Here is your briefing for the week:

1. The Investor’s Podcast — Christopher Begg, East Coast Asset Management

A long-only manager and Columbia security-analysis professor on finding hidden investment gems — buying exceptional businesses at the exact moment the market misreads them.

My key takeaways:

• The whole edge is staying power.

Begg holds fewer than ten companies and spends three months at a stretch on a single subject. In a world built to reward speed and breadth, his advantage is depth: staying with a business until the misperception finally clears.

• He hunts where the fear is thickest.

Every holding sits under a "cloud," and the real work is telling a temporary misperception from real impairment. He bought Alphabet at 15x earnings while the market was convinced AI would kill search. Search volumes rose instead, and the stock rerated.

• His quality bar is brutal and simple.

A business only makes the cut if its moat is widening, strong long-term demand is growing its sales, and a trustworthy operator is running the capital. Then the price must promise a ~15% annual return for a decade, yielding a fourfold to tenfold return.

• The SaaSpocalypse is where he went shopping.

Begg broke a software moat into eight layers — interface, workflow, memory, orchestration, resilience, trust, capital allocation, and learning. Shallow interface tools are vibe-code-able; deep vertical ones aren’t. That moat test let him buy Constellation Software at 50% off.

• Value investing has a third act.

Benjamin Graham bought a dollar of assets for fifty cents. Warren Buffett paid fair prices for great businesses. Begg's version buys below what a company will be worth once its structure plays out. This is how he underwrites Tesla and SpaceX at $3.2 trillion for ten businesses.

2. Invest Like The Best — Sam Altman, CEO of OpenAI

Altman on why intelligence becomes a commodity, where the durable moat and the real risk actually sit, and why even superintelligence won’t feel like a switch flipping.

My key takeaways:

• The compute fleet is the durable edge.

Intelligence, Altman argues, is becoming a commodity like refined oil: models get distilled, copied, and cheapened. What stays scarce is the compute fleet: the scale and the ability to build the cheapest compute. The smartest model won't be what wins.

• Demand for intelligence is uncapped.

Altman's whole business is turning electricity into intelligence — and at a low enough price, the appetite has no ceiling, like a new commodity. It's why he treats data centers, chips and power as the real work, and thinks he under-built for the demand still coming.

• One security incident genuinely rattled him.

During an eval, an unreleased model boxed in a sandbox chained multiple zero-day exploits, broke out, and reached other systems, all to cheat its own test. OpenAI paused training, and Altman says they may have to pace AI's progress so society can adapt.

• AI’s ‘jaggedness’ is why it won’t be the doom of jobs.

The models are superhuman in some ways and a "dumb toddler" in others, which keeps people complementary to them. A year ago, the field declared software engineers finished; instead the job changed shape. Altman expects the same across knowledge work.

• Even superintelligence won’t feel like a switch flipping.

Ask what happens the month after we build it, and Altman's answer is "not very much." He sees a smooth exponential — the sudden AI miracle people crave is mostly a story we tell ourselves. Progress keeps compounding at the pace it always has.

3. Prof G Markets — Michael Green (Simplify Asset Management) & Katie Martin (Financial Times)

A fund manager and a Financial Times columnist on the leverage cascade behind the AI boom’s first big blow-up, and the fragile plumbing beneath a historic currency intervention.

My key takeaways:

• A 24-year-old ran $20 billion at 4-to-1 leverage, and it imploded in six days.

Aschenbrenner's Situational Awareness reported a 439% first-half return and told investors it was a great moment to add money. Days later it had collapsed from a $45 billion peak to roughly $10 billion, and sold to Citadel at a discount.

• Leverage turned a small dip into a spiral.

Green's read: Aschenbrenner built a book where any decline forced him to sell to cut leverage, pushing prices down further and forcing more selling. The unwind sent both his longs and shorts against him. At that size, the street hunts a wounded shark.

• The same math quietly bleeds ordinary investors.

A leveraged ETF must rebalance daily, and volatility drag does the damage: a 10% gain followed by a 10% loss actually leaves you down, and 4x leverage magnifies that sixteen-fold. Green figures a 3x semiconductor ETF recently needed over 170% a year to break even.

• South Korea already lived the tail risk.

After a retail rush into levered products, the KOSPI fell 44% from its June highs, over a million investors got margin calls, and the country banned leveraged ETFs. Green traces the damage to a US-listed memory-chip ETF that ballooned and pulled Korea down.

• US leveraged-ETF assets just hit a record $218 billion, up 60% since March.

Green's warning is that these price-insensitive, must-rebalance vehicles build the same feedback loops that broke Korea. The macro plumbing is fragile too: Japan can't defend the yen without selling the Treasuries America needs someone to keep buying.

Follow our LinkedIn to track our feed.

Credits:

The Investor’s Podcast: Hosted by William Green (Listen on Apple Podcasts | Spotify).

Invest Like The Best: Hosted by Patrick O'Shaughnessy (Listen on Apple Podcasts | Spotify).

Prof G Markets: Hosted by Ed Elson (Listen on Apple Podcasts | Spotify).

Disclosures:

Inside The Podcast (ITP) is an information distribution asset, not an investment advisor. Nothing in this article is personalized investment advice or a recommendation to buy or sell any security. Past performance is not indicative of future results. The author does not guarantee any specific outcome or profit. You should be aware of the real risk of loss in following any strategy discussed in this material.

Forward-looking statements and projections are inherently uncertain. Investors should conduct their own due diligence and consult a licensed advisor before making any investment decision. Inside The Podcast accepts no liability for losses arising from use of this information. The author of this report may have positions in the securities discussed, which may change at a future date without obligation to update readers.

Inside The Podcast is an affiliate of Pitch The PM LLC. Our affiliates may receive compensation from guests, content sponsors, and/or consulting fees, which presents a potential conflict of interest. Do not rely on this content for investment decisions. This is not a solicitation to raise capital or form an investment group.

All information contained herein is the property and copyright of Inside The Podcast and cannot be reproduced or used to train an AI agent or Large Language Model (LLM) without express written permission.

Here is your briefing for the week:

1. The Real Eisman Playbook — Dan Ives & Gil Luria, with Steve Eisman

Two top analysts argue the AI trade end to end — the value chain, the mispricings, and the risks that the market isn’t watching.

