SHOWNOTES

A brisk tour through oil, gas, and U.S. fiscal mechanics, followed by a biotech history chapter and a deep dive on NVIDIA’s sustained dominance. We close with a big healthcare development—“Trump-RX”—and quick hits from Harrow Health’s investor day.

[00:00] Intro

[00:28] Exhibit C – Oil: $62 is Doing Its Job

Hunt frames today’s oil tape: production is rolling off in the U.S., OPEC, and non-OPEC; geopolitical thawing could keep prices soft near term. With Saudi capacity around 12 mb/d and surplus capacity trending toward ~2.5 mb/d by 2026, a “normally supplied” market implies oil trades in the $70s—though swoons into the $50s can happen on the way there.

[02:05] Exhibit B – Natural Gas, Power Burn & LNG

Gas underperformed, but storage is robust. Power burn was flat ’25 vs ’24 despite “AI, AI” chatter—a surprise tied to coal plant life-extension and a near-term lull in new wind/solar coming online. LNG demand stepped up this year; if power burn resumes climbing, a durable ~$4 gas band becomes more plausible.

[03:22] Exhibit A – U.S. Cash Flow & the Shutdown

The team walks through the continuing resolution dynamics: House bills vs 60-vote Senate reality, daily vote tactics, and the case for finishing all 12 appropriations. Beyond the political scoreboard, the focus is on process: budgeting via CRs isn’t how the system should work—and markets will parse timing of data releases like jobs.

[05:44] Biopharma History (Part 4): From Genzyme to Monoclonals

Jason spotlights Genzyme’s rare-disease model (boosted by the Orphan Drug Act), the Human Genome Project’s 13-year blueprint, and the 1996 Dolly milestone foreshadowing cell reprogramming. Genentech’s late-90s antibodies (Rituxan, Herceptin) ushered in targeted oncology—monoclonals remain the highest-grossing class today.

[12:09] Next Week: OpenAI—A Controversial History

Programming note: the crew will tackle OpenAI’s origin story and flashpoints next week, with Oracle and NVIDIA history episodes likely to follow.

[14:19] NVIDIA’s Moat: Cash, CUDA, Cadence (p. 3)

Hunt updates the NVIDIA page: zero net debt, >$50B cash, ~$85B run-rate FCF on ~$170B revenue, and modest capex vs hyperscalers. Mike explains why CUDA’s ecosystem and annual product cadence (powered by ~$15B R&D) create a TCO advantage so strong that even “free” AMD GPUs lose in practice; the window for rivals narrows as workloads stay variable.

[21:20] Post-Break: ASICs vs. AMD & What’s After LLMs

Jason notes the market’s pivot: when buyers do switch from NVIDIA, they’re more often going custom (ASIC/TPU) than to AMD. He adds that frontier researchers expect architectures beyond today’s LLMs to power the next step in reasoning—another reason flexibility favors NVIDIA’s stack for now.

[22:43] Healthcare Shake-Up: “Trump-RX” and PBMs

A new Medicare direct-to-consumer platform (Pfizer as first partner) could remove PBMs from many transactions, standardize pricing, and simplify Part D decisions. The panel discusses “gross-to-net” dynamics (often ~50% industrywide) and why UnitedHealth’s PBM economics could face pressure; potential annual savings to Medicare could be very large.

[27:25] Harrow Health Investor Day: Buying Back Melt

Harrow fully acquires the rest of Melt (anesthetic asset) for $8M upfront plus royalties. The team praises the capital allocation arc—spin out in the lean years, re-acquire after Phase 3 success once cash flows stabilize—highlighting optionality beyond ophthalmology (e.g., dentistry).

[28:41] Wrap-Up & Next Week

Sign-off with a reminder to check the memo. Next episode: OpenAI’s history—then likely Oracle and NVIDIA deep dives.

Thanks for listening! If this episode helped clarify markets, share it with a friend and subscribe. Grab the Cash Flow Memo for all exhibits and the ~80-company financials.

This podcast and the information herein are intended for informational purposes only. The views expressed herein are the author’s alone and do not constitute an offer to sell, or a recommendation to purchase, or a solicitation of an offer to buy, any security, nor a recommendation for any investment product or service. While certain information contained herein has been obtained from sources believed to be reliable, neither the author nor any of his employers or their affiliates have independently verified this information, and its accuracy and completeness cannot be guaranteed. Accordingly, no representation or warranty, express or implied, is made as to, and no reliance should be placed on, the fairness, accuracy, timeliness or completeness of this information. The author and all employers and their affiliated persons assume no liability for this information and no obligation to update the information or analysis contained herein in the future.

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit telltales.substack.com

The Cashflow Memo

Download

SHOWNOTES

A fast-moving tour across energy, technology, and healthcare: oil spare capacity looks set to tighten, natural gas loses its backwardation, DC politics and immigration policy enter the chat, we revisit Genentech’s beginnings, and then unpack Apple, Oracle/Broadcom, and NVIDIA through a cash-flow lens.

Stick around for a sailing community shout-out from Oakcliff and a healthcare segment weighing Tylenol headlines against mixed evidence.

[00:00] Intro

[00:19] Disclaimer

[00:26] Exhibit C — Oil: spare capacity on track to shrink

Hunt walks through surplus crude capacity by country (Saudi, UAE, etc.) and argues that by next year the cushion could slip back to ~2.5 mb/d. With demand growth modest, geopolitics (Ukraine, Iran) and the “$62 oil did its magic” supply response suggest the window for sub-$50 oil may have passed.

