The Cashflow Memo
SHOWNOTES
In this episode of the Telltales Podcast, Mike, Jason, and Hunt walk through the weekly Cash Flow Memo: oil, gas, and US government finances, then dive deep into Taiwan Semiconductor’s history and the future of chip manufacturing, before closing with tech headlines and high-impact healthcare updates. Along the way they connect physics, policy, and business models back to cash flows and the durability of different franchises.
[00:00] Intro
[00:20] Disclaimer
[00:30] Exhibit C: Oil Demand Slows, Supply Investment in Focus
Hunt reviews Exhibit C and notes that oil prices and the forward curve are essentially flat, with no contango or backwardation. Sanctions on Russia and Iran appear to have limited impact, while demand growth is slowing as China plateaus and developed markets remain flat, shifting global oil demand growth from ~1.0 million barrels per day to roughly 0.5 million. With natural decline rates of 8–10% on a 100+ million barrel per day base, he argues that supply replacement needs are massive, and at ~$60 oil, supply growth is likely to level off, tightening the balance over time.
[02:55] Exhibit B: Natural Gas, LNG & Power Demand
Turning to Exhibit B, Hunt walks through the natural gas outlook, highlighting a well-anchored 2026 strip price around $4.10. LNG demand growth is robust, while gas-for-power demand has been flat as coal plants run harder and compete when gas prices rise. He expects power demand for gas to resume growth by 2026 as the “one-year” coal reprieve fades, while a more permissive permitting environment for LNG terminals should support steady volume growth. Political outcomes after the 2028 election could matter, but for now he views gas demand as “pretty good” for investors.
[04:32] Exhibit A: US Deficit, Tariffs & Shutdown Drama
Hunt briefly reviews Exhibit A on US government revenues and expenses, noting progress toward resolving a government shutdown. He points out that part of the deficit improvement is coming from higher tariff revenues, with excise and “other” categories reflecting roughly $250 billion in tariff-related receipts over the current fiscal window. A Supreme Court loss for Trump-era tariffs could force around $100 billion in refunds, but even after that, Hunt still expects the deficit trend to improve from here.
[05:29] History of TSMC & Morris Chang’s Foundry Revolution
Hunt sets up the “10-minute episode” on Taiwan Semiconductor, whose financials sit on page 3 of the memo, emphasizing its ~$23 billion free cash flow run rate even after ~$30 billion in annual capex. Mike then tells the story of Morris Chang: surviving multiple wars, studying engineering at MIT, and spending 25 years at Texas Instruments driving wafer yields from 10% to over 90%. Chang observed the stability of TI’s contract manufacturing side business and later proposed a radical model for Taiwan—a pure-play foundry that would manufacture chips for everyone without ever competing with its customers.
[08:31] TSMC vs the Old Guard: From “Real Men Have Fabs” to Neutral Foundry (p. 3)
Mike contrasts Chang’s vision with the vertically integrated titans of the 1980s—IBM, Intel, Motorola, NEC, AMD, and the “real men have fabs” mindset. While Intel doubled down on only making its own CPUs and later missed mobile, GlobalFoundries tapped out of leading-edge nodes due to capital intensity, and Samsung struggled to be seen as a neutral partner. TSMC quietly focused on process excellence and reliability, evolving from a modest fab into the default manufacturing partner for the world’s most ambitious chip designers.
[10:53] Apple, EUV & TSMC as a Geopolitical Chokepoint (p. 3)
The episode highlights Apple’s shift from Samsung to TSMC for its A-series iPhone chips, with TSMC effectively “betting the company” on Apple by building dedicated fabs and tuning processes around the clock. Mike argues this was a turning point that shifted the center of gravity in chipmaking to Taiwan, and once Apple trusted TSMC, companies like AMD, Nvidia, and Qualcomm followed. With EUV machines from ASML costing ~$350 million each and enabling features smaller than a virus, TSMC’s leading-edge fabs—clustered within 100 miles of mainland China—have become both an economic linchpin and a geopolitical choke point.
[13:45] The Future of Fabs: EUV vs X-Ray & Substrate’s Bet (p. 3)
Jason gives a physics primer on lithography, explaining how EUV systems use lasers and tin droplets to generate very specific UV wavelengths, pushing optical lithography to its limits around 13–30 nanometers. He contrasts EUV with X-ray approaches, noting that the industry chose EUV in the 1990s to extend traditional optics, but that advances in compact particle accelerators now make X-ray lithography more plausible. A secretive startup, Substrate, aims to build X-ray-based lithography and stand up a fab by 2028 that could produce wafers at one-tenth of TSMC’s cost—moving the challenge from pure engineering toward execution and capital.
[17:20] Economics, Power Needs & Who Adopts X-Ray Fabs First (p. 3)
The hosts discuss the capital intensity of an X-ray fab, with estimates running into the billions of dollars—Substrate has reportedly raised more than $100 million so far. Jason notes that while X-ray tools might simplify some elements of EUV, particle accelerators will still be power-hungry and fabs will face the same water and infrastructure constraints. Hunt muses that Nvidia’s Jensen Huang could be highly interested in any path to cheaper wafers, given Nvidia’s dependence on TSMC, and the team speculates that China may be exploring alternative lithography tracks in response to ASML export controls. They also note that ASML itself likely has X-ray work on the drawing board, and any true disruptor will have to be both technically superior and cheaper.
