Fidelity veteran says AI malinvestment is 17x dot-com bubble, SK Hynix chairman on the memory supercycle, JP Morgan strategist debunks the dedollarization, Jason Zweig remembers Benjamin Graham
NFL veteran is making sports investable for retail, financial historian on 300 years of financial advice that didn't work, Coinbase Vice Chairman on the "Everything exchange", Zero to One deep dive
July 10 Investing Scuttlebutt Digest
Table of contents
Roundup
A roundup of top investors’ and CEOs’ appearances this week — interviews, podcasts, events, and the topics they discussed.
Investor Appearances
Chris Camillo on Dumb Money Live
On the latest Dumb Money Live, the hosts asserted that Fable AI demonstrates human-level creative nuance, evidenced by a professional designer’s resignation; defended Bloom Energy against fraud claims by citing robust institutional contracts with Brookfield and Oracle; identified the Brazilian rubber flip-flop market as a high-potential retail play with Alpargatas as a primary beneficiary; and proposed that Gen Z women’s podcasts represent the current premier media growth opportunity. The hosts emphasize their “social arb” investment philosophy.
Mentioned: Alpargatas; Birkenstock; Crocs; VFC Corp; Deckers; Bloom Energy; The Gap / Old Navy; Mercado Libre; Amazon.
Steve Eisman on The Real Eisman Playbook
On this episode of The Weekly Wrap, Steve Eisman makes the case that diversification in today's market is largely an illusion and that the entire market has effectively become a single bet on AI. Steve covers the collapse of Circle and Nike's uninspiring quarter. He dismisses residential real estate as a viable investment and suggests that historical literacy—specifically the study of patterns in power and human irrationality—serves as a far more effective tool for navigating financial volatility than standard business literature. Steve closes with two mailbags: one on his personal investment portfolio and one on book recommendations.
Mentioned: Circle, Tether, Stripe, Visa, Mastercard, Coinbase, BlackRock, Autonomous Research, Nike, Apollo, S&P 500, Amazon, Google, Meta, Oracle, Nvidia, OpenAI, Microsoft, Anthropic, Samsung, Jones Trading, Queen Anne’s Gate, Unicus Research, Strategus.
Brian Dunne on The Markets by Goldman Sachs
Goldman Sachs FX options trading head Brian Dunne told The Markets podcast that dollar strength this year has been driven by three forces: the US-Iran conflict, an intact AI trade reinforcing "US exceptionalism," and a potential hawkish shift at the Fed. He flagged that Fed dot plots moved from zero voters favoring a hike in March to nine voters favoring at least one hike by June — a shift he called still underpriced by markets. Dunne's top trade idea: long dollar versus the Swiss franc, using call spreads he says offer seven-to-eight-times payout given historically low volatility.
Mentioned: US dollar, Japanese yen, the Brazilian real, the Egyptian pound, the Swiss franc, Chinese yuan, dollar-Swiss call spreads, and one-year dollar-CNH calls.
Marques Colston on Making the Sports Asset Class Accessible on Bloomberg Radio
Former New Orleans Saints wide receiver Marques Colston discusses the Champion Fund, an SEC-registered interval fund built to open the sports asset class to everyday investors rather than institutions only. The core problem the fund is designed to solve, in Colston’s words: “the athletes, the fans... the coaches, administrators, they create all the value but get none of the equity.” The fund’s thesis centers on what Colston calls “meteorites”—media rights and capital displaced from legacy broadcasters as streamers like Amazon and Netflix enter sports, with that displacement flowing downstream to boost emerging leagues such as the WNBA and international soccer.
Jason Zweig on Wealthtrack
Benjamin Graham learned through experience and observation that investors are their own worst enemy. Jason Zweig, editor of the 75th-anniversary edition of “The Intelligent Investor,” explains how Graham’s advice is even more important today.
George Noble on Bloomberg Radio
Fidelity veteran George Noble of Noble Capital Advisors puts a number on the AI bubble: citing economist Julian Garrett, he says current malinvestment in AI is 17 times the size of the dot-com bubble. He calls Tesla "the biggest misallocation of capital at scale in the history of stock markets, perhaps only surpassed by SpaceX" — a company he says priced its IPO at roughly 120 times revenue. He also flags SpaceX's staggered share unlock schedule (5% float today, climbing to 100% by December) as "manipulative," names Oklo as "one of the biggest frauds out there on the market right now," and lays out why he thinks energy and gold — not chips — are the best risk/reward trades on the table.
Mentioned: South Sea bubble, Dutch tulip bulbs, dot-com, SpaceX, Bitcoin, Tesla, Chrysler, Ford, GM, SSRM, Oklo
DoubleLine Capital
DoubleLine's Portfolio Manager Eric Dhall and Analyst Mark Kimbrough break down a second quarter where the technology sector ripped 43.49% while nothing else came close — the industrial sector was a distant second at 15%, and the S&P 500 equally-weighted index (stripping out mega-cap influence) gained just 11.4%. Their warning: this is "a very thin market really driven by the AI animal spirits," and the question now is how long it can run. They also flag a stark reversal already underway in July, a labor force participation drop that pulled 750,000 people out of the workforce in one month, gold down over 13% in Q2 while copper jumped 10%, and a preview of who's sitting on the Fed's newly announced task forces — including someone from Anthropic on the "productivity and jobs" panel.
Mentioned: S&P 500, Russell 1000 value, Russell 1000 growth, S&P 500 equally weighted index, Russell 2000, U.S. dollar, WTI crude, gold, copper, Bitcoin, Anthropic, Xbox, Walmart.
Michael Cembalest: Why He’s Still “Analytically a Patriot” — Debunking Dollar Doom, China Bulls, and AI Bubble Fears | The Compound and Friends (Ep. 250)
On episode 250 of The Compound and Friends, Michael Batnick and Josh Brown are joined by Michael Cembalest, the Chairman of Market and Investment Strategy at JP Morgan Asset Management and author of Eye on the Market.