[03:13] Exhibit B — Natural Gas: futures tone weakens

Backwardation is gone; later-dated gas sits around ~$3.95 (’26), with a ’25 average ~ $2.65 and prompt near ~$3.10. Supply ticked down when prices averaged ~$2.40–$2.75, but rebounds in ’25 (Permian leads; Waha even flipping negative). Despite AI power buzz, gas-fired demand has flattened as coal plants linger and wind/solar come online.

[05:35] Exhibit A — DC budget brinkmanship & policy crosscurrents

House passes a short extension; Senate dynamics keep shutdown risk alive around October 1 and again into Thanksgiving without a broader deal. Conversation turns to a White House move to hike H-1B employer application fees from ~$2,500 toward ~$100,000—raising questions about talent, wages, and whether price hikes would throttle volumes or better filter for “best and brightest.”

[11:12] Biopharma History (Part 3): Genentech, insulin & the tools that unlocked biotech

Jason traces Boyer/Cohen’s 1973 gene-splicing breakthrough and Genentech’s path from somatostatin to human insulin (Humulin) and a blockbuster 1980 IPO alongside peers like Biogen and Amgen. Mike spotlights PCR’s origin story—Kary Mullis’s unconventional inspiration—and how DNA amplification reshaped diagnostics, forensics, and research.

[17:33] Oakcliff Sailing

From Newport’s Sailing Museum to fundraising for sails and a long-term home in Oyster Bay, Dawn Riley shares updates and an invitation to visit, support, and get involved.

[19:52] Healthcare Update — Tylenol headlines vs. mixed evidence

Trump/HHS draw a headline-grabbing link between prenatal Tylenol use and autism. The crew weighs the literature (often ADHD-focused, mixed conclusions), real-world usage since the 1950s, and pregnancy tradeoffs when Tylenol is the only approved painkiller—underscoring how nuance gets lost in big-policy moments.

[22:09] Page 1 — The case for $AAPL

Hunt contrasts hyperscaler CapEx (Amazon ~$90B; Alphabet/Microsoft $65B each) with Apple ($12B) and asks whether Apple can win via on-device/app-centric AI without giant model spend. Jason, a hardware/software skeptic of late, notes a packed Apple Store and a rare OS update that actually feels faster—hinting at green shoots on experience even as security/AI integration remain hard problems.

[25:17] Page 2 — Broadcom vs. Oracle: cash flow, CapEx, and OpenAI exposure

Oracle is running heavy CapEx with negative FCF tied to AI data centers and a massive OpenAI compute commitment; leadership shifts elevate cloud execs. Broadcom, by contrast, throws off FCF with comparatively little CapEx and benefits when hyperscalers deploy non-NVIDIA accelerators. The debate: is Oracle’s risk/reward compelling—or just risky?

[27:57] Page 3 — NVIDIA’s lead & the demand question

NVIDIA’s equity value (~$4.4T) and striking FCF margins prompt the sustainability question. Mike thinks supply execution looks solid—each new chip lowers cost-per-token and sustains ROI—but demand relies on platform wars, funding cycles, and whether value accrues beyond NVIDIA. If financing tightens, who blinks first—and do balance-sheet giants like Google gain relative advantage?

How did you like this week’s Telltales? Your feedback helps me make this great.

Loved | Great | Good | Meh | Bad | ….. If you liked this post from Telltales, why not share it?

Share

This podcast and the information herein are intended for informational purposes only. The views expressed herein are the author’s alone and do not constitute an offer to sell, or a recommendation to purchase, or a solicitation of an offer to buy, any security, nor a recommendation for any investment product or service. While certain information contained herein has been obtained from sources believed to be reliable, neither the author nor any of his employers or their affiliates have independently verified this information, and its accuracy and completeness cannot be guaranteed. Accordingly, no representation or warranty, express or implied, is made as to, and no reliance should be placed on, the fairness, accuracy, timeliness or completeness of this information. The author and all employers and their affiliated persons assume no liability for this information and no obligation to update the information or analysis contained herein in the future.

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit telltales.substack.com

SHOWNOTES

This episode tackles the week’s cash-flow movers across energy, tech, and healthcare—then time-travels through biopharma’s biggest breakthroughs. We close with practical EV takes, Big Tech updates, and why Mars might be more interesting than the headlines suggest.

[00:00] Intro

[00:30] Exhibit C: World Oil Supply/Demand

IEA’s latest revision lifts 2025 demand growth to ~0.7–0.8 mb/d (vs. ~0.4–0.5), implying the surplus could tighten from ~3.3 mb/d toward ~2.0—mostly in Saudi spare capacity (10.1 mb/d run-rate vs. ~12 mb/d capability). Futures sit near ~$62 as macro/geopolitical risk (Russia distillates, Iran supply) caps downside but keeps the curve flat in the low-60s.

[03:34] Exhibit B: U.S. Gas Demand/Supply

Production hovered near ~106 while power burn stayed frustratingly flat (~37.8→38.5). Spot printed $3.04 last Thursday, with a contangoed strip around ~$3.83–$3.90 for 2026—disappointing versus the old ~$4 anchor given modest data-center uplift to date.

[04:38] Exhibit A: U.S. Government Finances & Rates

Congress passed a CR to Nov 20; Democrats seek Medicaid restorations, making a shutdown plausible. The Fed delivered a quarter-point cut—arriving “too late” in Hunt’s view—after holding steady post last fall’s 50 bps move; politics around timing get a candid airing.

[06:50] Biopharma History: From Penicillin to mRNA

Alexander Fleming’s accidental penicillin find (1928/29) led to wartime scale-up (Pfizer’s deep-tank fermentation; doses >2M by 1943) and a “Golden Age of Antibiotics” (streptomycin for TB, tetracycline). Genentech’s 1980s recombinant insulin marks the modern biotech inflection; BioNTech’s mRNA pivot shows how far we’ve come—even as viruses and antibiotic resistance remain unsolved frontiers.

