source models. A company with a few hundred employees can spend $200,000 to $300,000 on hardware, run a sophisticated downloadable model in house, and reduce its ongoing AI cost to electricity (17:01). He reads China’s giveaway strategy as classic dumping — no one will trust a Chinese-hosted account, so distribution has to be free — aimed squarely at forcing the labs to cut token prices to compete (18:44). The other risk is data center backlash choking off the buildout. Watch open-source adoption inside enterprises as the mechanism that compresses lab revenue without any bubble bursting.
The Bond Market Is Telling Warsh to Hike, and Bianco Agrees: On Fed funds futures, the probability of a hike at the July meeting moved from 9% a week earlier to 34% at the time of recording, and Bianco reads the September meeting as carrying a far higher implied case for a hike (31:19). His argument is that the Fed cannot print oil but it can address demand: easing into a supply constraint only widens the imbalance and pushes prices higher (32:26). With Brent up 36% since July 2nd, nominal growth rising and the economy not slowing, the neutral rate is moving up — which makes holding steady a de facto ease (28:50, 33:06). He supports Warsh’s abolition of forward guidance on the grounds that the Fed has repeatedly promised paths it did not deliver. Do not position for cuts.
The 2022 Comparison Is the Trade — Panic Is Bullish for Bonds: Bianco’s sharpest observation is a simple pair of numbers. In all of 2022, with inflation running at 9%, the 10-year yield never exceeded 4.23%, because the Fed was visibly panicking and hiking 75 basis points a meeting (35:41). Today the 10-year is 4.65% and the 30-year 5.15% with inflation far lower — his explanation being that the Fed is not taking the inflation fight seriously, so the bond investor has to (36:24). Against a backdrop where the ECB, Australia and Japan have already raised and he expects the Bank of England to follow, continuing to resist the global trend simply removes the incentive to own US duration (37:34). Inflation has now run above 2% for 64 consecutive months, which he treats as a regime, not a lag (22:25). A little Fed panic is what stops long yields going up.
Investment Focus
Bianco is the most useful stress test available for the bear case in EP 1, because he agrees the spending is enormous and disagrees only about whether the revenue arrives. The investment template: (1) be short duration or underweight long bonds — his explicit view is that the path from here runs to 5.5% and then 6% on the long end unless the Fed hikes, which makes a hawkish surprise bullish for bonds and a dovish hold bearish; (2) do not use pre-2020 recession heuristics, and note that the Conference Board leading index fell for 32 consecutive months into 2024 without producing the recession every model forecast (21:34); (3) separate the AI infrastructure trade from the AI application trade, and recognize that his own thesis is bearish incumbent software by construction — the money for AI has to come out of somebody’s existing budget; (4) treat free Chinese open-source models as the specific mechanism that could compress lab economics without any accounting scandal or credit event; (5) watch the September FOMC as the live event, since Bianco’s framework says the neutral rate is rising with nominal growth and energy, and following it higher is neither restrictive nor a policy error — the error is sitting still while it moves.
EP 3 - The US Bails Out Japan — Are US Stocks Next?
David Nicoski — Vermilion Research CIO
David Nicoski, in conversation with David Lin, supplies the tape evidence for everything the macro guests are arguing. He is a technician and refuses to call a top, but his relative-strength work says the rotation out of technology is real rather than cosmetic, that concentration risk in US equities has reached levels with only one historical precedent, and that everything investors sold to fund the technology trade has been bottoming for four to five weeks. Recorded on Monday August 3rd, the conversation also captures the weekend’s significant event: a US intervention to buy Japanese yen, the first at this scale since 1998.
Actionable Bullet Points
The Rotation Is Real, and It Is Showing Up in What Was Sold to Fund Tech: Since the end of June, Nicoski has seen relative strength improve across financials, healthcare and to a lesser degree staples, and his framing is deliberately precise: everything investors sold to get into the technology trade is now outperforming (1:10). He will not call a top, because his discipline requires long-term trendlines to break first, but he argues you do not need to call one — you need only recognize that breadth is improving in the areas that were abandoned (3:33). Correlations across the market are unusually low and single-stock skew is unusually high, which he reads as a stock picker’s market rather than an index market (11:58). Trade the dispersion rather than the direction.
Concentration Risk Now Has Only One Historical Analogue: Technology is roughly 42% of index weight against about 40% at the dot-com peak, which Nicoski describes as starting to get egregious even though it can persist (4:02). The semiconductor comparison is starker: semis were about 9% of the market during the dot-com bubble and peaked near 22% this cycle, and that figure understates the exposure because it excludes the hyperscalers designing their own silicon — companies effectively competing with the semi index while sitting outside it (9:11). On the percent-above-200-day measure the group reached extremes matched only by the dot-com period. His conclusion is a probability statement rather than a forecast: with that information, do not carry 22% or 26% in semis when other sectors are bottoming on relative strength (10:11).