Michael Cembalest unloads a data-driven takedown of nearly every popular bear case making the rounds — the dollar's collapse, China's rise as a reserve-currency alternative, and gold hoarding by central banks. His verdict: "none of the pillars of the dollar as the world reserve currency are shifting." He also flags a specific worry that's crept up on him only in the last three months: whether semiconductor stocks have quietly become "too big proportionally" relative to the hyperscalers actually buying their chips — a dynamic he compares directly to the telecom equipment rollover that preceded the dot-com crash. Plus: his real timeline for when U.S. federal interest and entitlement spending consumes 100% of tax revenue, why he ranks OpenAI as carrying more "business uncertainty" than Anthropic or SpaceX, and a preview of a forthcoming cybersecurity report — code-named "Patchmageddon" — written for the 20-30 companies inside Project Glasswing.
Mentioned: Nvidia, Google, Apple, Alibaba, Gemini, Anthropic, SpaceX, OpenAI, Cisco, Verizon, Micron, Meta, Oracle, Amazon, Salesforce, Blackstone, Samsung, and SK, U.S. dollar, gold, Chinese yuan, Japanese yen, British pound, Singapore dollar, Norwegian krone, U.S. Treasuries, leveraged ETFs
Louis-Vincent Gave: “China’s Building Dubai, the US Is Building Fortresses” — On Oil, AI, and the Coming Dollar Crack | RiskReversal Podcast
Dan Nathan brings together Peter Boockvar (Chief Investment Officer, OnePoint BFG Wealth Partners) and Louis-Vincent Gave (CEO, Gavekal) for a conversation that moves beyond the AI hype cycle and into the hard constraints—and crowded trades—shaping institutional capital allocation today. They unpack why cheap, open-source models threaten to commoditize OpenAI and Anthropic's trillion-dollar valuations, whether a $6.5 trillion AI CapEx buildout is sustainable, and why record semiconductor concentration in global indexes is creating a fragile, overcrowded position. The discussion then turns to longer-term structural shifts: the energy crunch driving refining shortages worldwide, the case for gold and commodity stockpiling in a post-Hormuz world, and which bond markets may crack first (spoiler: it may not be the U.S.). They close on where Boockvar and Gave are actually deploying capital right now: financials and cyclicals over crowded tech.
Mentioned: OpenAI, Anthropic, Goldman Sachs, JP Morgan, Valero, Marathon Petroleum, Micron, CXMT, Yangy Memory Technology, Alibaba, NVIDIA, Oracle, SpaceX, energy stocks, semiconductors, gold, long-term bonds, US Treasuries, financials, the AI arms race, the dollar’s structural decline, energy scarcity, the data center buildout, and the inflationary boom.
C-Suite Appearances
SK Group Chairman Chey Tae-won Discusses SK Hynix's ADR Debut and the Case for an "AI Era" Memory Supercycle on Bloomberg Radio
SK Group Chairman Chey Tae-won explains why SK Hynix chose now for its $26.5 billion US listing — after 15 years of waiting for access to American capital markets — and lays out why he believes the memory business has permanently broken from its old boom-bust cycle. His central claim: AI inference and "KV caching" have turned memory demand into something structurally different from the smartphone-driven cycles of the past, since a single user could soon run "tens or hundreds" of AI agents, each requiring its own memory pool. He also reveals SK Hynix is doubling total capacity within five years — and says customers are telling him that's still not enough — plus his direct answer on whether AI is a bubble, and where things stand with the Trump administration on U.S. investment.
Coinbase’s New Vice Chairman, Ryan VanGrack, on Crypto Legislation and the “Everything Exchange” on Fox Business
Newly appointed Coinbase Vice Chairman Ryan VanGrack — stepping into the role as longtime Chief Legal Officer Paul Grewal steps down — makes his case that crypto's roughly 50% market cap decline since late 2025 is masking positive underlying trends: institutional adoption, new partnerships, and what he calls unprecedented regulatory momentum behind the CLARITY Act, which he expects to reach the Senate floor by early August. But the conversation turns sharp when host Maria Bartiromo raises JPMorgan CEO Jamie Dimon's warning that the bill could destabilize the monetary system — VanGrack fires back that Dimon's stablecoin concerns are self-serving and that "no one is going to bow down" to him, while separately defending Coinbase CEO Brian Armstrong against Dimon's criticism as the industry's most persistent advocate for real regulation.
Pelgo CEO Chieh Huang Discusses AI, Jobs, and the New American Dream on Bloomberg Radio
CEO Chieh Huang argues the American entrepreneurial dream remains intact globally even as the U.S. no longer holds a monopoly on opportunity, pointing to companies raising $29 billion as evidence capital is flowing freely across borders. On AI’s labor impact, he stakes out a contrarian near-term position: “I’m in that camp” that believes AI will be “a net destroyer of jobs” in the short run, comparing the disruption to what happened to America’s steel belt—while noting the U.S. lacks the safety nets other countries offer displaced workers. His firm works with companies to provide outplacement and support for employees displaced by AI adoption.
Mentioned: Pelgo, Volkswagen, and electric vehicles.
Business and Investing Podcasts
This week in AI: Grok is Back, ChatGPT 5.6, Meta’s new models, China’s Chip Gambit on Limitless Podcast
The hosts discussed many new AI releases this week. Including xAI’s Grok 4.5 and OpenAI’s GPT 5.6, their coding, reasoning, and benchmark performance. OpenAI also introduced a new full-duplex voice mode, which lets users speak and listen at the same time, and may preview future AI interfaces; Chinese AI labs are building their own chips to reduce reliance on NVIDIA and imported hardware. And finally, a few updates on Meta’s new image and video models, ByteDance’s SeeDream image model, and Cloudflare’s payments API for micropayments and AI agent access.
Companies mentioned: SpaceX, xAI, OpenAI, Anthropic, Amazon, Cursor, Meta, ByteDance, Cloudflare, Nvidia, Huawei, DeepSeek, Jepu, Etched, Thinking Machine Lab
David Senra Breaks Down Zero to One: Peter Thiel’s Blueprint for Building a Monopoly on Founders Podcast
Founders podcast host David Senra revisits Peter Thiel's Zero to One for the first time in four years, calling it "probably the only business book, maybe the only business book worth reading." His core takeaway: forget competing, forget "minimum viable products," and forget copying what worked for someone else. Thiel's argument is that every great company is built around a secret — an unexploited truth about the world — and that the goal isn't to win a competitive market; it's to avoid one entirely by building what Thiel calls a "creative monopoly." Senra pulls apart how Steve Jobs, Jeff Bezos, and even Rockefeller applied this, why "competition is for losers," and why Thiel thinks the most dangerous and most powerful companies are led by founders with extreme, contradictory personality traits.