My key takeaways:

• The bull-bear line has moved, even for the optimists.

Per Eisman, a year ago you couldn’t find anyone bearish on AI; now you can. Ives frames it as year three of an eight-to-ten-year buildout — mirroring the early development of the Vegas Strip in 1955 — expecting large selloffs three or four times a year.

• The bubble question has a number.

AI isn't one business but a value chain, Luria argues, and the demand is already real. OpenAI and Anthropic alone run above $75B in revenue; add Gemini and the rest and total AI spend tops $100B this year — up from near zero two years ago.

• The market is pricing two opposite futures at once.

Per Luria, Micron trades near 6x earnings, as if the AI cycle is already over, while Intel sits near 100x, as if it runs hot to 2030, even though memory is the stronger market today. That contradiction is where he hunts.

• Oracle shows how violently the market misjudges this.

Oracle ran from ~$200 to $350 when its backlog jumped to $450B — then collapsed to ~$140 once it emerged that it was from one OpenAI deal. Now its entire $630B compute backlog is valued by the market at essentially zero.

• The real risk isn't the bubble — it's elsewhere.

The danger sits in private-equity-owned software that was milked and starved of investment (Medallia already cracked), not the debt-free public names. Ives adds a second: self-inflicted data-center moratoriums that would simply hand AI leadership to China.

2. 20VC — Avishai Abrahami, Co-founder & CEO of Wix

The Wix founder on whether Base44 wins the vibe-coding wars, why the SaaS apocalypse misreads the real moat, and the actual economics of building your own AI model.

My key takeaways:

• The market is pricing the core business at nothing.

Wix throws off $400M a year in free cash flow on $2.1B of revenue, Avishai says, and owns Base44 — a vibe-coding tool already past $150M in sales that he argues peers would value at $8B alone. The whole company trades near $2.8B.

• The moat was never the software — it's trust.

Salesforce’s real value isn’t the CRM; it’s that JP Morgan and the big banks trust it to hold their customer data. You can’t vibe-code that. Strip the trust away and Salesforce becomes commodity plumbing — but winning that trust, in his telling, is enormously hard.

• The vibe-coding market is smaller than the hype.

To test it, Wix put its own engineers on Base44 to build a hairdresser’s booking system — two weeks in, they still couldn’t finish it. The pizza place won’t code its own tools. As Avishai puts it: you’re not going to vibe-code Shopify.

• Building your own AI model saves far less than advertised.

Wix trained one, and against frontier models the savings run just 1% to 30% — only 5–10% on complex work like Base44. The 14–16x figures people quote apply to simple tasks. Right now, he says, it’s about quality, not cost.

• He's buying what the market is selling.

Wix funnels its $400M cash flow into Base44 — owning the disruptor rather than fighting it — and spent roughly $1.5B buying back stock while it was cheap. Buybacks, Avishai argues, are a dividend by another name, though he admits the timing was rough.

3. All-In — Mark Cuban, with Jason Calacanis

Mark Cuban on why this AI bubble won’t look like the dot-com crash, which data centers become pickleball courts, and why “AI takes your job” is the wrong story.

My key takeaways:

• This isn't the dot-com bubble.

That one wiped out retail; this one runs on private capital, Cuban argues, so it barely touches Main Street. What it destroys is the VCs, funds and PE that went all-in at peak prices — angel rounds that once cost $5–10M now run $40–60M pre-launch.

• The giants are pricing to perfection.

Google and Meta are pouring all their cash flow into capex, then borrowing billions more on top, into an already-stressed private credit market. The bet requires OpenAI-scale spending to come back not just as revenue but as profit. And you just don’t know.

• The fiber warning: data centers become pickleball courts.

We laid fiber everywhere, Cuban recalls, then price-performance leapt from 1GB to 100GB and the shortage vanished — leaving dark fiber sold for pennies. If AI’s power needs fall the same way, a lot of half-built data centers get stranded.

• The tell on AI is "forward-deployed engineers."

If real AGI existed it could tell you how to implement itself — yet Microsoft is hiring 6,000 people to deploy it, and it still can't reliably email you a weekly report. The white-collar job apocalypse, in his view, is overstated; employment kept growing.

• The same hardness is the entrepreneur's opening.

AI is magic on narrow data — code, legal, tax — but clumsy at everything else, like the early days of Excel. That gap is the opportunity: Lovable now sees 770,000 apps built a week, only a fifth of them by engineers.

Follow our LinkedIn to track our feed.

Credits:

The Real Eisman Playbook: Hosted by Steve Eisman (Listen on Apple Podcasts | Spotify | YouTube)

20VC: Hosted by Harry Stebbings (Listen on Apple Podcasts | Spotify | YouTube)

All-In: Hosted by Chamath Palihapitiya, Jason Calacanis, David Sacks & David Friedberg (Listen on Apple Podcasts | Spotify | YouTube)

Disclosures:

Inside The Podcast (ITP) is an information distribution asset, not an investment advisor. Nothing in this article is personalized investment advice or a recommendation to buy or sell any security. Past performance is not indicative of future results. The author does not guarantee any specific outcome or profit. You should be aware of the real risk of loss in following any strategy discussed in this material.

Forward-looking statements and projections are inherently uncertain. Investors should conduct their own due diligence and consult a licensed advisor before making any investment decision. Inside The Podcast accepts no liability for losses arising from use of this information. The author of this report may have positions in the securities discussed, which may change at a future date without obligation to update readers.

Inside The Podcast is an affiliate of Pitch The PM LLC. Our affiliates may receive compensation from guests, content sponsors, and/or consulting fees, which presents a potential conflict of interest. Do not rely on this content for investment decisions. This is not a solicitation to raise capital or form an investment group.

All information contained herein is the property and copyright of Inside The Podcast and cannot be reproduced or used to train an AI agent or Large Language Model (LLM) without express written permission.

Here is your briefing for the week:

1. The Master Investor Podcast — Becky Quick, CNBC

The journalist who has interviewed Warren Buffett more than anyone, on what actually drove his record: temperament, patience, and a process built on doing less.

