Exponential Learning

An initiation on Axon Enterprise (AXON), from Alpha Brands Research's Oct 19 report: the company owns every link in the public safety chain — TASER, cameras, drones, and the evidence cloud — with $3.22B of revenue growing 34.6%, $1.64B of SaaS recurring revenue, 126% retention, and $15.1B already contracted. The problem is cash: just $133M of free cash flow against $634M of stock comp, shares 56% below their high. The 12-month view is a hold; Q3 (expected ~Nov 4) and Q4 cash flow are the test of the thesis. For informational and educational purposes only — not financial advice.

A portfolio construction episode: 31 names in 5 sleeves built from the Flaningam, Noren/Apollo, and Hosking letters. Own the scarce inputs (the full power chain, ASML, Synopsys), rent the buildout with exit discipline, overweight the proven adopters — plus beloved brands and military modernization. Includes the weight sketch and the three kill switches that would change the thesis.

Trevor Noren (Sage Road Research) and Apollo ask the demand-side question: the marginal cost of compute exceeds its marginal benefit, corporate token spending is slowing, and hyperscaler ROIC has halved from ~38% to ~20%. We cover the prisoner's dilemma of the buildout, the private-flight-vs-cell-phones analogy for agentic AI, the fiber-optic precedent — and why it all still points to owning the tollbooth.

Eric Flaningam (Timeless) published the most rigorous bull-case math on AI cap ex: clouds need ~$600B in third-party AI revenue by late 2027 and $1T+ by 2028 at 35% margins to earn their cost of capital. We steelman the piece — the depreciation treadmill, the three conditions (frontier breakthroughs, faster diffusion, surviving margins), and the 'apps decide what runs short' thesis — then deliver our conclusion: own the tollbooth, rent the factory.

Hosking Partners' Q3 2026 letter, via MastersInvest: a London value manager beat global markets by 6.3% YTD owning almost none of the AI winners — and its diagnosis of the AI buildout is the sharpest we've read. Cap ex doubled to T (a railway-boom share of GDP); 00B+ of IG debt issued in 2026 alone, .1T more by 2030; credit investors now run the industry through SPVs and off-balance-sheet leases while token prices fall. The core claim: the data centre is a factory, compute is a commodity — load-the-plant economics will push token pricing toward marginal cost, the way solar was commoditized (and China-dominated). The .5T OpenAI/Anthropic flotations are the exam. Plus: how they won without AI (Greek refiners, Norwegian tankers +440% since 2017, Saga's 25% ROIC, 80 Japanese small caps), the honest 260bp cost of missing mega-cap tech, a full steelman of the other side, and why the factory thesis actually CONFIRMS the energized-land view — factories commoditize, tollbooths don't.

Alpine Macro's Chen Zhao on the three macro stories that matter right now. (1) America: the bond selloff has likely crested — 10Y down 30+ bps to 5.1%, and the contrarian evidence (bearish crowd, 6% calls, Economist cover, heavy shorts) says yields topped; hot CPI/PPI produced curve flattening, not a selloff — the long end fears growth. Tech rout is 'healthy'; roadmap stays 'wobble now, new highs later'; rebalance 70/30 to 60/40. (2) France: OAT-Bund 140-150bp, a state spending 58% of GDP, 5.4% deficit, 120% debt, borrowing at 4.8% vs 2% nominal growth — the debt-trap math; Le Pen's plan lacks a 'how'; no Draghi moment yet — short the CAC on strength, EUR toward parity in 2027. (3) China: easing + weak economy + record-low yields + one of the world's strongest currencies — explained by internal depreciation (15% undervalued REER) and a Japan-1990s-style capital recycling problem (00B+ surplus with nowhere to go); the CGB bull continues while CNY stays strong, and China's excess savings quietly subsidize global bond yields. Plus: where the thesis could break — oil shock, French contagion, China's flows reversing.

A sharp new essay — 'A Defense Primer: The Economics of Having One Customer' — argues defense investors keep making the same mistake: valuing a genuine military requirement as durable cash flow without checking which pipe the money flows through. Alex and Jordan walk the budget machine (authorization vs appropriation, CRs, supplementals, reconciliation), three wars that show what funding paths do to contractors (the textbook Gulf War supplemental, Iraq's 0B-a-year parallel budget, and today's 'rabbit-duck' 84B acquisition budget that's really 95B in base), and the core concept: Funding Path Risk. Plus Taiclet vs Warden at the Morgan Stanley conference — same demand story, two very different exposures — and why, for primes, politics is the prism. Note: the essay's account of 2026's Iran operation is presented as the author's case study; details not independently verified.

Five weekly research decks, six weeks, one regime shift: from scarcity of growth (pay any price for AI earnings) to scarcity of affordable, durable growth (cash generation, financing strength, reasonable valuations). Alex and Jordan walk the five chapters — the illusion of calm, the macro squeeze, the fault line, the AI refuge, the great rotation — then unpack three contradictions (Fed relief without long-end relief, AI earnings without AI stock returns, October breadth unlike September resilience), separate durable signals from noise, and lay out four tests for whether October's rotation is real. Bottom line: the market may be moving from growth at almost any price toward growth and cash flow that justify their price in a 5%-plus rate regime.

Eric put a deposit on the Scout Terra pickup and asked the two questions every reservation holder should ask: who's backing this company, and will they ever build the thing? Alex and Jordan dig in — Volkswagen's B bet, the B South Carolina factory at 90%, why 87% of reservation holders chose the gasoline range-extender, the slipped 2028 timeline, the dealer lawsuits, and the serious risks from VW's profit warning to a brutal EV winter.

Alex and Jordan go deep on Sonu Varghese's Carson Group piece 'It's All One Big Trade Now: AI or Bust': why hot inflation sent yields plunging, the AI credit stress (CoreWeave, Oracle, Meta CDS), the inflation supercycle vs the 1990s, concentration (42 stocks driving two-thirds of returns), plus a stress-test of the thesis — the bull case, private credit's role, and what would prove it wrong. For informational and educational purposes only; not financial advice.