[17:16] Oakcliff Sailing

Community update and Annapolis Boat Show plans; proceeds benefit programs via racingcalendar.org.

[19:29] Page 1: Mega-Cap Check-In

Quick scan of the “Page 1” five. Tesla highlights include Musk buying ~$1B of shares and Nevada approval to test robotaxis; Amazon FC tour preview tees up automation talk; Microsoft’s OpenAI arrangements continue to solidify.

[19:44] Apple: Translation AirPods & the AR Path (p. 1)

AirPods Pro 3 impress—live in-ear translation hints at Apple’s real-world AR strategy while iPhone “Air” appeals to China’s thin-phone fad and preps a foldable future. New “liquid glass” UI lands better in hand than on stage; the market’s “not enough AI” reaction misses the iteration-to-platform arc.

[21:51] Microsoft & OpenAI: Deal Momentum (p. 1)

No big surprises—just steady progress that reinforces Azure’s AI gravity and inference economics, with edge vs. cloud tradeoffs a watch-item.

[22:28] Oracle: Building AI ‘Factories’ (p. 2)

Larry Ellison leans into capacity for inference—Oracle wants to be a go-to “AI factory” supplier. Jason flags an overbuild risk if inference migrates to edge devices, but legacy database strengths gain new relevance in the AI era.

[25:10] Page 3: NVIDIA & China Strategy

China signals intent to reduce NVIDIA reliance while simultaneously pressing a prior remedy about uninterrupted supply—talk vs. action tension. Bottom line: NVDA’s constraint is still TSMC wafer output; global demand ex-China remains plenty.

[27:26] Life on Mars?

NASA reports organic carbon and iron compounds consistent with microbial biosignatures; alternative non-biologic origins remain unconvincing. Sample-return logistics could unlock definitive answers—surprising this didn’t dominate the news cycle.

[29:09] Tesla Model Y Door Safety (p. 1)

Emergency latches exist but aren’t intuitive, especially rear seats; several accidents spotlight the issue. Expect regulatory pressure for a clearly mechanical primary interior release.

[30:16] Used EVs: Who Should Buy What? (p. 1)

Charging at home is the gating factor; in California it’s often a no-brainer. For used, Tesla’s deeper field data and software stack make it the safer pick; typical battery life guidance ~150k miles, with many vehicles exceeding that—non-Tesla software still lags.

Thanks for listening! If this helped you think in cash-flow, share the episode and grab the full memo for charts, exhibits, and ~80 company updates. Subscribe so you never miss the next deep dive.

This podcast and the information herein are intended for informational purposes only. The views expressed herein are the author’s alone and do not constitute an offer to sell, or a recommendation to purchase, or a solicitation of an offer to buy, any security, nor a recommendation for any investment product or service. While certain information contained herein has been obtained from sources believed to be reliable, neither the author nor any of his employers or their affiliates have independently verified this information, and its accuracy and completeness cannot be guaranteed. Accordingly, no representation or warranty, express or implied, is made as to, and no reliance should be placed on, the fairness, accuracy, timeliness or completeness of this information. The author and all employers and their affiliated persons assume no liability for this information and no obligation to update the information or analysis contained herein in the future.

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit telltales.substack.com

SHOWNOTES

A brisk, cash-flow-first tour through energy, technology, and healthcare: oil and gas realities, the federal balance sheet, big labor revisions, and the AI build-out—plus biotech’s backstory and fresh FDA/advertising policy moves.

[00:00] Intro (n/a)

[00:19] Disclaimer

Reminder that the episode is informational only and not investment advice. Listeners should do their own work and assume no guarantee of completeness or accuracy.

[00:30] Exhibit C: World Oil Supply/Demand

Oil pricing “stuck” near the low-60s amid geopolitical tensions and weak macro catalysts. Spare capacity is drifting lower mainly from decline curves and reduced capex; U.S. activity (rigs/completions) is down, with 2025 U.S. output guided lower and 2026 ~12.7–12.8 mb/d. New projects look unlikely at current prices, keeping a floor under crude absent major macro breaks.

[04:22] Exhibit B: U.S. Gas Demand/Supply

Despite questions about recovery to $70s oil, weaker demand growth—especially in China and Asia—tempers the upside. The team doesn’t see a $50 handle as likely given supply dynamics and geopolitical backdrop, but also isn’t banking on a quick return to higher prices.

[05:06] Exhibit A: U.S. Government Cash Flow

Targeting progress rather than perfection: discussion of trimming the deficit toward ~$1–1.5T over the next couple years versus COVID-era blowouts. Debt-to-GDP math: if deficits stabilize, the ratio can flatten near ~102% before drifting down. Expect noise as Congress wrestles with funding via continuing resolutions to avoid shutdown-style disruptions.

[07:55] Jobs Revision: The Labor Data Reality Check

Roughly ~1M jobs revised away (~76k/month on average), implying two negative months and several weak ones last year. Vanguard’s vantage via 401(k) enrollments/contribution trends hinted at a softer market earlier; poor inputs complicate Fed timing and reinforce criticism that policy cuts ran “too late.”

[10:14] Biopharma History (10-Minute Sprint)

From ancient fermentation and early antibiotic uses to Pasteur’s germ theory, vaccination experiments, and Mendel’s genetics—laying the groundwork for modern biopharma. Clarifies “biotech” vs “pharma” and notes this series will span multiple episodes.