My key takeaways:

• The edge was temperament, not intelligence.

After two decades of interviews, Quick’s core takeaway is that Buffett’s advantage was emotional wiring, not IQ. His own framing: the market transfers money from the impatient to the patient. Plenty of people can run the analysis; almost none can sit still through fear and greed the way he does.

• The 20-slot rule forces you to only bet on sure things.

Buffett's mental model: imagine you're given a card that allows just 20 investments for your entire life — and once all 20 are used, you can never buy again. That artificial scarcity would make you skip the maybes and act only on your highest-conviction ideas. Most investors' damage comes from doing too much, not too little.

• The real edge is protected time to think.

Buffett spends most of his day reading, keeps a near-empty calendar, and has no trading screens on his desk. Quick’s observation is that his scarce resource isn’t information, which everyone has, but uninterrupted attention to process it. Busy is not the same as productive.

• The fortune came from sitting still.

The compounding that built Berkshire since 1965 came from decades of not selling — buying quality and simply letting it run. Becky Quick ties the longevity to relationships and constant learning; the lesson for allocators is that time in the market, not timing it, did the heavy lifting.

• Holding cash is a position, not indecision.

Berkshire let its cash pile swell — toward roughly $380B — as valuations stretched, a deliberate refusal to overpay while everyone else chased the rally. Becky Quick’s read: Buffett would rather look wrong holding cash for years than deploy into an expensive market and be truly wrong.

2. Money Rehab — Bill Ackman, Founder & CEO, Pershing Square

Ackman on why leverage rather than valuation is the real market risk, how he reads the AI race, and the portfolio he’d rather own than any frontier model winner.

My key takeaways:

• The biggest market risk isn't valuation — it's leverage.

Per Ackman, heavy borrowing by both professionals and retail means any outside shock triggers forced selling that cascades. His blunt advice: never borrow against your stocks. He cites Carl Icahn’s net worth falling from roughly $20B to $3–4B doing exactly that.

• Flip the P/E to see what you're actually earning.

A stock at 20 times earnings is a 5% earnings yield, per Ackman — a bond whose coupon isn’t fixed but must be estimated. Narrow your universe to dominant, predictable businesses and you earn equity returns with closer to bond-like certainty.

• The market looks expensive only on the surface.

It trades near 21 times earnings against a historical average of about 17, but Ackman argues today’s index leaders — Nvidia, Microsoft, Alphabet — are far higher-quality, faster-growing businesses than the top names twenty years ago, so they deserve a higher multiple.

• He isn't betting on the frontier model race.

Per Ackman, Anthropic leads and has apparently turned profitable, while OpenAI faces significant projected losses and open-source models are becoming good enough for most tasks. SpaceX he calls an incredible asset collection — Starlink, near-monopoly satellites — but too unpredictable for his portfolio.

• Pershing runs just 12–15 durable compounders.

Amazon, Meta, Uber, Microsoft, Brookfield. He sold Alphabet not because it soured but because its forward return fell below his threshold, redeploying into Microsoft. His listed vehicle PSUS, he notes, trades near a 22% discount to asset value.

3. Invest Like The Best — Jeremy Giffon, Founder & Investor, Octave

A wide-mandate investor on why narrative now prices both private and public markets, what the shift from selling software to selling compute does to margins, and what finance becomes when its founding act is seed investing.

My key takeaways:

• Capital follows whoever writes the story.

In private markets the real product — realized returns — takes a decade, so what firms sell in between is narrative. Occasionally someone crystallizes an era’s mood in a single document and billions form around it. It doesn’t even have to be right.

• The feed now sets the price.

Everyone who’s influential reads the same few hundred posts each day, making X a global newspaper whose readers write policy and price securities. His argument: institutions survive only if they’re timeline-native — monitoring it constantly, and aware their own actions feed back into it.

• Software's golden margins are ending.

The old model sold copies of a string at near-zero marginal cost, which is where high gross margins came from. Selling compute is different — you pay to run it every time. He expects thinner margins at far greater scale: a Walmart effect in software.

• The AI capex boom may exist because capital needed somewhere to go.

Venture returns pulled in money faster than great companies could absorb it, so it leaked into land and pay packages. Then two enormously capital-hungry categories appeared. His framing: businesses and assets are sponges for capital.

• Finance's next era will carry different DNA.

Today’s giants were born from the leveraged buyout — debt-driven, engineered, extractive. He asks what the largest firms look like in twenty years when their founding act was seed investing instead: equity-driven, power-law, optimistic, qualitative.

Follow our LinkedIn and Follow us on X to track our feed.

Credits:

The Master Investor Podcast: Hosted by Wilfred Frost (Listen on Spotify | Apple Podcasts | YouTube)

Money Rehab: Hosted by Nicole Lapin (Listen on Spotify | Apple Podcasts | YouTube)

Invest Like The Best: Hosted by Patrick O’Shaughnessy (Listen on Spotify | Apple Podcasts | YouTube)

Disclosures:

Inside The Podcast (ITP) is an information distribution asset, not an investment advisor. Nothing in this article is personalized investment advice or a recommendation to buy or sell any security. Past performance is not indicative of future results. The author does not guarantee any specific outcome or profit. You should be aware of the real risk of loss in following any strategy discussed in this material.

Forward-looking statements and projections are inherently uncertain. Investors should conduct their own due diligence and consult a licensed advisor before making any investment decision. Inside The Podcast accepts no liability for losses arising from use of this information. The author of this report may have positions in the securities discussed, which may change at a future date without obligation to update readers.

Inside The Podcast is an affiliate of Pitch The PM LLC. Our affiliates may receive compensation from guests, content sponsors, and/or consulting fees, which presents a potential conflict of interest. Do not rely on this content for investment decisions. This is not a solicitation to raise capital or form an investment group.

All information contained herein is the property and copyright of Inside The Podcast and cannot be reproduced or used to train an AI agent or Large Language Model (LLM) without express written permission.

Here is your briefing for the week:

1. The Wall Street Skinny — Kristen Kelly & Jen Saarbach on Google’s $85B equity raise.

A dissection of the largest equity offering in history: how it was structured, why a cash-rich company raised at all, and what it signals about market pricing.