Mentioned: Apple, Yahoo, Facebook, Polaroid, Fairchild Semiconductor, Google, Amazon, Constellation Software, Standard Oil, Dell, and Nvidia.
Historian Dr. Joseph Moore on 300 Years of Financial Advice That Worked (and Didn’t) on The Meb Faber Show
Historian Joseph Moore spent over a decade in the archives to write a book on three centuries of American financial advice — and his central finding upends nearly everything people assume about money. Real estate "did not always go up," he says — in fact, an inflation-adjusted home in most American cities cost the same in the 1990s as it did in the 1890s. The claim that "stocks always beat bonds" over the long run is also false, he says, citing academic research showing bonds actually beat stocks for a roughly 60-70 year stretch of the 19th century. Moore also reveals his own real estate horror story (human traffickers, a shotgun-wielding tenant), the moment he made his own fake billion-dollar cryptocurrency to prove a point about net worth, and why he thinks Bitcoin's real legacy won't be what its founders intended.
Mentioned: Canal Bank of New Orleans, Bitcoin, US Treasury bills, residential real estate, stocks, bonds, gold, REITs, ETFs, cryptocurrency
Economist Jim Paulsen Discusses Why the AI-Driven Bull Market Looks "Like Dot-Com" to Him on Excess Returns Podcast
Jim Paulsen, PhD economist and 40 years as a Chief Investment Strategist, joins to explain why weakening economic momentum, tightening financial conditions, and extreme AI enthusiasm could set the stage for a 10% to 20% stock market correction. They discuss labor market weakness, the growing divide between technology and the broader economy, fading tech leadership, market complacency, bond yields, and the demographic forces that could keep US growth and inflation lower for years.
Jim also explains why he does not expect a recession or the end of the long-term bull market, but believes investors may need to reduce their concentration in AI and technology stocks as leadership quietly shifts toward the broader market.
Mentioned: S&P 500, Magnificent Seven, technology sector, communication services sector, information technology, information processing equipment, intellectual property products, utilities, consumer staples, healthcare, REITs, US Treasury yields, US dollar, WTI crude oil
Deep Dives
Detailed breakdowns of today’s must-listen investing and finance podcasts — key arguments, notable quotes, and takeaways from each episode, condensed so you can catch up without listening.
Investors Appearances
Chris Camillo on Dumb Money Live
The Flip-Flop “Social Arbitrage” Trade
A $750 flip-flop from Mary-Kate and Ashley Olsen’s label The Row went viral in the spring of 2025, worn by Jennifer Lawrence, Kendall Jenner, Zoë Kravitz, Jonathan Bailey, and Elizabeth Olsen. By summer 2026, the look had migrated downstream into a visually similar $20 mass-market flip-flop, creating a full price ladder from $20 to $750 that lets every income bracket participate. The hosts trace the trend to its purest public-market expression: Havaianas, whose square-toe silhouette most closely echoes the viral shoe, combined with the broadest global distribution of any comparable brand. Havaianas isn’t independently listed, though — it sits under its Brazilian parent, Alpargatas (ticker ALPA4), which trades only on Brazil’s B3 exchange and draws an estimated 98–99% of revenue from Havaianas. As one host framed it, “this is about as pure of a pure trade as it gets.” Cited Q1 2026 figures (stated on-air, unverified): sales +12%, volume +8%, U.S. sales +22%, U.S. volume +161%. The hosts note most of the actual trend impact likely landed in just the last 60 days, meaning that quarterly data probably understates the current run rate.
Why It’s Hard to Trade.
There is no actively traded U.S. ADR for Alpargatas — the hosts explicitly warn listeners against any ADR ticker that surfaces in a search, believing it to be inactive or illiquid. Schwab doesn’t offer direct Brazil market access; Interactive Brokers is presented as the only realistic U.S. retail path, and even that requires a funded account most retail investors won’t already have. Notably, neither host had money in an Interactive Brokers account at the time of recording — they aired the idea anyway rather than delay for a personal head start, and disclosed they held no position. Currency conversion adds further friction, at roughly 1% each way. The hosts used the episode to publicly needle Robinhood’s CEO over the access gap: “we are desperate for it as retail investors,” one said, arguing that tokenized international market access would be “a massive segment” if built.
Sympathy Trades — Mostly Rejected.
VF Corp and Deckers were ruled out; both divested their strongest sandal assets (Reef, Sanuk), and sandals now make up a minimal slice of Deckers’ revenue mix. Reef itself has strong distribution but is privately held, taking it off the table. Crocs surfaced in related search queries but wasn’t dominant enough to matter. Old Navy and Gap were dismissed as low-margin traffic drivers rather than genuine trend beneficiaries. MercadoLibre came up as a tangential Brazil/Latin America e-commerce long, disclosed separately by one host — explicitly not part of the sandal thesis.
Birkenstock — the one real sympathy play.
One host disclosed an actual long position, built on: the Arizona sandal’s square-toe styling sitting loosely adjacent to the trend; Google search interest for “Birkenstock sandal” roughly doubling year-over-year this June, versus a slower, steady climb in prior years; SimilarWeb traffic up meaningfully year-over-year; and an upcoming July 16 collaboration with French ballet-shoe brand Repetto across the Arizona, Ribbon, Scallop, and Opera lines, expected to lift brand attention even though the limited run won’t move revenue on its own. Tariff-driven margin pressure is treated as already priced in, with pricing power — price increases not denting demand — as an offset. But the host was careful to size the conviction correctly: “I wouldn’t call it a high conviction trade on Birkenstock by any means. This would be a medium conviction trade for me.”
The Conviction Framework.
The hosts grade trades across three tiers of evidence quality. In their words: “what we have high confidence in... is that fashion attention and search momentum are... exceptionally strong.” Retail sell-through data earns “moderate conviction.” Extrapolating the trend to the entire sandal category, rather than the specific fashion sub-trend, is where they have “less conviction” — and they explicitly warn against making that leap. This week’s flip-flop trade sits at the high-confidence tier bordering moderate; the broader “all sandals are hot” read is the low-conviction tier, and they say so directly.
Why the Trend Might Have Legs.