Korea Is the Template, Not the Exception: The KOSPI fell 33% in July alone — worse than 1998, worse than 2008, worse than the dot-com unwind — after a run that made it the world’s best performer, and roughly 1.2 million accounts were margin called, close to 3% of the population (2:12). Nicoski had called that market’s bottom in December 2024 and January 2025 when names traded at four and five times earnings, and watched breadth collapse from 70% of names above the 200-day to 15% as the index was held up by two stocks (3:26). The tech-heavy markets — Korea, Taiwan, the Nikkei — share the same signature: a parabolic advance now breaking its short-term uptrend. If the US breaks its longer-term uptrends, he treats that as the vulnerable signal that America follows.
The Yen Intervention Is an Attempt to Force the Carry Trade Unwind: The US bought yen for the first time since 1998 after the currency touched a 40-year low, and Nicoski reads the motive as pre-empting an Asian-contagion repeat, since Japan imports essentially all of its oil and seizes up if the currency keeps weakening (12:28, 20:16). The mechanics are what matter: yen strengthening and Japanese yields rising simultaneously are both anti-carry-trade, so the intervention is effectively speeding up an unwind rather than preventing one (20:53). What surprises him is the absence of an equity market reaction, given that the classic trade is to borrow yen cheaply and buy US assets — leaving open whether positioning was already correct or something is still coming (14:38). Separately, he flags the dollar breaking its uptrend through the 100–101 support zone, and recalls that when the dollar broke a parabolic uptrend on April 12, 2002 the S&P fell 25% — a repatriation call he has not made in 24 years (17:49, 18:31).
Energy Is the Positioning Asymmetry; the Consumer Is the Warning: Nicoski notes that most investors do not realize XLE has outperformed XLK from the COVID lows, and that with energy at 3% of the market, simply being market weight is a 3% position most managers do not carry — so a 1% shift in fund flows is a 33% move in the sector (22:57, 23:30). He does not expect $40 oil, with crack spreads near record highs effectively pricing $140 crude, which he translates into higher inflation expectations and bond yields that keep moving up unless a lower equity market forces money into bonds (22:32, 24:20, 24:35). On the consumer, the discretionary sector sits at a 14-year relative strength low, McDonald’s at a 14-to-16-year low, and Target has handily outperformed both Walmart and Costco over six months — while anything requiring a five-year payment, from RVs to boats, sits at the bottom of his rankings (25:51, 27:03, 30:00). Consumers are buying experiences, not obligations.
Investment Focus
Nicoski’s value here is that he is not arguing a narrative — he is reporting what price is already doing, which makes him the timing check on the week’s more thesis-driven guests. The investment template: (1) reduce semiconductor and technology concentration toward or below market weight rather than waiting for a top to be called, since his own book remains overweight tech only until long-term trendlines violate (31:24); (2) add where relative strength is bottoming — financials and banks, insurance names with limited exposure to private debt and AI funding, and European banks such as UBS and Deutsche Bank that have outperformed for four and a half to five years since exiting negative rates (31:41, 32:48); (3) in healthcare, he highlights medical devices, naming Medtronic, Abbott and Baxter as improving over the last month and a half as funds trim technology weightings (33:10); (4) own energy for the flow asymmetry rather than the oil forecast, given a 20-year base on XLE and free-cash-flow-positive constituents; (5) watch credit default swaps on AI-linked names as the risk tell he is monitoring, and gold at the 4,100 level — where the 200-day sits and he is interested — with enthusiasm fading if it breaks 4,000 (32:33, 34:15).
EP 4 - I Cut Tech. Here’s Why.
Barry Knapp — Ironsides Macroeconomics Founder
Barry Knapp, speaking with Maggie Lake on Wealthion, offers the clearest mechanical explanation of the K-shaped economy available anywhere this week: it is not a social phenomenon but the residue of a Fed that eased with its balance sheet and tightened with its policy rate. That asymmetry, he argues, is precisely what Kevin Warsh was appointed to unwind, and he lays out the three-step sequence he expects. He then does something most strategists avoid — he tells you exactly where he has put the money, and it is a large, uncomfortable underweight in the sector that has driven every index return for two years.