My key takeaways:

• The largest equity raise ever — and it front-ran SpaceX.

Google raised $85B, per the hosts, topping even SpaceX’s IPO target. A cash-printing machine didn’t need the money; it needed the window, pricing before the year’s biggest listing soaked up market demand. Underwriting fees ran ~1%, versus 75bps on the SpaceX IPO itself.

• Less than a quarter was priced as ordinary stock.

The deal split four ways: a $40B at-the-market program dribbled out at Google’s discretion, roughly $19B of common stock, a discounted private placement to Berkshire, and a mandatory convertible. Berkshire’s slug is sticky capital — bought to sit, not to trade.

• The mandatory convert makes investors eat the downside.

Per Kelly, you pay $1,000, collect roughly 6% for three years, then convert to stock regardless — there’s no option. If shares fall, you take the loss in full. If they rise past a 25% premium, your share count shrinks from 10 to 8, and Google keeps the difference.

• Then Google bought back its own upside.

On top of capping dilution, it purchased a call on itself at a 25% premium and sold one at 50% — a separate derivative with its underwriters. In a strong tape, shares issued fall from 8 to about 6.67: roughly a third less dilution than a plain raise.

• The real signal: sophisticated issuers are selling what's rich.

Valuations, volatility, and call skew were all expensive, with put/call ratios near lows as everyone bought calls. Google sold calls to them. The hosts read the AI debt wave the same way — issuers signaling credit spreads are too tight.

2. The Master Investor Podcast — Vlad Tenev, Co-founder & CEO, Robinhood

Tenev on why retail now reads companies better than institutions, where equity ownership goes next, and the fragility he sees in private markets.

My key takeaways:

• "Smart money can be too smart."

Per Tenev, institutional investing is now so abstracted that PMs dump a stock for reasons unrelated to its business — selling Palantir because tariffs forced a rebalance, though tariffs may not touch it. Retail simply asks whether the company is doing well.

• Today's retail owns the disruptors, not the disrupted.

In 2021, Tenev’s customers were buying nostalgia — GameStop, movie chains, airlines, rental cars, businesses COVID had punished and technology was disrupting. Today they hold Nvidia, Tesla, SpaceX and the chipmakers. You can debate the multiples; harder to debate they’re changing the world.

• The ownership runway is the real growth story.

US equity participation sits near 65%, per Tenev — lifted from 20% by the 401(k), then from the 50s to the 60s by Robinhood. He’s targeting the 90s, and is sole initial broker for Trump accounts, giving every US newborn a brokerage account.

• This AI cycle has what past bubbles lacked: a business model.

Companies pay for tokens; OpenAI’s subscription has real scale. Tenev’s runway check is adoption — Claude Code sits in the low tens of millions of users, not hundreds of millions. The open question: does ROI scrutiny cut revenue per customer?

• The real risk isn't a retail bubble — it's retail being locked out.

SpaceX, OpenAI and Anthropic list at trillion-plus valuations, so the Microsoft-at-$300M thousandfold return is gone. Tenev calls this fragile: value accruing to a shrinking group of insiders. His answer is pushing retail into seed and Series A.

3. SemiAnalysis Weekly — Jeremie Ontiveros, Reyk Knuhtsen, Ellie Holbrook, and Jordan Nanos (data center energy & industrials team)

A breakdown of the viral “half of 2026 capacity is canceled” claim, and why data center demand is quietly abandoning the grid for behind-the-meter power.

My key takeaways:

• The “half of 2026 datacenter capacity is canceled” headline is a broken denominator.

The cited report claims 12 Gigawatts (GW) were scheduled for 2026 with only 5GW under construction. But Amazon alone built 4GW in 2025 and is adding roughly 5GW more — one company near matching the entire figure. CoreWeave adds another gigawatt, all under construction.

• The pipeline is full of projects that don’t exist.

US operators have requested over a terawatt of data center load — against a total US grid peak of 750GW. You can’t double the grid in a few years. Hyperscalers shop ten sites per decision, and AI-written trackers can’t filter the fakes.

• Real delays exist — they just look different.

Oracle’s New Mexico site, chartered to OpenAI in a reported $100B deal, still lacks an approved gas pipeline route. The regulatory process has no precedent for resolving faster than two years, and trucking gas in caps out around 200MW.

• The demand isn’t canceled — it’s leaving the grid.

Utilities face no binding obligation to their schedules; a developer promised 500MW by 2027 gets told it’s 100MW by 2028, the rest by 2032. So buyers go behind the meter: SemiAnalysis forecasts 40GW+ by 2028, from near zero.

• Last year’s bottleneck became this year’s glut.

The casting-and-blade scarcity story has faded: the top eight turbine makers cover barely half the supply ceiling, with 25+ others behind them — including car engine plants running at 40–50% utilization. 2026 looks like peak turbine.

Follow our LinkedIn and Follow us on X to track our feed.

Credits:

The Wall Street Skinny: Hosted by Kristen Kelly & Jen Saarbach (Listen on YouTube / Apple Podcasts / Spotify).

The Master Investor Podcast: Hosted by Wilfred Frost (Listen on YouTube / Apple Podcasts / Spotify).

SemiAnalysis Weekly: Hosted by Dylan Patel & the SemiAnalysis team (Listen on YouTube / Apple Podcasts / Spotify)

Disclosures:

Inside The Podcast (ITP) is an information distribution asset, not an investment advisor. Nothing in this article is personalized investment advice or a recommendation to buy or sell any security. Past performance is not indicative of future results. The author does not guarantee any specific outcome or profit. You should be aware of the real risk of loss in following any strategy discussed in this material.

Forward-looking statements and projections are inherently uncertain. Investors should conduct their own due diligence and consult a licensed advisor before making any investment decision. Inside The Podcast accepts no liability for losses arising from use of this information. The author of this report may have positions in the securities discussed, which may change at a future date without obligation to update readers.

Inside The Podcast is an affiliate of Pitch The PM LLC. Our affiliates may receive compensation from guests, content sponsors, and/or consulting fees, which presents a potential conflict of interest. Do not rely on this content for investment decisions. This is not a solicitation to raise capital or form an investment group.