Six reinforcing factors were cited for durability beyond a single season: luxury legitimization cascading into other designer houses; a broader 90s/Y2K materials revival; the wide price ladder letting every budget participate; a broad and growing celebrity endorsement pipeline; expanding use cases beyond beach and pool into going-out styling; and high product-iteration bandwidth — colors, materials, toe shape, embellishment — giving brands room to keep the look fresh across multiple seasons rather than burning out in one.
Bloom Energy: the short thesis, and why the hosts aren’t buying it.
A short seller from Hunter Brook Research alleged on CNBC that Bloom Energy lacks sufficient scandium supply for its fuel-cell technology, drawing a Theranos comparison; the stock fell roughly 20% on the report. Bloom’s response, per an SEC disclosure cited on the show, asserts sufficient non-China scandium supply for current demand and backlog, with visibility into 25 GW of annual production. The hosts’ core pushback: major counterparties — Brookfield (partnership expanded to $25B), Oracle (contract scaled to 2.8 GW), Nebius (up to $2.6B committed) — would all have had to skip their own diligence for the short thesis to hold. “It’s theoretically possible... but it’s extraordinarily unlikely,” one host said. They also point out that the manufacturing yield and efficiency data needed to actually verify the scandium claim is proprietary, meaning outsiders on either side of the trade are working with incomplete information. One host disclosed he isn’t selling and may add to the position, stating plainly: “I’m willing to take that risk.”
Beyond public markets: a women-centered podcast bet.
Late in the episode, one host used the flip-flop trade as a springboard into a broader point about "social arbitrage" thinking — that spotting a cultural opportunity doesn't always end in a stock trade. He disclosed he's developing a new podcast built around what he called "the number one talent solo content creator" in the Gen Z women's space, framing it as a multi-year bet rather than a side project. His reasoning: AI is going to increase, not decrease, demand for distinctly human personalities, since raw authenticity is "the one thing that AI can't yet fully deliver." He argued the bigger opportunity specifically sits in women's podcasting because of a structural gap — "do you know 70% of podcasts are men's podcasts?" — driven in his view by higher risk-aversion among women creators facing the technical friction of cameras, editing, and production. He drew a direct parallel to his own history of acting on a cultural read outside public equities, referencing an earlier venture (a Pokémon-themed trade show) built from a thesis with no investable public equivalent at the time. The takeaway he offered listeners: training yourself to think like a "social arb investor" has implications well beyond the stock market — sometimes the trade is a business, not a ticker.
Steve Eisman on The Real Eisman Playbook
It’s All One Trade: AI’s Grip on the Market
The GDP math. Citing Apollo chief economist Torsten Slok, Eisman broke down the roughly 2% US GDP growth expected in 2026: about 100 basis points comes from AI-related capex and data center spending, 30bps from reshoring/infrastructure, and roughly 90bps from tax refunds tied to Trump’s “big beautiful bill.” In other words, AI accounts for about half of this year’s growth.
AI accounts for about half of this year’s growth.
The market math is worse. Eisman argues AI’s market impact dwarfs its economic impact. Nvidia alone is 38% of the S&P 500; add Google, Amazon, and other tech-adjacent names, and you clear 50%. That breaks the assumption that index ownership equals diversification.
Bonds aren’t a hedge either. By Eisman’s numbers, 15% of all existing corporate debt is AI-related, and 50% of all newly originated corporate debt in 2026 is AI-related — meaning a standard 60/40 stock/bond split doesn’t provide the diversification investors assume it does. “It’s all one trade. It’s all AI,” he said, adding that if AI fails to deliver or takes much longer than expected to become profitable, he expects the US to tip into recession with markets falling sharply.
How the Bull/Bear Debate Has Evolved
The GFC parallel. Eisman drew a direct line between how his subprime short thesis developed over 12+ months in 2006–07 — starting from anecdotal evidence, then delinquency data, then a tip about Wall Street’s own balance sheet exposure — and how the AI debate is now maturing in real time as new facts surface.
Cracks in the bull case. He flagged growing signs of an AI price war: cheaper Chinese AI models are drawing corporate interest, and AI companies had been subsidizing token pricing well below actual cost to drive adoption. That subsidy is fading — corporate AI budgets reportedly blew through annual plans within months, and companies are now pulling back the “all you can eat” AI access they gave employees.
Where money is rotating. Investors have been shifting out of hyperscalers and into semiconductors, semiconductor equipment, and AI power plays — though even that trade is showing strain. Eisman pointed to Samsung’s 1,800% operating profit jump this week, which still sent the stock down 7% on fears that slowing hyperscaler growth will eventually pressure semi pricing.
This Week’s News
Circle / stablecoins. Circle fell 17.5% on June 30 after a consortium including Visa, Mastercard, Stripe, Coinbase, and BlackRock announced a competing stablecoin ecosystem — a serious threat given Circle’s position as the second-largest stablecoin issuer behind Tether.
Nike. Adjusted EPS came in at 20 cents versus 14 cents a year ago and 13 cents expected, with revenue of $10.97B also beating estimates — but revenue was still down 1% year-over-year, and cautious management commentary sent shares lower after hours. Eisman’s take: a beat, but not a turnaround story.
Iran. Attacks on shipping in the Strait of Hormuz and on US sites in Bahrain and Kuwait drew US retaliatory strikes, with Eisman describing the ceasefire as increasingly fragile.
Mailbag
On portfolio construction: A subscriber asked how Eisman’s disclosed 16-stock portfolio fits into his broader asset allocation. His answer: it is his entire long book. Beyond that, he holds mostly cash in a money market fund, no bonds, no precious metals, and doesn’t count his home as an investment. His hedges are held in short positions he hasn’t yet disclosed.
On required reading: Asked for a recommended reading list, Eisman pushed back on the premise — he rarely reads business books, favoring history and historical analysis instead. Picks mentioned: The Nation That Never Was (Kermit Roosevelt), Intellectuals (Paul Johnson), The Guns of August (Barbara Tuchman), The Cultural Revolution (Frank Dikötter), and Colleen McCullough’s First Man in Rome. His reasoning: broad, cross-domain reading builds the analogies he draws on when identifying patterns in markets — “being well read is my superpower.”
Brian Dunne on The Markets by Goldman Sachs
On the Fed and rates. Dunne said real (inflation-adjusted) rate differentials still favor dollar strength even in a scenario where the Fed simply holds rather than hikes. He noted six to seven basis points are priced for July and slightly more than a full hike by year-end, but stressed the committee’s shift in its dot plot is the bigger signal than near-term pricing.