[22:50] Tech News: Amazon’s Data Center Build-Out & AI Power Constraints (p. 1)
In the tech news segment, Jason flags Amazon’s plan to double its data center capacity over the next two years, adding roughly 10 gigawatts—an enormous engineering and infrastructure task. Hunt wonders whether this implies Amazon will push CapEx beyond its already massive ~$90 billion run rate, and Mike expects higher spend. The hyperscalers are currently constrained by power availability for new data centers, but longer term they still see chips as the bottleneck, reinforcing how demand for AI compute flows back into semis and fabs.
[24:05] Uber + Nvidia Robotaxis & The AV “Three-Way Contest” (pp. 1, 3, 20)
Jason highlights a new Uber–Nvidia partnership to co-develop a fleet of around 100,000 autonomous vehicles built on Nvidia’s Drive platform and produced by OEMs such as Mercedes. Mike notes Mercedes has been working with Nvidia’s stack for years, making the tie-up logical. Hunt frames the robotaxi landscape as a three-way contest among Alphabet’s autonomous efforts, Tesla’s FSD, and Nvidia as the enabling platform behind multiple OEMs and fleets.
[25:10] Tesla FSD v13 Test Ride – Getting “Very Good” (p. 1)
Jason shares that the team took a 40-minute ride in a new Tesla Model 3 running the latest full self-driving software and, for the first time, it completed the trip with no observable mistakes. The anecdote underscores how quickly the software is improving and how that progress interacts with the chip and data center investments discussed earlier. For investors, it’s another data point in assessing the durability and optionality of Tesla’s software and autonomy revenue streams.
[25:31] Harrow: Lawsuit Overhang vs Branded-Drug Growth (p. 20)
Shifting to healthcare, Jason explains why Harrow’s stock has traded down after California filed a lawsuit over its compounding pharmacy, alleging deviations from good manufacturing practices. He notes that Harrow has addressed similar findings from the FDA in the past and continues to operate in California while negotiating a resolution, and that compounding is now a smaller part of the business. The core branded drug portfolio continues to perform well, especially its dry eye product, which large PBMs are adding to formularies—sometimes as the preferred first-line therapy—potentially lifting average selling prices as reimbursement replaces fixed cash-pay pricing.
[27:30] Harrow’s Melt Pharmaceuticals Deal & Capital Allocation (p. 20)
The hosts then discuss Harrow’s acquisition of the remaining stake in Melt Pharmaceuticals, a company originally spun out of Harrow that developed an anesthetic lozenge used in eye surgery and increasingly by dentists and oral surgeons. Phase 3 data so far look strong, and a second Phase 3 readout is pending before regulatory submission. Mike emphasizes that Harrow structured the deal with minimal upfront cash and milestone-based payments contingent on approval and sales, showcasing disciplined capital allocation that could give them full upside to a potentially significant product without overextending the balance sheet.
[28:57] Vertex: Kidney Disease Breakthrough & Fast-Track Timeline (p. 15)
Hunt tees up Vertex’s new kidney disease therapy, and Jason explains that the company presented positive Phase 2 results at a major conference, even though the trial wrapped up about a year ago. The drug targets a progressive kidney disorder that often leads to failure over 15–20 years, and Vertex’s molecule inhibits both key proteins involved, leading to best-in-class efficacy relative to competitors that only hit one. The Phase 3 program is already fully enrolled, the FDA has granted multiple expedited designations and accepted a rolling submission, and Vertex plans to use a priority review voucher—potentially enabling approval as soon as mid-next-year, with major implications for patients and Vertex’s revenue mix.
[31:16] Vertex’s Non-Opioid Pain Drug & Reimbursement Uncertainty (p. 15)
In closing out Vertex, Hunt asks about uptake of its non-opioid pain therapy, and Jason describes it as slower than hoped. He attributes the drag to reimbursement uncertainty: investors and management are watching to see whether coverage ultimately comes via the federal “No Pain Act” subsidy or through PBMs and private insurers. The recent government shutdown delayed expected policy clarity, illustrating how regulatory timing can impact the pace at which even clinically strong drugs translate into cash flows.
[31:52] FDA Leadership & Regulatory Stability
Jason notes that the FDA has appointed a new head of drug evaluation, a long-time insider who previously led the oncology division. He views this as a constructive pick for continuity and stability at an agency that sits at the center of many of the pipelines discussed on the show. Hunt wraps the episode by wishing listeners well and confirming they’ll be back in seven days with another memo and discussion, followed by the standard legal disclaimer.
To keep up with future episodes of the Telltales Podcast, download the weekly Cash Flow Memo at telltales.us and subscribe wherever you listen. If you’re enjoying these deep dives into cash-flow-driven investing across energy, technology, and healthcare, consider sharing the show with a friend or colleague.
$TSM $TXN $IBM $INTC $MSI $NIPNF $AMD $GFS $PHG $AAPL $SSNLF $NVDA $QCOM $ASML $AMZN $UBER $MBGAF $GOOGL $TSLA $HROW $VRTX
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?
This post 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.