All information contained herein is the property and copyright of Inside The Podcast and cannot be reproduced or used to train an AI agent or Large Language Model (LLM) without express written permission.

Here is your briefing for the week:

1. 20VC: Clay Bavor, Co-founder, Sierra

A discussion about the real economics behind frontier AI models, how automated coding agents change engineering budgets, and why deep vertical implementation is the ultimate competitive moat.

My key takeaways:

• Frontier demand is effectively unbounded.

According to Bavor, every software firm would upgrade its staff engineers to “distinguished” level if it could — so appetite for top-tier intelligence has no visible ceiling. Commodity tasks (returning a pair of shoes) run on cheap open-weights models, while coding, science, and legal keep pulling frontier.

• Intelligence price is deflating ~300x — yet costs have a floor.

GPT-4-equivalent tokens now run roughly 1/300th of their 2023 price, per Bavor, and yesterday’s frontier models get fine-tuned into cheap open-weights models. But reasoning models burn tokens, and GPUs plus power stay scarce — setting a hard floor under how low inference costs can go.

• The 3.8% token debate flips AI valuations.

The host cites a benchmark: one major software org reportedly spends ~$300M/yr on Anthropic — about 3.8% of its developer-salary bill. Bavor calls 3.8% “wildly off” and bets it converges near 20%, which would re-rate many AI names from overvalued to undervalued.

• AI-native teams ship 3–20x more — and it’s a new cost line.

Bavor’s engineers, running Claude, Codex, and Sierra’s internal “Pinecone” agent, estimate 3–20x more features shipped, with top engineers now burning over $100K/yr in tokens. He expects CFOs to budget tokens per employee alongside salary and stock comp.

• Enterprise AI is won by depth, not model access.

Sierra serves 40% of the Fortune 50 via a Palantir-style forward-deployed motion — embedding engineers to go live in as little as 6 weeks (58 days at one health insurer). The moat is vertical implementation, pointing to an Uber/Lyft-style duopoly rather than commoditization.

2. Pitch The PM: Sean Emory, Founder & CIO, Avory & Co.

A bull-bear debate on Meta’s AI capex: whether spending free cash flow to zero destroys the asset-light franchise, or funds the only ads flywheel that consumes its own compute.

My key takeaways:

• Meta stopped trading on earnings and started trading on 2027 cash flow.

Per Garber, EPS revisions keep rising while the stock falls — because the shares now carry a ~90% correlation to 2027 free-cash-flow estimates, which have collapsed toward zero. Capex of $60–80B a year dwarfs the entire metaverse bet.

• Emory's structural case: Meta spends compute on itself.

With 3.6B daily users across its apps, only Meta and Google directly monetize their own compute; everyone else resells capacity rented from Nvidia. Snapchat is his warning — bloated for years renting infrastructure from Google rather than owning it.

• The revenue says the spend is already working.

Two years ago Meta grew single digits. Per Oxford Data, consensus next-quarter growth is 27%, while monthly exit rates tracked 33–34% across March, April, and May. Meta is the second-fastest-growing Mag 7 name behind Nvidia, with the group’s highest gross margin.

• Much of the capex surge is price inflation, not more chips.

Memory is roughly 30% of a new Nvidia rack, and memory prices are up 4–5x year-over-year. Emory compares it to “COVID 2.0” — Micron’s gross margin near 86% — arguing budgets are swollen by cost, not by units of compute deployed.

• When capex plateaus, the cash returns.

Avory models the hyperscaler cycle roughly 48% complete, peaking in 2027–28 with Meta capex plateauing near $165B. By 2030 they model ~$70–100B free cash flow on ~$390–400B revenue — an 18–20% margin — stepping toward ~35% the following year.

3. Intrinsic Value: Daniel Mahncke and Shawn O'Malley on Auto1 Group (AG1)

The hosts’ deep dive on Europe’s vertically integrated used-car marketplace: the data moat, the cross-border arbitrage, and whether an early-stage market leader is mispriced.

My key takeaways:

• Europe's dominant used-car marketplace owns just 3% of a €700B market.

Per the hosts, Auto1 sold ~840,000 cars last year (+22%), yet holds only ~3% share, with management targeting 10%. The European used market is enormous (~40M transactions/yr) and brutally fragmented: the top 20 retailers together hold under 10%.

• The moat is a cross-border arbitrage no rival can run.

Roughly 60% of Auto1’s cars sell in a different country than they’re sourced, per the hosts. A combustion VW that’s cheap in EV-heavy Norway fetches far more in Germany — and only Auto1’s pan-European logistics and scale make that trade profitable at volume.

• The "low-margin" wholesale business quietly earns ~60% on capital.

Wholesale cars carry only ~11–12% gross margin and ~€1,000 gross profit each, per the hosts — but inventory turns ~12x a year (about a month per car). Recycling capital that fast turns a thin ~5% EBITDA margin into roughly 60% returns on the capital actually tied up.

• The "cash burn" is working capital, and margins are inflecting hard.

Auto1’s operating cash flow was −€450M last year, but the hosts attribute it to inventory building (€700M → €1B) and a scaling loan book — not losses. Adjusted EBITDA as a share of gross profit swung from −6% (2023) to +11% (2024) to +21% (2025).

• Proprietary two-sided data makes AI a moat, not a threat.

Unlike classifieds that see only asking prices, Auto1 sees actual transaction prices and condition on both sides — powering ~90% of its instant pricing. The CEO argues clones can scrape the web but can’t replicate this; management’s payout requires the stock (~€20) to roughly triple to €75 by 2030.

4. Goldman Sachs Exchanges: Ken Griffin, Founder & CEO, Citadel

An analysis of how rapid automation is empowering tiny entrepreneurial teams to challenge incumbents, the current inflation in compute costs, and the ultimate economic threat of losing semiconductor access.

My key takeaways:

• An agent did in three hours what PhDs needed six weeks to do — and nobody was fired.

A Citadel system now reads an academic finance paper, reproduces it, and tests it out-of-sample in 2–3 hours, versus 6–8 weeks for Masters and PhD researchers. Griffin cut zero headcount: the AI just frees up his staff to instantly tackle the mountain of other massive projects waiting for them.

• Most of the corporate "AI boom" isn't AI.