On dollar debasement fears. He separated two worries investors have raised: 2025’s tariff policy (reserve diversification is continuing but hasn’t accelerated beyond expectations) and petrodollar risk tied to Iran. On the latter, he flagged a real tail risk — Iran could start charging yuan rather than dollars for transit through the Strait of Hormuz — but said there’s no evidence yet that this is actually shifting reserve currency dynamics.
On Japan and the yen. Despite Japanese yields hitting 30-to-40-year highs, Dunne argued policy is still too easy relative to inflation to support the currency, and that recent long-end moves reflect inflation expectations more than tightening. His house view: continued yen weakness with episodic intervention — unless Japan’s pension system pivots away from dollar assets toward domestic holdings, which he called the biggest swing factor.
On positioning and trades. Dunne warned the market is heavily crowded into carry trades this summer, with EM carry (Brazil, Egypt) in favor. His own highest-conviction ideas: long dollar vs. Swiss franc as a core position, and long dollar vs. Chinese yuan (one-year dollar-CNH calls struck above 7) as a hedge against renewed Fed hawkishness or a re-escalation in the Middle East pushing energy prices higher.
Marques Colston on Making the Sports Asset Class Accessible on Bloomberg Radio
The fund structure and mission
The Champion Fund is structured as an SEC-registered interval fund, a format that allows it to pool capital from a broader investor base than the institutional-only vehicles that have historically dominated sports investing. Colston frames his own background—as a player, then an owner with operational experience over roughly seven years—as giving him a multi-lens view of the asset class: “I’ve been able to see this asset class from a bunch of different lenses.” That perspective shaped the fund’s mission, which he describes plainly as making “the sports asset class accessible to every investor.”
The five sub-asset classes
The fund breaks its exposure into five buckets: emerging sports teams (with specific interest in leagues like Serie A, which Colston calls “slightly underrated from a media value perspective”), growth-stage sports-tech ventures (fan engagement platforms, ticketing technology), media and services businesses, real estate and hospitality tied to sports (stadium development, sports-anchored mixed-use projects), and fund-of-funds positions in established managers with their own track records and thesis.
The “meteorites” thesis
Colston’s central growth argument is that media rights holders displaced by streaming entrants act as “meteorites” that push capital and attention downstream into leagues and properties previously overlooked by linear broadcast: “as meteorites kind of displace and find their way down downstream, we see... a high tide effect.” He points to the WNBA’s rapid media-rights renegotiation over the past two to three years as a live example, calling the league’s growth “just extraordinary.”
Where the fund stands today
The fund is currently working to secure anchor institutional investors. Asked what’s next, Colston kept it simple: “It’s really just continuing to tell the story, continuing to find those anchor investor partnerships, and just continuing to get the word out.” He described the reception from prospective partners and portfolio companies as strong, noting portfolio companies value access to “a distribution network of investors that typically they wouldn’t have access to.” No specific fundraising targets, minimums, or closing timeline were disclosed in this conversation.
Jason Zweig on Wealthtrack
The Investor as Their Own Worst Enemy: Zweig explains that Graham’s most profound observation is that investors often fail due to their own psychological impulses rather than market forces. Modern distractions like social media, trading apps, and the constant urge to compare oneself to others exacerbate these self-defeating behaviors.
Definition of the Intelligent Investor: Intelligence, in Graham’s view, is not about high IQ or academic credentials, but rather good judgment, skepticism, independence of thought, and emotional discipline. It is more akin to wisdom than raw intellect.
The Need for Structure: To protect against impulsive decision-making, Zweig advocates for the use of an Investment Policy Statement and strict checklists. These structures force investors to pause, evaluate their goals, and analyze investments objectively before acting.
Investing in the Digital Age: Many modern trading platforms are designed to mimic the addictive nature of casino slot machines. Zweig suggests avoiding trading during market hours, ignoring market noise, and focusing on long-term objectives to maintain clarity.
Total Wealth Perspective: Zweig encourages investors to think beyond just their financial capital. He highlights the importance of viewing wealth through three primary buckets: financial capital (investments), physical capital (possessions), and human capital (career and earning potential), and ensuring all three are managed and integrated effectively.
George Noble on Bloomberg Radio
On the AI bubble — and why he thinks it’s worse than 2000
George Noble frames the AI trade through a historical lens: “there’s really nothing new under the sun,” pointing to the South Sea Bubble and Dutch tulip mania as precedents for the same fear-and-greed cycle now playing out in AI. He’s careful to say he isn’t “bearish on AI per se” and expects near-universal adoption, but his test is simple: “show me the money. Where’s the ROI? And I don’t see it.” He cites Julian Garrett of Micro Macro Strategy Partners, who calculates the current malinvestment at 17 times the dot-com bubble, and argues the scale relative to the real economy — not just the dollar figures — is what makes this cycle more dangerous.
On SpaceX’s valuation and share unlocks
Noble is sharply critical of the SpaceX IPO, noting that buying companies above ten times revenue “usually ends very badly” — and SpaceX priced at roughly 120 times revenue, with some projections showing the company cash-flow negative for years. His bigger concern is structural: a staggered lock-up schedule starting the month after the next quarterly earnings, with 20% of shares unlocking shortly after and additional tranches of 7% unlocking every 20-30 days, taking the float from roughly 5% to 100% by December. He calls this dynamic “manipulative” and argues regulators should intervene: “it’s not a question of what’s legal or illegal. It’s just not right,” warning that retail 401(k) holders are effectively providing exit liquidity for insiders.
On Tesla and Elon Musk
Noble calls Tesla “probably the biggest misallocation of capital at scale in the history of stock markets, perhaps only surpassed by SpaceX.” Drawing on his own background as an auto analyst under Peter Lynch — he visited Chrysler, Ford, and GM in Detroit in 1981 — he argues fundamental analysis simply hasn’t applied to Tesla’s valuation, and that professional auto analysts with hold/sell ratings “have been wrong” against a market narrative of “juy buy it, it’s Elon.” He also floats — with the caveat “I never speculate” — that a SpaceX-Tesla merger is “a reasonable speculation.”