Actionable Bullet Points
The K Is a Policy Artifact, and That Means It Can Be Fixed: Knapp traces the K-shaped economy to the Fed easing primarily through the balance sheet during the pandemic and tightening primarily through rate policy afterward, leaving what he estimates as half a point to three-quarters of a point of excess accommodation at the back end of the Treasury curve (1:28). The consequence is that homeowners, equity holders and high-quality long-term fixed-rate borrowers — including the hyperscalers and the big banks — receive accommodation, while floating-rate borrowers, small banks and paycheck-to-paycheck households face restrictive policy. The evidence is in the spread: the return-on-equity gap between regional and money-center banks is roughly 3% where it would typically be half that, and small business employment is contracting about a quarter percent annually while large business employment grows at 1.5% (2:17, 2:43). Because the cause is policy design, the cure is policy design.
Four Adverse Demand Shocks Explain Why Growth Is So Narrow: Knapp counts four deliberate hits to aggregate demand from the current administration: slower immigration, government spending growth cut from 11% in the prior administration’s final year to 3%, tariffs that both suppressed demand and compressed margins, and the war-driven energy price shock (3:02). Staples company profit margins are at their lowest levels in more than 30 years as a direct result (3:31). Those headwinds are beginning to fade but still linger, which is why growth is concentrated in the AI theme and its capital spending while the rest of the economy muddles. Read narrow growth as policy-induced rather than structural, and expect breadth to improve as the shocks age out.
The Warsh Sequence Is Three Steps and Roughly Two Quarters Away: Knapp is confident this policy imbalance is the reason Warsh has the chair, and describes the plan as: cut the policy rate toward 3% to relieve floating-rate borrowers; stop reinvesting maturing securities into 10s and 30s and start reinvesting at the short end to remove the accommodation embedded in a $6.5 trillion long-duration portfolio; and deregulate the banking system so banks can absorb the securities, which is what Michelle Bowman is already doing (5:43). He expects three to six months for the task forces to build the academic and political justification, putting implementation later this fall as inflation readings come down (7:28). He also argues the market will gain confidence once the process starts, because the current uncertainty — a sword hanging over the long end since Yellen’s 2023 duration extension pushed 10s to 5% and forced a retreat into bills at roughly a third of issuance against a 15–20% recommendation — is worse than the unwind itself (9:32). Position for a steeper curve, not a bond rally.
Inflation Stabilizes Near 2.5%, and the 2% Target Was the Original Error: Knapp argues the 2% target adopted in January 2012 was a mistake of timing — it followed two years in which the Fed’s preferred deflator ran a median 1.57% in the aftermath of a debt crisis, a period the academic literature predicts produces years of disinflation (16:11). Since the early 1960s no business cycle other than that one has run inflation below 2%, and the median postwar CPI increase is 3% (18:45). Near term he sees benign readings: goods prices flat since the effective tariff rate peaked last September, core services seasonally softening, and rents still being calculated near 3% by CPI and the PCE deflator while every alternative measure says two or lower (22:15). His landing zone is 2.5% — stable enough to support a broad capital spending boom and a manufacturing renaissance, and low enough that investors stop pricing inflation risk. Do not position for a return below 2%.
He Cut Tech to 25% Against a 37% Index Weight — and Says Why: Knapp’s view is not that the technology cycle is over but that the rate of change of capital spending growth is set to slow, and his strongest evidence is price: the four big spenders all reported on April 29th and their shares fell throughout the subsequent quarter after raising capital spending (29:17). His preferred metric is capex as a percentage of cash flow, which peaked near 80% in telecom in 1999–2000 and again in energy during the 2014–15 shale boom; the big spenders are at 90% today (29:34). Google — the furthest along in monetizing the investment — fell 7% the day after reporting negative cash flow for the quarter (30:19). He is therefore at 25% technology against a 37% index weight and 5% communication services against 10, which he acknowledges is a very large bet, adding that 50% of a portfolio in one theme borders on recklessness (30:38).
Investment Focus
Knapp is the week’s most implementable macro framework because the policy call and the portfolio are the same call. The investment template: (1) overweight the non-AI capital spending story — industrials, energy and materials — on the manufacturing renaissance thesis, while acknowledging his own caveat that industrials are expensive because the crowd already agrees, and that this is one where the crowd is probably right and gets paid over time (31:29); (2) overweight financials specifically as the Warsh trade, since deregulation plus a steeper curve plus a lower policy rate is a powerful combination for bank profitability, and regional banks still screen cheap despite a strong year (35:38); (3) underweight the consumer — both staples and discretionary — on margin pressure and the K, and note his sharpest valuation point: paying roughly 30 times earnings for a staples business growing earnings 5% is not a defensive trade, it is a silly one (34:19); (4) hold meaningful cash, because rising real rates can trigger a 10% drawdown at any point and midterm years are when that typically happens (33:04); (5) treat the reversal signal as explicit — when the Fed is definitively unwinding the imbalanced policy, he would rebuild consumer exposure and would consider taking technology back up, though probably not to market weight.