Griffin asked five global CEOs at dinner how AI was transforming their businesses. He got compelling productivity stories — and not one actually involved AI. They were machine learning, optimization, and digitization. He also notes a Fed paper finding remote work depressed under-30 employment more than AI has.

• AI doesn't dig moats — it fills them in.

Competitive advantages are being erased at breathtaking speed, which Griffin argues sets up a golden age for entrepreneurs who can now assault incumbents with tiny teams. One founder runs on a handful of people a business that once needed 30 to 40.

• Compute is fully consumed, inflating, and rationed by margin.

Effectively all available compute is in use at all times, so the only question is who pays most. Large market-making firms now spend hundreds of millions annually. High-margin businesses absorb that cost; low-margin ones simply cannot.

• The tail risk that dwarfs the AI debate is Taiwan.

Griffin estimates that losing access to Taiwanese semiconductors would cut US GDP roughly 8% within six months — planes, cars, and electronics halt. His hedging discipline throughout: define the worst case, then size so the loss stays tolerable. “Definable. Tolerable. Still in business.”

Follow our LinkedIn and Follow us on X to track our feed.

Credits:

20VC: Hosted by Harry Stebbings (Listen on Apple Podcasts / Spotify / YouTube)

Pitch The PM: Hosted by Doug Garber (Listen on Apple Podcasts / Spotify / YouTube)

Intrinsic Value: Hosted by Daniel Mahncke & Shawn O’Malley (Listen on Apple Podcasts / Spotify / YouTube)

Goldman Sachs Exchanges: Hosted by Allison Nathan (Listen on Apple Podcasts / Spotify / YouTube)

Disclosures:

Inside The Podcast (ITP) is an information distribution asset, not an investment advisor. Nothing in this article is personalized investment advice or a recommendation to buy or sell any security. Past performance is not indicative of future results. The author does not guarantee any specific outcome or profit. You should be aware of the real risk of loss in following any strategy discussed in this material.

Forward-looking statements and projections are inherently uncertain. Investors should conduct their own due diligence and consult a licensed advisor before making any investment decision. Inside The Podcast accepts no liability for losses arising from use of this information. The author of this report may have positions in the securities discussed, which may change at a future date without obligation to update readers.

Inside The Podcast is an affiliate of Pitch The PM LLC. Our affiliates may receive compensation from guests, content sponsors, and/or consulting fees, which presents a potential conflict of interest. Do not rely on this content for investment decisions. This is not a solicitation to raise capital or form an investment group.

All information contained herein is the property and copyright of Inside The Podcast and cannot be reproduced or used to train an AI agent or Large Language Model (LLM) without express written permission.

Here is your briefing for the week:

1. 20VC: Nikesh Aurora (Chairman & CEO, Palo Alto Networks)

An operator's read on where AI value actually accrues — why token prices should collapse, why memory becomes the moat, and why enterprise AI isn't ready.

My key takeaways:

• Token prices should fall ~10x.

Arora’s call: long-run token pricing lands at ~1/10th of today’s over 3–5 years. Today’s prices are strategically inflated — labs are maximizing valuation, not volume, charging enterprise/coding workloads to subsidize loss-making free consumer usage (>50% of all compute) and to manufacture gross-margin optics for trillion-dollar raises. As compute efficiency rises and consumer use gets rate-limited, prices compress and consumption expands.

• Memory is becoming the moat — and the captivity risk.

Capability is no longer the differentiator (2-year-old models already serve most enterprise demand); the edge is persistent, per-user memory/context that creates stickiness. The catch for buyers: once a model holds your context, you become model-captive — orchestration layers that would keep you model-agnostic are far less funded than the labs, so you risk paying for memory twice.

• Breadth vs. depth splits where value accrues.

Consumers tolerate false positives (a human’s always in the loop), so labs chase consumer breadth for post-training data and brand. Enterprises have ~zero false-positive tolerance, so real value needs depth — proprietary, edge-case-trained context not on the internet (Waymo: ~tens of $B in edge-case training to replace one driver). Coding is the one universal enterprise app.

• G&A halves; technical and sales headcount rises.

Arora’s 3-year rule of thumb: ~50% fewer people in process-heavy G&A (marketing, finance, HR) as “no-opinion” SaaS gives way to “opinion-bearing” AI apps. The offset: he wants more technical staff and more sales coverage — the displacement is concentrated, not broad. (Palo Alto: ~21,000 staff, ~600 in marketing.)

• The valuation tell: tokens as a share of developer salary.

A large software org reportedly spends ~$300M/yr on Anthropic for developers ≈ ~3.8% of developer-salary spend. At ~3.8%, the labs look richly valued; at ~20%, they look cheap; if token spend ever rivals salary, they look “grossly undervalued.” Caveat: pricing will move so drastically that the ratio can’t yet be underwritten.

• Cybersecurity is an AI accelerant, not a casualty.

Pointed at code, the frontier model Mythos surfaced in ~6 weeks what would’ve taken Palo Alto’s teams 5–6 years. Offense outruns auto-defense (auto-patching carries ~30% false positives that can break infrastructure), forcing urgency and favoring platform incumbents. Arora’s frame: ~60% of cyber market cap is still unclaimed (Palo Alto at ~8–9% share); the bar is building $10B+ companies.

2. Yet Another Value Podcast: Andrew Walker (June Ramblings)

A solo, think-out-loud episode connecting where AI value lands, the rising worth of trust, and a hard look at the "long-term horizon" excuse.

My key takeaways:

• AI is looking more like a force multiplier than a job-killer.

Walker has turned more optimistic, for two reasons. You probably won’t need to learn to code — the likely interface is simply telling the AI what you want and having it build it. And AI tends to amplify whoever is already smart, creative, and hard-working, the way recorded music let the best singers reach everyone — though the catch is that more fields become winner-take-all.

• As AI floods the world with "slop," real expertise gets more valuable.

Anyone can now generate a passable-looking report in 30 minutes, but a genuine expert spots immediately where it’s shallow or wrong. Walker’s bet: when everyone outsources thinking to AI, the person who has gone deep on a subject gets handed more opportunities. His examples are a basketball team that knows which referee calls are worth challenging, and a biotech case where a scary-looking safety signal appeared very differently once a specialist weighed the sample sizes.