Where he sees value: energy, gold, and commodities
Away from AI-adjacent names, Noble is constructive on traditional value sectors. He flags energy as offering “very little downside and potentially a lot of upside” given a widening gap between financial and physical oil markets, and calls the setup “the biggest energy dislocation in history.” He’s bullish on gold stocks, citing one name (SRM) trading at seven times earnings, not cash flow, and likes select copper and commodity names. On AI-adjacent power plays tied to data-center demand, he’s more cautious — “a lot of the derivative power plays... are going to have a big problem” — and says his fund has been short Oklo, which he calls “one of the biggest frauds out there on the market right now,” for a year.
On investing internationally
Asked whether overseas investing requires a weak dollar, Noble says no — the key variable is that “economic cycles across different geographies vary enormously,” requiring a top-down view layered on bottom-up stock picking. He notes markets are far more correlated today than when he started at Fidelity in the early 1980s, reducing some of the diversification benefit international investing once offered.
DoubleLine Capital
Equities: an extremely narrow rally
The technology sector’s 43.49% second-quarter return drove the S&P 500’s 15.2% quarterly gain, with the industrial sector as the only other standout at roughly 15%. Russell 1000 Value rose 14%, Russell 1000 Growth rose about 17%, and the equal-weighted S&P 500 gained 11.4% — a meaningful gap versus the cap-weighted index that signals concentration risk. Energy and communication services were the weak spots, with energy still up over 23% year-to-date despite pulling back, and utilities essentially flat despite AI-driven power demand narratives. Month-to-date in July, the pattern is reversing: tech is down about 2.5%, while consumer discretionary and communication services lead. The Russell 2000, which gained 21.5% in Q2, is down 1.6% month-to-date.
Fixed income: a flattening curve
The 10-year yield rose 15 basis points in Q2 while the long end moved just 4 basis points, but the 2-year and 5-year sold off 38 and 28 basis points respectively — a belly-of-the-curve selloff tied to shifting inflation and Fed expectations. The 2s10s spread compressed from 52 to 29 basis points. Despite the yield moves, the Bloomberg Agg still returned over 60 basis points in Q2, led by emerging market debt (up over 4.5%), high yield corporates (up ~2.5%), and investment grade corporates (up ~1.5%). Month-to-date, the Agg is down about half a percent on duration-driven losses, with investment grade corporates the worst performer (-82 bps) and bank loans the best (+81 bps) given minimal rate sensitivity.
Commodities and currencies
The Bloomberg Commodity Index fell about 8% in Q2 (still up 18% year-to-date) as energy dropped over 13% amid an easing Middle East conflict, with WTI crude pulling back to roughly $71/barrel from over $100 during the height of tensions. Precious metals sold off sharply — gold fell over 13% in Q2 and sits down more than 5% year-to-date — while industrial metals gained, with copper up over 10% in the quarter and 9.5% year-to-date. The dollar index rallied to roughly 100.8 on the new Fed chair’s inflation-fighting commitment, and Bitcoin had a rough quarter, closing Q2 near $58,600 (down 27%) before bouncing to around $63–64,000, still down 14% year-to-date.
Labor market: stable, but with a wrinkle
Non-farm payrolls rose just 57,000 — a four-month low — while the household survey unemployment rate (U3) surprised lower at 4.2% versus an expected 4.3%. The explanation: labor force participation dropped from an expected-flat 61.8% to 61.5%, meaning roughly 750,000 people left the workforce, about a third of whom were previously counted as unemployed — mechanically pulling the unemployment rate down rather than reflecting underlying strength. JOLTS job openings for May came in at 7.66 million, pushing the openings-to-unemployed ratio to 1.04, the highest since January 2025. ISM Services for June printed 54.0, with new orders, backlog, and employment all showing expansion.
Fed watch and the week ahead
Rate traders are pricing in roughly 1.5 cuts through year-end, reflecting a hawkish tilt into next week’s CPI print. The Fed also named leaders for five new task forces — on communications, balance sheet policy, data, productivity/jobs, and inflation — with notable names including Nobel laureate Thomas Sargent on the inflation framework committee, former Bank of England governor Mervyn King on communications, and members with ties to Anthropic and Walmart’s former leadership on the productivity/jobs and data committees, respectively. Next week brings June CPI (est. -0.1% m/m, core steady at 2%), PPI (est. -0.1% m/m), and retail sales (est. +0.3% m/m).
Michael Cembalest: Why He’s Still “Analytically a Patriot” — Debunking Dollar Doom, China Bulls, and AI Bubble Fears | The Compound and Friends (Ep. 250)
Debunking the dollar bears
Cembalest argues that dollar-collapse warnings routinely fail a basic empirical test: are there actual signs of it happening? Tracking the dollar’s share of world trade, FX reserve allocations, SWIFT payments, and cross-border loans, he finds “there’s really no change” over the past five years — no currency is gaining meaningful share, not the yen, pound, or yuan. He’s equally skeptical of gold-hoarding narratives, explaining that the surge in gold’s share of central bank reserves is almost entirely a price effect — gold rising from roughly $1,500 to over $4,000 an ounce — rather than central banks actually accumulating more ounces: “that is 95% of the reason why gold as a share of foreign exchange reserves has gone up.”
Why China can’t replace the dollar
Cembalest’s clearest argument against a Chinese alternative centers on capital controls: China’s money supply has “massively outstripped” its accumulation of central bank assets and foreign exchange reserves, meaning that if China ever loosened capital controls, pent-up money would immediately flee the country. He notes China reportedly saw $800 billion to $1 trillion in capital outflows in 2025 despite existing controls, and argues a true reserve currency can’t carry that kind of structural risk.
The real debt timeline
Cembalest’s biggest long-term concern isn’t a sudden crisis but a specific calendar date: he projects that by 2031 or 2032, “100% of federal tax revenue” will be needed just to cover interest on the debt and entitlement spending, with credit rating agencies likely to threaten a downgrade two to three years before that point unless entitlement reform occurs. He contrasts this with the Eisenhower era, when the U.S. reduced its debt-to-GDP ratio from roughly 100% to 60% not through tax hikes or spending cuts, but through strong nominal growth.