[17:25] Page 1: Tesla’s $1T Pay Package (p. 1)

Long-dated, “outlandish-but-plannable” milestones: $8T valuation, robotaxis, humanoid robots. Hosts lean over 50/50 on operational targets over a 10-year window; valuation ultimately rests with markets. Governance sidenote: Delaware ruling, Texas reincorporation, and how a prior plan could affect today’s package.

[21:27] Page 2: Software—CRM / NOW / SNOW / ORCL / AVGO (p. 2)

Oracle pops ~30% on AI-related backlog/commentary (notably OpenAI). Debate: Can capital markets really fund another ~$100B for OpenAI? One view: software innovation may pivot, forcing a spending rethink; another: token demand far exceeds supply (HBM bottlenecks, NVIDIA shipping everything, hyperscalers/“neo-clouds” deals), so pricing power and demand migration across providers (MSFT/Gemini/etc.) sustain the buildout.

[26:03] Sponsor Break: Oakcliff Sailing

Quick update from the water: race results, training, and a fall foiling series—follow along on social feeds.

[27:44] Healthcare News Roundup

Executive order seeks to curb pharma ads by reverting to pre-1997-style disclosure—potential hit to legacy media and to DTC channels that bypass PBMs. Policy fragmentation on vaccines at state levels draws criticism. FDA moves to accept single-arm studies for ultra-rare diseases (<1,000 patients) to cut costs/time, and plans transparency by publishing complete response letters.

[31:00] Sign-Off

Thanks for listening! If this helped your investing process, subscribe, share with a friend, and grab the latest Cash Flow Memo to follow the numbers each week.

This podcast and the information herein are intended for informational purposes only. The views expressed herein are the author’s alone and do not constitute an offer to sell, or a recommendation to purchase, or a solicitation of an offer to buy, any security, nor a recommendation for any investment product or service. While certain information contained herein has been obtained from sources believed to be reliable, neither the author nor any of his employers or their affiliates have independently verified this information, and its accuracy and completeness cannot be guaranteed. Accordingly, no representation or warranty, express or implied, is made as to, and no reliance should be placed on, the fairness, accuracy, timeliness or completeness of this information. The author and all employers and their affiliated persons assume no liability for this information and no obligation to update the information or analysis contained herein in the future.

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit telltales.substack.com

SHOWNOTES

This week we open with the Google antitrust decision, then dive into Exhibits C (Oil), B (Natural Gas), and A (U.S. finances) before a tech deep-dive on AI capacity, TPUs vs GPUs, and close with healthcare policy shifts.

[00:00] Intro

[00:38] Disclaimer

General investing discussion only; do your own work. Views are the hosts’ alone.

[00:51] Exhibit C – Oil (Exhibit C)

Geopolitics keeps crude “stuck” around $62–64 with a flat curve (no backwardation). Spare capacity from Saudi/UAE plus tepid demand (+~0.4 mb/d 2H) cap prices; at these levels, U.S. completions slow and ~3 mb/d global surplus could burn off over ~18 months. U.S. output near ~13 mb/d (from ~11.9 in ’22) likely flattens.

[03:07] Exhibit B – Natural Gas

Near-month gas dips under $3 after holding $4–4.25 for ’25–’26; shoulder-month seasonality adds pressure. Dry gas production trending ~108 (from ~106) looks entrenched; power demand is surprisingly flat (coal retirements slower), while LNG rises—netting out to softer pricing.

[04:37] Exhibit A – U.S. Government Cash Flow

Budget brinkmanship keeps Exhibit A in the headlines into Oct. Targeting a deficit move from just under $2T to ~$1.5T would be progress, with $1T the real goal. Ten-year rate resilience matters; if deficits bend down and debt/GDP stabilizes, the U.S. could outperform Europe/Japan/China.

[05:28] Macro: Rates, Dollar & Gold

Beyond the policy rate, markets set the rest—carry dynamics have limits without adding risk. Central banks now hold more gold (by value) than USD reserves, framing the dollar as the “least bad currency.” AI-driven productivity could help “grow our way” out of deficits.

[09:08] Google Antitrust: Defaults Without Exclusivity

The ruling labels Google an illegal monopoly in search but stops short of break-ups; exclusive search deals must end, yet paid defaults continue (Apple’s ~$20B+/yr remains in play). Google may need to license index/interaction data at marginal cost, potentially aiding competitors (e.g., OpenAI). Non-exclusive deals could cut Google’s payments—or invite a Meta bid to become the iPhone’s AI-first gateway (with Apple’s 30% app economics inside the App Store). Parallel talks: Siri ↔ Gemini; Meta exploring Google capacity underscores a broader compute crunch.

[16:22] TPU vs GPU: Who Wins Inference?

China’s domestic stack (e.g., homegrown accelerators) plus energy policy could brute-force parity despite lower-spec chips. Google’s vertically integrated TPUs may lower inference cost; NVIDIA counters with rapid cadence and an inference-tilted roadmap (GB300, “Rubin” series). Ultimately, all-in “cost per token” decides share.

[20:13] Healthcare News: Labels, SNAP & CDC

Texas moves to flag ingredients banned abroad on food labels, restricts SNAP “junk food,” and expands nutrition education—watch for copycat states or FDA action. CDC leadership churn sparks debate; quick course-correction can be healthy if fit is wrong. Broader reviews (and even autism-cause inquiries) spotlight a needed return to open scientific debate.

[24:36] Next Week: Biotech History Tease

From penicillin back to early natural-product hunts—when scientists literally brought home dirt to find new drugs. A fun look at the explorer mindset that seeded modern biotech.

[26:28] Outro & Disclaimer

Standard disclosures; information believed reliable but not guaranteed. Hosts may hold positions mentioned.