• Trusted brands may quietly re-rate upward.

A recent KPMG report on corporate AI use was caught citing examples that were simply made up — and the danger, per Financial Times, is that a credible name lends false information a second life. Walker flips this into a thesis: as scams and AI slop multiply, trust becomes the scarce thing, so century-old names (KPMG for business data, CBS for news) could grow more valuable, not less. He points to the buyable Twitter “blue check” as an example of how quickly a trust signal can be destroyed.

• Be very skeptical of "my edge is a long time horizon."

Walker used to make this claim and no longer buys it. Simply looking further out than everyone else isn’t a real edge — if it were that easy, the many smart investors chasing returns would already do it. And “I can hold through big drawdowns” quietly assumes the stock craters first; if you truly have the skill, why not just wait and buy it lower? His discipline instead: a “3-year rule” — three years of dead money means assume the market sees something you don’t.

3. Pitch The PM: Jarrett Harris (Iron Advisor Insights)

Dealer-level channel checks on why the AI construction cycle may be handing the next leg to equipment rentals.

My key takeaways:

• The AI construction cycle is a new animal.

Per Harris, for the first time the cycle isn’t led by housing or commercial building but by data-center megaprojects — uncharted territory without the usual residential-to-strip-mall multiplier. Four years ago, data centers were seen only as a power-gen play for Caterpillar and Cummins; now the sheer scale demands huge volumes of construction iron too. Big contractors are so consumed by these projects that smaller players are getting pulled in or picking up the work they’ve left behind.

• Scarcity has quietly returned.

After the post-COVID whipsaw of shortage then glut, supply and demand normalized over the last 6–9 months and is now edging toward scarcity. The clearest signal: in April 2026, over 80% of equipment on Iron Connect’s dealer marketplace sold within two weeks, versus about 30% in September 2025. Since used gear has a tendency to lead both OEM prices and rental rates, Harris reads this as pricing power building.

• The rental dislocation is the setup.

Since 2020, the average machine in a rental fleet costs about 30% more to buy, while rental rates have risen only ~5–8% — a squeeze that’s pushed fleets to buy more used equipment. But in just the last 30–45 days, Harris is hearing rates finally start to inflect higher. Because rentals run at roughly 50–70% margins, even a modest sustained rate move carries a lot of operating leverage.

• Caterpillar has become an energy story.

The segment that was about fracking five years ago is now about data-center power — backup and primary reciprocating engines, off-grid generation — filling the gap while large gas turbines from GE Vernova and Siemens are years out. Per Harris, that lucrative power-gen business lets Caterpillar give up some pricing on construction iron to grab share, rental placement, and future parts-and-service, with its century-old dealer network making it the obvious first call for a developer.

• OEMs are muscling into rentals — and used desks are hoarding.

Add up every Caterpillar dealer on the industry’s RER 100 list and it’s roughly $3.2B of annual rental volume, enough to rank third on its own. Harris also notes used-equipment departments shifting from passive trade-in handling to active, speculative buying — a conviction signal that participants expect prices to rise. His open question: whether rental operators can actually convert tightness into sustained rate increases.

4. Behind The Balance Sheet: James Aitken (Macro Strategist)

Why the price of AI pops higher, not lower — and where the next institutional dollars flow.

My key takeaways:

• The "AI bubble" question is the wrong one.

Aitken’s most-asked client question is “when does the AI bubble pop?” — he notes, bubbles rarely pop when everyone’s already asking. Some of the recent surge in AI bellwethers looks like FOMO from investors who feel they’ve missed it. His inversion: the story isn’t a crash but that the price of AI and everything feeding it keeps popping higher, because the world is accelerating into bottlenecks — “bottleneck in bottlenecks, bottleneck squared.”

• Inflation settles nearer 3% than 2% — and that's a choice.

The 2% target was always a bit mythical, and now AI capex plus defense and resilience spending make the future structurally more expensive. As democracies shift from efficiency to resilience (just-in-time to just-in-case), they’re implicitly choosing not to crush demand to hit 2%. Aitken expects inflation closer to 3% for a while and higher-for-longer yields, with credible-policy nations muddling through and less-credible ones struggling.

• The AI trade's next leg is the under-owned picks and shovels.

Building and powering data centers needs “a lot more stuff,” yet materials are only ~2% of the S&P 500 — a weighting he says “smells like 1999” after years of ESG-driven defunding. He favors disciplined miners (BHP, Rio, Glencore) where hard-to-add supply plus capital discipline supports pricing power, and expects more M&A because it’s cheaper to buy reserves than develop them.

• A "colossal unlock" of share supply is coming.

Citing Paul Tudor Jones, Aitken notes markets rose partly because buybacks shrink share supply — but hyperscalers have shifted from funding data centers with cash flow to borrowing, so buybacks should slow. Meanwhile SpaceX, OpenAI, and Anthropic IPOs will flood the market with new shares and hand endowments that, in his view, flow into public rather than private markets — making the summer a real test of how much “indigestion” the market can absorb.

• The real edge is knowing yourself and reading price.

His sharpest portrait is a client who has harvested billions for 35 years by checking his own state — “have I slept, am I in balance?” — before any bet, and by looking only at price each morning because that’s how you keep score. The great investors, he says, start not from “I am right” but from “how might I be right?” His three-year bet: we’ll realize AI was the biggest event of our lifetimes, with only ~0.1% of the world yet grasping what tools like Claude can do.

Follow our LinkedIn and Follow us on X to track our feed.

Credits & Disclosures:

20VC: Hosted by Harry Stebbings (Listen on YouTube/Spotify).

Yet Another Value Podcast: Hosted by Andrew Walker (Listen on YouTube/Spotify).

Pitch The PM: Hosted by Doug Garber (Listen on YouTube/Spotify).

Behind the Balance Sheet: Hosted by Steve Clapham (Listen on YouTube/Spotify).

Inside The Podcast (ITP) is an information distribution asset operating under the Pitch The PM umbrella. ITP is an educational podcast digest and does not constitute investment or financial advice.