On AI, chips, and a potential 1999 echo
Cembalest describes the current market catalyst as more technically complex than any he’s covered in 30 years, but his most pointed concern is structural: hyperscaler capital spending has begun shifting from internally generated cash flow toward debt and equity financing, with free cash flow margins declining sharply. He draws a direct historical parallel to 1999, when communications equipment makers kept climbing even as the internet service providers buying their gear began rolling over — a warning sign, in his view, now potentially playing out with semiconductor stocks continuing to run even as some hyperscaler names show signs of strain. He notes hyperscalers’ own in-house chips (Google’s TPUs, AWS’s Trainium, Microsoft’s Maia, Meta’s custom silicon) are reportedly cutting total cost of ownership by 30-40% versus Nvidia GPUs — part of why Nvidia’s forward P/E has compressed to around 18-20 times.
Leveraged ETFs and Korean retail as late-cycle signals
Cembalest flags the explosive growth in double-leveraged ETF notional exposure — roughly $500 billion — as a source of the sharp intraday market swings investors have been experiencing, since providers must mechanically buy into rallies and sell into selloffs to stay hedged. He also singles out surging margin lending on Korea’s KOSPI index tied to Samsung and SK Hynix as a signal worth watching, recalling similar retail euphoria patterns from 1997 and 2008: “Korean retail is basically the... scariest market indicator... by the time those guys start to pile into stuff, you’re really, really late in the game.”
Ranking OpenAI, Anthropic, and SpaceX on business uncertainty
Asked to rank the business uncertainty of the major AI companies, Cembalest places OpenAI highest, followed by Anthropic, then SpaceX — citing OpenAI’s revenue concentration in consumer subscriptions with limited enterprise and advertising diversification. He draws a comparison to AOL in the early internet era: “the network effect is pretty powerful, but... Anthropic showed how quickly those kind of things can change.”
Market health check
Despite pockets of frothy behavior, Cembalest describes current market internals as broadly healthy — citing strength in regional banks, transports, industrials, small caps, and Wall Street banks, along with better-than-expected retail sales and solid bank loan growth. His preferred coincident indicator, the ratio of new manufacturing orders to inventories across the PMI data JPMorgan tracks, currently reads favorably, though he flags rising credit card and auto loan delinquencies as a softer spot in the consumer.
A cybersecurity warning ahead
Cembalest previewed an upcoming JPMorgan research piece — internally titled “Patchmageddon” — written in partnership with the firm’s cybersecurity team as part of Project Glasswing, the group of roughly 20-30 companies with early access to advanced AI models for security research. He described AI-discovered vulnerabilities as often defying human intuition entirely, warning that the pace of newly discovered vulnerabilities is currently outstripping the pace of patching, and that individual companies should expect a wave of AI-enabled breaches in the coming period even if it doesn’t become a market-wide event.
Louis-Vincent Gave: “China’s Building Dubai, the US Is Building Fortresses” — On Oil, AI, and the Coming Dollar Crack | RiskReversal Podcast
Fortress vs. Dubai: two AI strategies
Gave’s central framing: “OpenAI and Anthropic are essentially looking to build medieval fortresses with huge walls and say come into our fortress and you’ll be safe” — serving enterprise clients like Goldman Sachs and JPMorgan. China, by contrast, is “building Dubai... everybody can come in, everybody can come out.” He argues China was forced into the open-source strategy rather than choosing it: unable to buy the chips needed to compete on raw compute after U.S. export controls, Chinese labs had no option but to open-source their models and let global developers improve them for free — a strategy Gave calls “completely anathema to the usual Chinese way of doing” business, which normally just throws money at problems.
Oil, China’s storage strategy, and the $65-$100 range
Gave argues China’s decision to fully withdraw from the seaborne oil market during the Iran conflict — leveraging an estimated 1.8 billion barrels of storage capacity, larger than officially reported — was the single biggest reason oil didn’t spike toward $150 as many predicted. He frames China’s oil-buying behavior as mechanical: “at 65, they buy as much as they can... at 100 they stop buying,” effectively creating a de facto price ceiling and floor. He views the resulting $65-$100 range as bullish for most energy producers’ cash flow, but flags refining margins (crack spreads) as unusually stretched due to Russia-Ukraine refinery strikes, the loss of six major Gulf refineries including Bahrain’s during the Iran conflict, and China restricting refined product exports — with Russia’s diesel export halt (roughly 11% of world supply) adding further pressure.
Why China isn’t trying to destabilize the world order
Pushing back on the idea that China, Russia, and Iran form a coordinated “axis” benefiting from U.S. entanglement in the Middle East, Gave argues China’s overriding priority is domestic social stability, not geopolitical maneuvering — and that a $150 oil price hurting the U.S. consumer would actually hurt China more, since it depends on that consumer. He frames China’s shipbuilding buildup as a response to a specific lesson from the Iran conflict: “the US no longer controls the world’s sea lanes,” because modern drone warfare has broken the economics of naval protection.
Commoditization risk in AI and semiconductors
Gave’s most direct market call: both AI models and semiconductors are “priced to essentially never be commoditized,” despite China’s rapid entry into both markets. He points to Chinese memory makers CXMT and Yangtze Memory Technology as underappreciated threats to Micron’s roughly 85% gross margins, and notes token pricing has already shifted from “token maxing to token minimizing” in just the past few months — evidence, in his view, that customers won’t tolerate a bait-and-switch from cheap to expensive AI access.
The capex math and “no rainbow in year four”
Gave challenges the assumption that hyperscaler AI spending is a temporary three-year buildout followed by harvest years, citing Oracle spending effectively 100% of revenue on capex and pointing to embedded long-term costs — 30-year real estate leases, constant GPU upgrade cycles, and rising maintenance capex — that don’t disappear once the initial buildout ends. He cites a McKinsey estimate of $6.5 trillion in cumulative AI infrastructure spending between now and 2030.
Earnings quality concerns
Gave flags a widening gap between S&P 500 reported earnings and national income earnings (a proxy drawn from tax receipts), calling large divergences historically associated with either bubble-era financial engineering or bear-market bottoms. He notes recent quarterly S&P earnings growth of roughly 28% falls to about 17% excluding “other income” (investment gains from stakes in companies like OpenAI and Anthropic), and to mid-single digits excluding semiconductors entirely.