This podcast and the information herein are intended for informational purposes only. The views expressed herein are the author’s alone and do not constitute an offer to sell, or a recommendation to purchase, or a solicitation of an offer to buy, any security, nor a recommendation for any investment product or service. While certain information contained herein has been obtained from sources believed to be reliable, neither the author nor any of his employers or their affiliates have independently verified this information, and its accuracy and completeness cannot be guaranteed. Accordingly, no representation or warranty, express or implied, is made as to, and no reliance should be placed on, the fairness, accuracy, timeliness or completeness of this information. The author and all employers and their affiliated persons assume no liability for this information and no obligation to update the information or analysis contained herein in the future.

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit telltales.substack.com

SHOWNOTES

A fast-moving episode: Hunt opens with Exhibits C/B/A (oil, gas, U.S. finances), then we sprint through a crisp history of AI from Turing to Transformers before closing on healthcare—Berkshire’s UNH stake and how MFN drug pricing pressures PBMs.

[00:00] Intro (—)

Welcome to Telltales with Mike, Jason, and Hunt. Grab the weekly Cash Flow Memo and follow along as we hit energy, technology, and healthcare through a cash-flow lens.

[00:57] Disclaimer (—)

Informational only; do your own work. No investment advice.

[01:22] Exhibit C: World Oil—Supply, Sanctions & the Risk Premium (Exhibit C)

Iranian output is running stronger than assumed while sanctions dynamics on Russia/India/China and “shadow fleet” enforcement remain key swing factors. Hunt frames oil pricing as fundamentals plus a $5–$20/bbl risk premium, shaped by geopolitical uncertainty and sanction compliance pathways (e.g., ship classing, SWIFT visibility). Saudi policy and the restoration of curtailed OPEC barrels are tracked against this risk overlay.

[07:10] Exhibit B: U.S. Gas & Producer Cash Flows (Exhibit B; p. 11)

Strip prices are modest vs last year, and Hunt notes how lower commodity realizations cascade into lower producer cash flow. Capital spending tends to track cash generation; he flags names on p. 11 where production rose despite tighter cash (EOG, Magnolia).

[07:56] Exhibit C (cont.): U.S. Shale & OPEC Strategy (Exhibit C)

U.S. crude peaked near 13.3 mbpd in ’24 and could drift lower if prices hang in the low-$60s. With OPEC+ restoring prior curtailments and Saudi’s voluntary cuts unwinding, non-U.S. and U.S. declines may eventually absorb surplus capacity—unless geopolitical risk fades abruptly.

[10:12] Exhibit A: U.S. Finances—Deficit Path, “All Other,” and Tariffs (Exhibit A)

Baseline shows the deficit improving from ~$1.9T to $1.5T, still elevated vs pre-COVID. “All other” outlays ballooned post-COVID; trimming continues to be the lever. Medicaid ($600B federal share) has been flat; tariffs add revenue at the margin, though sky-high projections are debated.

[13:21] Rates, the Fed, and the Long End (Exhibit A)

Debate: Should the Fed cut soon? Hunt expects policy rates drifting toward ~3–3.25% next year while the 10-year could stay ~4.5–4.75% given heavy Treasury supply and QT runoff. Jason worries about services inflation and sees labor data stronger beneath the surface; rate-cut timing remains contentious.

[18:43] Software History (Part 7): From Turing to Transformers (—)

Jason and Mike trace AI’s arc: 1950s Turing test; rule-based “expert systems”; 1990s machine learning and Deep Blue’s hybrid approach; then 2012’s ImageNet breakthrough as GPUs turbo-charged neural nets; finally 2017’s Transformer architecture underpinning today’s LLMs. Big idea: the future blends deterministic (classical) and probabilistic (AI) programming—knowing when to use each unlocks real productivity.

[31:53] Healthcare Quick Hit: Berkshire Buys $UNH (p. 19)

UnitedHealth pops onto the memo as large investors disclose stakes. The crew riffs on whether a Berkshire position could nudge governance and strategy—even as skepticism about UNH’s model lingers.

[32:48] MFN Pricing & the PBM Squeeze (p. 19)

“Most favored nation” pressure pushes pharma to sell more directly, compressing the gross-to-net gap historically captured by PBM rebates. If PBMs lose take-rate and transparency rises, margin structures across the channel could reset.

[33:50] MAHA, FDA, and Bending the Cost Curve (p. 19)

Positive read-throughs on FDA process changes; big win is healthier inputs (nutrition) that reduce downstream “sick-care” spend. The team sees a potential inflection toward better outcomes per dollar.

[35:14] Next Week & Wrap (—)

Biotech history is teed up for an upcoming 10-minute segment. Keep an eye out for stitched “software history” and “chip history” compilation episodes in the feed.

This podcast and the information herein are intended for informational purposes only. The views expressed herein are the author’s alone and do not constitute an offer to sell, or a recommendation to purchase, or a solicitation of an offer to buy, any security, nor a recommendation for any investment product or service. While certain information contained herein has been obtained from sources believed to be reliable, neither the author nor any of his employers or their affiliates have independently verified this information, and its accuracy and completeness cannot be guaranteed. Accordingly, no representation or warranty, express or implied, is made as to, and no reliance should be placed on, the fairness, accuracy, timeliness or completeness of this information. The author and all employers and their affiliated persons assume no liability for this information and no obligation to update the information or analysis contained herein in the future.

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit telltales.substack.com

SHOWNOTES

A wide-ranging tour this week: oil demand vs. spare capacity, U.S. gas and fiscal tracks, and the cloud era’s security shocks that boosted both AWS and CPU demand. We close on healthcare policy, Intel’s maneuvering, Nvidia’s China fee, and what Snowflake/Salesforce/ServiceNow signal for AI ROI.