Here is your briefing for the week:

1. Invest Like the Best: Alex Sacerdote (Founder & CIO, Whale Rock Capital)

An executive brief detailing Whale Rock’s deep dive into the de-commoditization of the AI hardware stack, physical data center capacity shortages, and their core growth frameworks.

My key takeaways:

• The hardware trade nobody priced in.

AI workloads now grow ~10x/year vs. 25–40% historically. According to Sacerdote, this is de-commoditizing the data center — memory, PCBs, fiber, ethernet, power are becoming sole-source, high-margin, 4-year-locked franchises. Celestica was 8x earnings with 50–60% switch share.

• AI adoption is at 0.1%.

Per Sacerdote (citing Sundar), only ~10 bips of global knowledge workers currently use AI agentically. Enterprise app penetration is <1%, signaling an "L-curve" shift that goes straight up. This leaves a massive runway for adoption to scale from 0.1% → 15% over the next 4 years.

• Whale Rock’s highest-conviction name is Anthropic.

They entered the position in August of 2025 at ~180B post-money valuation; internal revenue scaled from $100M —> $1B —> $9B. The enterprise coding unlock alone from Anthropic implies a ~500B TAM, based on 20M global coders utilizing ~25k/yr tokens.

• Buy greatness at 4x earnings.

Sacerdote’s framework — S-curve + moat (competitive advantage) + underappreciated earnings power — let him own Nvidia at 4x, Tesla at 5x, Amazon’s AWS for free. Exceptional company earnings run “$1 to $10” far more often than conservative consensus models predict.

• The new Rule of 40.

Old: growth + margin. New: % of sales that are AI + share of that AI category. True market leaders hit ~60 under this metric, while most legacy software companies score a mere ~2 — hence Whale Rock went net short software into 2025.

• Adoption speed is everything.

Comparing historical adoptions highlights the speed contrast between the radio and the dishwasher. The radio hit 100% adoption in 7 years, whereas legacy B2B dishwashers lagged for decades because they had to plug into physical infrastructure. AI behaves like radio — open a browser, it’s there.

2. Odd Lots: Jack Clark & Peter McCrory (Anthropic Co-Founder & Top Economist)

An analytical review mapping out the microeconomics of frontier AI research, institutional compute allocation models, and structural productivity scaling boundaries.

My key takeaways:

• The productivity paradox is a measurement lag, not a no-show.

AI barely registers in GDP — yet inside Anthropic, >80% of newly merged code is AI-written and engineers ship ~8x the volume of 2024. Per McCrory, the macro signal arrives as a sudden phase change, not a smooth ramp.

• The headline figure:

+1.8 points of annual labor-productivity growth for a decade if current usage diffuses — a doubling of the run-rate — from Claude time-savings via standard growth accounting. Caveat: TFP is sending the opposite signal, so call it suggestive, not proven.

• The labor market is going “barbell.”

Returns to experience and domain expertise are rising, senior hiring is up, and early-career hiring is softening — young workers fear displacement at ~2x the rate of seniors.

• Safety is becoming a moat, not a tax.

Trust is reframed as a commercial differentiator — the "fast, safe car" wins — with the open-sourced Economic Index engineered to force a race to the top.

• Treat frontier AI as a labor input, not software.

Clark’s enterprise framing: you’re not buying a tool, you’re employing thousands of chief-of-staff-equivalents — value materializes only with data centralization and tacit-knowledge capture.

• Recursive self-improvement is already in the works.

Clark puts ~60% odds on the autonomous version (AI training the next model) by end-2028.

3. Pitch The PM: Sonu Chawla (Co-Portfolio Manager, TimesSquare Capital Management)

An analytical review of TimesSquare’s Snowflake thesis — why the team read the April software selloff as a mispricing, and how they sourced conviction from a product the company itself left out of guidance.

My key takeaways:

• The mispricing setup.

During the April 2026 “SaaS selloff,” de-rated to ~7x forward revenue vs. Datadog’s ~14x — despite >30% revenue growth and mid-20s% FCF margins. A late-March CRO departure (Mike Gannon) and a tough G2 comp (prior-year Q2 had accelerated +600bps) compounded the angst. Per Chawla, the team ~doubled its position into that drawdown.

• The catalyst left out of guidance.

Snowflake’s data coding agent Cortex Code ("Coco”), positioned as “Claude Code for the Snowflake environment,” went GA in Feb 2026 and was excluded from fiscal-Q1 guidance (management couldn’t yet size a brand-new product). Chawla’s thesis: even a 1-2% consumption contribution would be meaningful upside. Per her account, the print delivered a ~15% guidance revision and an ~87% subsequent move.

• The adoption curve, read off LinkedIn + channel checks.

CEO Sridhar Ramaswamy disclosed (LinkedIn, early April) that within 9 weeks of launch, 50% of customers were using Coco, with early-adopter consumption +11% on average. Cited customer anecdotes: one with 90% of code AI generated via Coco; another claiming ~500 saved engineering hours (~$100K of value) in the first 20 days.

• Why Coco matters — accelerating timelines and query volume.

It compresses migration timelines from ~6 months to ~6 weeks (faster migrations) and turns every business user into a data analyst (more queries). Both feed Snowflake’s consumption-based model —> higher usage intensity and workload breadth (Snowflake serves 11,000+ customers, ~800 of the Global 2000).

• Sales-model pivot: headcount and topline are dissociating.

New CRO Jonathan Beaulier (10-yr Snowflake veteran, AI-fluent) is shifting from the prior headcount-driven, “feet-on-the-street” model to flat-headcount, AI-driven efficiency while targeting ~30% growth; management raised operating-margin guidance +100bps. Chawla’s framing: the old “sales headcount x productivity = growth” equation is breaking.

Follow our LinkedIn and Follow us on X to track our feed.

Credits & Disclosures:

Invest Like the Best: Hosted by Patrick O’Shaughnessy (Listen on YouTube/Spotify).

Odd Lots: Hosted by Joe Weisenthal & Tracy Alloway (Listen on Bloomberg/Spotify)

Pitch The PM: Hosted by Doug Garber (Listen on YouTube/Spotify).

Inside The Podcast (ITP) is an information distribution asset operating under the Pitch The PM umbrella. ITP is an educational podcast digest and does not constitute investment or financial advice.