[00:00] Intro

Mike sets the table for another 30-minute sprint through energy, tech, and healthcare, and points listeners to the Cash Flow Memo for exhibits and company pages.

[00:00] Disclaimer

Standard reminder: informational only; do your own work.

[00:00] World Oil Supply/Demand (Exhibit C)

Hunt notes oil “hanging in at 60” as OPEC+ countries aren’t meeting raised quotas—implying true spare capacity may be overstated. Demand growth has cooled: 2025 is up only ~0.5 mb/d vs. +1.4 mb/d in 2023, with China the swing factor (15.2 → 16.1 → 16.2 mb/d). This tempers bullishness even as quota compliance lags.

[02:08] U.S. Gas Demand/Supply (Exhibit B)

U.S. gas output averaged ~106 bcfd in 2025 and is running ~107; 2026’s 108 looks plausible. Gas-for-power use is flat YoY after years of +~1.5 bcfd growth, with fewer coal retirements and more solar capacity changing the stack. Policy/tax-credit shifts could slow new wind/solar adds, keeping an eye on balances.

[03:11] U.S. Fiscal Picture (Exhibit A)

Projected deficit improvement from ~$1.9T to ~$1.5T in FY26 reflects lower rates, modest spending trims, and targeted savings. Near-term risk is a pre-Oct 1 continuing resolution. Mike flags BLS job revisions softening the labor picture, and Hunt notes markets will parse the Fed Chair’s Jackson Hole speech for a potential 25 bp path.

[05:43] Software History Part 6: Cloud Computing

Jason traces AWS’s 2006 origin story (holiday-driven overbuild → rentable compute → managed services), plus REST and the API economy making zero-CapEx startups viable. OAuth tokens improved security UX but created token-theft risk; Spectre/Meltdown forced software mitigations that slowed CPUs ~10–20%—pushing some customers to buy more compute and organically lifting AWS spend.

[13:32] AWS Segment Breakout & Lessons (p. 1)

Hunt recalls the period when AWS was “other” before driving a re-rating as its profits eclipsed the legacy retail engine. The takeaway: durable, high-margin infra with years of head start can reset how the market values a conglomerate’s sum of parts.

[15:04] AWS Moat & Switching Costs (p. 1)

Mike and Jason discuss AWS’s growth and high switching costs; once landed, enterprise workloads tend to stay. Anecdotes include a CIA contract where AWS beat IBM on perceived security despite higher price, underscoring trust and managed security as part of the moat.

[18:00] Healthcare News & Policy (p. 15, 19, 20)

Items include the “Patients Deserve Price Tags Act” (hospital price transparency) and FDA disclosure trends (more CRL data) that could train LLMs and compress regulatory timelines. Noted too: an Eli Lilly exec voicing support for MFN dynamics to realize better EU pricing rather than U.S. cuts.

[19:47] Memo p.3: Intel—Governance, Strategy, and Scale (p. 3)

Hunt and Mike recap leadership scrutiny around China ties and a White House visit aiming to stabilize plans. Intel’s 10-Q flags that proceeding on the next leading-edge node likely requires a cost-sharing anchor customer—else more TSMC outsourcing. Scale economics dominate; policy levers (ownership stakes, tariffs) may be used to bridge the gap.

[23:03] Nvidia’s China Fee & Industrial Policy (p. 3)

Mike notes a 15% fee on Nvidia chips sold into China; Hunt/Jason debate authority, goals, and realpolitik (e.g., rare earths). The crux: keeping China dependent on Western silicon while managing leakage of high-end parts (H100/Blackwell) vs. allowing “dumbed-down” SKUs.

[26:14] Tech Updates: Tesla, Blackwell TCO, and AI in SaaS (p. 3-4)

Tesla launches a six-seat Model Y variant. Early reads suggest GB200 data-center TCO ~1.6× H100, so performance must scale accordingly to justify ROI. Salesforce appears slow to commercialize AI externally even as Benioff touts 1/3 of internal work done by AI; ServiceNow cites ~5% opex reduction with AI features gaining traction.

[29:12] What Does Snowflake Do (for AI)? (p. 4)

Mike/Jason explain Snowflake’s role unifying messy enterprise data (e.g., Salesforce) for BI and now AI—becoming the “source of truth” and a natural enclave to run/training workflows across clouds. Dirty upstream data can bottleneck AI adoption, which helps explain Salesforce’s challenges and Snowflake’s positioning.

[31:22] Wrap & Next Week

The team plans to extend software history and compare Salesforce, ServiceNow, and Snowflake’s AI monetization paths. Stay tuned for deeper dives on AI ROI across the enterprise stack.

Closing: If you found this helpful, grab the full Cash Flow Memo (with Exhibits A–C and company pages) and share the episode with a friend. Questions for the hosts? Drop them in the comments for next week’s mailbag.

This podcast and the information herein are intended for informational purposes only. The views expressed herein are the author’s alone and do not constitute an offer to sell, or a recommendation to purchase, or a solicitation of an offer to buy, any security, nor a recommendation for any investment product or service. While certain information contained herein has been obtained from sources believed to be reliable, neither the author nor any of his employers or their affiliates have independently verified this information, and its accuracy and completeness cannot be guaranteed. Accordingly, no representation or warranty, express or implied, is made as to, and no reliance should be placed on, the fairness, accuracy, timeliness or completeness of this information. The author and all employers and their affiliated persons assume no liability for this information and no obligation to update the information or analysis contained herein in the future.

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit telltales.substack.com

SHOWNOTES

A fast tour across energy, technology history, and healthcare—then a spirited debate on Tesla’s path: robotaxis or humanoid robots. Plus: where oil and gas balances sit now, what built Web 2.0, and why some healthcare names look interesting on cash flow.

[00:00] Intro

Welcome to Telltales with Mike, Jason, and Hunt, setting up a deep dive into energy, technology, and healthcare with the companion Cash Flow Memo. Quick reminder: this conversation is for informational purposes only.

[00:00] Disclaimer

Standard investment disclaimer on doing your own work and limitations of the information shared.

[00:00] Exhibit C: World Oil Supply/Demand (Exhibit C)

Hunt frames oil as a macro story: a cordial Trump–Putin meeting likely won’t yield a Ukraine ceasefire, keeping sanctions risk—and prices—elevated. Saudi output is marked up, Russia down; spare capacity sits near 3 mb/d. OPEC’s demand growth looks optimistic vs. the memo’s ~0.5 mb/d step-up, and the curve has flattened with less backwardation. Base case: WTI hovering in the low-to-mid $60s for now.

[02:49] Exhibit B: U.S. Natural Gas Supply/Demand (Exhibit B)

Gas is “in a swoon”: 2026 pricing hangs near ~$4 while the prompt month slides toward ~$3, as dry gas production ticks from ~106 to ~108 bcf/d. Associated gas from the Permian surprises to the upside; Gulf Coast hub pricing has even averaged under $1 at times, signaling localized oversupply. Expectations of lower Permian activity didn’t dent gas as much as hoped.

[04:17] Exhibit A: U.S. Government Finances (Exhibit A)

Deficit glidepath discussed: ~$1.9T this fiscal year toward ~$1.5T next year, aided by tariff/“excise & other” revenues and some spending flattening. Debt held by the public sits ~102% of GDP vs. ~81% pre-COVID (2019); bending that ratio down would be a major achievement. Musk’s provocative view on trimming the deficit by ~$1T is cited directionally.

[06:28] Agenda & Segue

The crew carves out time for two threads: software history leading to modern AI, and a healthcare check-in before closing on Tesla.

[06:45] Software History: Web 2.0 & the LAMP Stack

Jason tracks the move from Web 1.0 to user-generated Web 2.0: blogs, wikis, YouTube, MySpace/Facebook. Google’s PageRank shift—links and probability (Markov chains) over keyword counts—reshaped discovery. The LAMP stack (Linux, Apache, MySQL, PHP) plus permissive OSS licensing let startups ship dynamic apps cheaply and consistently. Proprietary stacks (e.g., .NET) faced friction from licensing and per-core pricing, nudging adoption toward open source.

[16:22] Healthcare News & Stock Updates (pp. 15, 19)

A surprise FDA leadership twist gets a mention. Harrow’s earnings show an inflection toward profitability: early data points suggest Vivi prescriptions outpacing Miebo among adopters, and another franchise growing >60% YoY with focus on retina practices. GLP-1 pills disappoint at Novo Nordisk and Eli Lilly. Vertex research results were mixed, but the core CF franchise underpins valuation comfort. Lantheus draws lessons on management change, capital allocation, and now screens cheap (7–8× FCF) with potential strategic interest.

[20:54] Sponsor: Oakcliff Sailing Update

Dawn checks in from the road with regatta results, upcoming events, and where to find race coverage on Oakcliff’s Facebook page.

[22:01] Robots vs. Robotaxis: The Tesla Debate (p. 1)

Is Tesla’s trillion-plus valuation driven more by robotaxis or humanoid robots? Jason sees robotaxis arriving sooner; Mike argues humanoids may be closer than consensus. Training data is a bottleneck for general-purpose robots; roads have clearer rules than “the world.” If Tesla can deliver ~$0.25/mile robotaxi costs, mass adoption could crush car ownership economics. Musk’s claim that Tesla could surpass Apple’s value is weighed against timelines: it likely takes a decade—and a lot of autonomy deployed—to get there. Waymo may be the other credible FSD player, but Tesla’s commercialization path looks stronger after shifting from rules-based to deep learning.

[33:13] Wrap & Tease

They’ll return next week to finish software/AI history and dig into chip stocks. Closing reminders and standard disclaimer.

Thanks for listening—grab the Cash Flow Memo for full exhibits, charts, and the 80-company update. If you enjoyed this episode, subscribe and share it with a friend who follows energy, tech, or healthcare.

This podcast and the information herein are intended for informational purposes only. The views expressed herein are the author’s alone and do not constitute an offer to sell, or a recommendation to purchase, or a solicitation of an offer to buy, any security, nor a recommendation for any investment product or service. While certain information contained herein has been obtained from sources believed to be reliable, neither the author nor any of his employers or their affiliates have independently verified this information, and its accuracy and completeness cannot be guaranteed. Accordingly, no representation or warranty, express or implied, is made as to, and no reliance should be placed on, the fairness, accuracy, timeliness or completeness of this information. The author and all employers and their affiliated persons assume no liability for this information and no obligation to update the information or analysis contained herein in the future.

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit telltales.substack.com

This week the Telltales crew unpacks shifting energy fundamentals, tough healthcare headlines, the rise of Web 1.0—and whether Tesla’s robo-taxi dream justifies a half-trillion price tag.

[00:00] Intro

Mike welcomes listeners, frames the show’s cash-flow focus across energy, tech, and healthcare, and reminds everyone to grab the accompanying memo.

[00:47] Exhibit C – World Oil Supply & Demand

Hunt revisits OPEC+ production moves, Saudi-U.S. coordination, and secondary sanctions aimed at Russia. He argues excess supply could keep WTI/Brent in the $50s before tightening 18–24 months out.

[03:55] Exhibit B – U.S. Natural Gas Outlook

Gas futures slip as shoulder-season demand meets rising output (~108 Bcf/d). Despite near-term weakness, Hunt still sees 2024–25 averaging ~$4/MMBtu if production stabilizes.

[05:05] Exhibit A – U.S. Fiscal Picture

With Congress on recess and spending bills stalled, Hunt warns that continuing-resolution drama could spook capital markets. Political gridlock plus Epstein-file intrigue add to the uncertainty.

[07:59] Technology & Healthcare Agenda

The hosts outline three focus areas—healthcare, software history, Tesla—and promise a deeper Tesla dive later in the episode.

[08:19] Healthcare Earnings & Policy (pp. 15)

Jason highlights FDA leadership changes, a 250 % pharma tariff, and halted mRNA research grants. Vertex’s pain-drug miss and Lantheus’s pricing pressure contrast with fortress balance sheets and fresh buybacks, sparking a value-vs-value-trap debate.

[15:56] Software History Part 4 – Network Effects & Web 1.0

From Microsoft Office lock-in to AOL’s ubiquitous CDs, Jason tracks the 1990s standards (HTTP, SSL, JavaScript) that enabled Internet scale—and set the stage for Amazon, eBay, PayPal, Yahoo, and Google.

[20:40] Oakcliff Sailing Update

A quick on-the-water report from Oakcliff International match-racing, with shout-outs to volunteers and rising teams.

[22:18] Tesla Sum-of-Parts & Robo-Taxi Debate (p. 1)

Hunt admires Tesla’s debt-light, cash-rich balance sheet but questions its $1 T valuation on just $5 B free cash flow. The trio models a $500-600 B option on robo-taxis and humanoid robots, comparing cost-per-mile economics to Uber and DoorDash, and muse on venture-style fleet financing.

[32:30] Wrap-Up & Next Week Teaser

Running long on enthusiasm, the hosts promise more on Tesla, Uber, and healthcare in the next episode—and sign off with the usual disclaimer.

Catch the full memo at telltales.us and join us next week for deeper dives into markets that move your portfolio.

This podcast and the information herein are intended for informational purposes only. The views expressed herein are the author’s alone and do not constitute an offer to sell, or a recommendation to purchase, or a solicitation of an offer to buy, any security, nor a recommendation for any investment product or service. While certain information contained herein has been obtained from sources believed to be reliable, neither the author nor any of his employers or their affiliates have independently verified this information, and its accuracy and completeness cannot be guaranteed. Accordingly, no representation or warranty, express or implied, is made as to, and no reliance should be placed on, the fairness, accuracy, timeliness or completeness of this information. The author and all employers and their affiliated persons assume no liability for this information and no obligation to update the information or analysis contained herein in the future.

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit telltales.substack.com

In this episode we swing from macro storm clouds to silicon sunshine—probing debt risks, tracing software’s evolution, and spotlighting a Samsung-Tesla chip alliance that could turbocharge robotaxis.

[00:00] Intro

:00:47] Macro Risk: Could 2008 Happen Again? (Exhibit A)

Hunt draws parallels between today’s oil-market risk premium and the 4Q 2008 liquidity freeze. He warns that a repo-market hiccup—not a bond auction failure—could be the trigger, given money-market–fund flight. Jason advocates 7 % across-the-board federal cuts, while Mike finds optimism in June’s surplus and $27 B tariff haul.

[00:10:19] Software History – Part 3

Jason and Mike trek from IBM mainframes to the Apple Lisa and Windows 3.0. Highlights: Thomas Watson Jr.’s $5 B System/360 gamble, GM’s early OS, the birth of “software engineering” (1968), Xerox PARC’s WIMP GUI, and Bill Gates’ pivot from DOS to a ground-up graphical Windows.

[00:19:21] Top Mark Rankings Spotlight

Tyler introduces the Top Mark Olympic-class sailing ranking—bringing the race to LA 2028 and positioning Top Mark Capital as the sport’s data hub.

[00:20:08] Healthcare Round-Up (pp. 15, 19, 20)

Jason covers Roche’s direct-to-consumer MS drug move, the FDA exit of controversial pick Pradeep Vemuri, and Novo Nordisk’s growth-guidance cut amid GLP-1 competition. Discussion centers on PBM mark-ups, gene-therapy approvals, and Lilly’s competitive edge.

[00:23:37] Tesla × Samsung: A Fab Deal (Page 1)

Hunt lauds Musk’s “highest and best use” as Samsung commits billions to an Austin foundry for Tesla’s in-house AI chips. Mike recounts Tesla’s silicon lineage—from Nvidia to Jim Keller’s HW 5—and explains why owning capacity beats vying with Nvidia at TSMC. Jason calls it Intel’s missed opportunity.

[00:28:04] Next Week Teaser & Sign-Off

The crew previews Episode 2532: a deeper dive into Tesla’s 2 M-unit robotaxi strategy and the consumer shock of humanoid-delivered DoorDash orders. Listeners are urged to grab the memo and tune in next Wednesday.

Thanks for listening! Download the memo at telltales.us, leave a review, and join us next week for more market storytelling.

This podcast and the information herein are intended for informational purposes only. The views expressed herein are the author’s alone and do not constitute an offer to sell, or a recommendation to purchase, or a solicitation of an offer to buy, any security, nor a recommendation for any investment product or service. While certain information contained herein has been obtained from sources believed to be reliable, neither the author nor any of his employers or their affiliates have independently verified this information, and its accuracy and completeness cannot be guaranteed. Accordingly, no representation or warranty, express or implied, is made as to, and no reliance should be placed on, the fairness, accuracy, timeliness or completeness of this information. The author and all employers and their affiliated persons assume no liability for this information and no obligation to update the information or analysis contained herein in the future.

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit telltales.substack.com