The bond market and a “Damoclean sword” over global yields
Gave describes a structural bond bear market driven by persistent inflation and unsustainable fiscal deficits (the U.S. running roughly 7% of GDP deficits at full employment). He highlights Japanese investors holding an estimated $3.5 trillion in foreign assets, equal to roughly 80% of Japan’s GDP, and argues that as the U.S.-Japan yield differential narrows, Japanese capital could increasingly repatriate — pressuring bond markets in the U.S. and Europe alike. He argues debt-to-GDP is the wrong crisis indicator; the real risk factor is what share of a country’s debt is foreign-owned, citing Greece, Thailand, and Argentina as precedents, and views France and the UK (both around 30% foreign-owned debt) as more exposed than the U.S.
The dollar’s slow unraveling
Gave argues Magnificent Seven stocks have effectively become a global reserve asset, cited by institutions from the Swiss National Bank to sovereign wealth funds, and that if the AI/hyperscaler trade rolls over, “the dollar resumes its downward trend... absolutely.” He also flags China’s currency policy shift — the renminbi has appreciated roughly 50 basis points per month for 18 months — as creating tension with a weakening yen, a divergence he expects to eventually “snap.” He argues the DXY dollar index, roughly two-thirds euro-weighted, is a poor gauge of the dollar’s real trajectory since it excludes China and commodity currencies entirely.
On gold and the shift toward stockpiling commodities
Gave remains “a bull on gold” but says he’s “much more bullish on other commodities,” arguing the lesson of the Iran/Strait of Hormuz crisis wasn’t to hold gold or Treasuries but to stockpile physical necessities — citing India’s fertilizer shortage when China declined to sell reserves. He connects this to reported U.S. Department of Defense lithium stockpiling and a new German natural gas strategic reserve, framing a coming multi-year wave of inventory rebuilding as structurally inflationary for commodities broadly.
Positioning: overweight financials and cyclicals
Gave’s tactical view for U.S. equities favors financials over tech, citing stretched valuations and extreme investor positioning in tech versus more attractive valuations and building momentum in financials, alongside a global trend of bank stocks (U.S., European, Japanese, even Chinese) breaking out. He expects a steeper yield curve as central banks keep short rates low to manage rolling government debt, favoring cyclicals and a weaker-dollar trade over consumer defensives.
C-Suite Executives Appearances
SK Group Chairman Chey Tae-won Discusses SK Hynix’s ADR Debut and the Case for an “AI Era” Memory Supercycle on Bloomberg Radio
Why list in the US now
Chey says SK Hynix had wanted access to U.S. capital markets since he acquired the company roughly 15 years ago, and the $26.5 billion raise represents that ambition finally realized. He confirmed leadership has discussed doing follow-on ADR offerings with regularity, though he declined to give specifics, and said the priority now is keeping the stock price stable to support future demand before pursuing additional rounds.
The structural shift in memory demand
Chey’s core argument is that AI has changed memory from a cyclical, hardware-count-driven business into something closer to permanent structural demand. In the old model, chip demand tracked the number of devices per person — one phone, one PC. In the AI era, he argues each person may eventually run many different AI agents, and each agent generates substantial KV (key-value) caching that requires memory to store, meaning demand scales with usage rather than device count. He said he doesn’t know how long this dynamic will persist, but suggested it continues until society reaches some kind of “settlement” with AGI.
Capacity expansion and customer demand
SK Hynix announced five months ago it will double its total manufacturing capacity within five years — a plan Chey says outside observers view as risking oversupply, but which he says customers have told him is still insufficient; some are asking for five to six times more capacity rather than a simple doubling. He also noted the industry has shifted toward long-term supply agreements at customers’ request, which he says helps smooth out the historically cyclical, capital-intensive nature of memory manufacturing.
Betting on memory when others wouldn’t
Reflecting on the 2012 decision to acquire what was then a distressed company under a bank-led workout program, Chey said the memory business’s brutal cyclicality and constant multibillion-dollar capital expenditure requirements had scared off other buyers. His view was that mastering memory manufacturing — with hundreds of billions of transistors packed into a single chip — represented the most sophisticated manufacturing process achievable, and that mastering it would open doors elsewhere.
On the AI bubble question
Asked whether his view has changed since previously saying the “cost of AI” was too high, Chey said token cost has fallen roughly 80–90% over the past three years and expects continued declines as the ecosystem matures. On bubble concerns specifically, he distinguished between stock market overshooting — which he acknowledged happens — and the underlying technology, which he called real and still in an early, immature stage that will keep improving.
Investment plans, talent, and the US relationship
SK Group has already invested more than $35 billion in the U.S. across SK Hynix, its bio and battery businesses, and SK Telecom’s AI startup investments, including a new Indiana fab. Chey said his broader investment plans exceed that figure significantly, though he didn’t give a specific target. On competing for skilled semiconductor talent against companies like Micron, he pointed to the ADR listing itself, along with stock options, as tools to attract talent globally. He also addressed AI infrastructure plans, targeting roughly 15 gigawatts of AI data center capacity in Korea and an additional 5 gigawatts internationally, including potential U.S. sites, with power availability and electricity pricing as the key constraints.
On China and competition with Samsung
Chey described China’s AI strategy as diverging from the U.S. approach, built around a large domestic pool of STEM graduates and a focus on producing cheaper tokens rather than competing on token quality. He declined to directly address whether SK Hynix can catch Samsung, saying his priority is stakeholder outcomes broadly — balancing shareholder returns, customer pricing and supply needs, and employee wellbeing — rather than positioning against a specific competitor.
Coinbase's New Vice Chairman, Ryan VanGrack, on Crypto Legislation and the "Everything Exchange" on Fox Business
On the CLARITY Act and the regulatory momentum thesis
VanGrack frames the CLARITY Act as pivotal for the industry, describing it as a bipartisan effort with both Democratic and Republican senators working to get it across the finish line. He argues the “20th century rails” of banking-era regulation weren’t designed for today’s financial technology and that the goal isn’t less regulation but regulation that actually exists for crypto in the first place — something he says has been missing for too long. He expects passage to trigger not just market stability but a broader suite of follow-on rules and regulations.
The Dimon clash
Bartiromo pressed VanGrack on Jamie Dimon’s concern that the CLARITY Act could let crypto platforms effectively pay interest on stablecoin deposits without the protections banks provide, potentially destabilizing the monetary system. VanGrack rejected the concern directly, arguing the legislation offers no advantage to money market funds and that traditional and community banks won’t be harmed by it. He characterized Dimon’s opposition as an incumbent spending heavily to protect the status quo, and said the industry will fight back regardless of whether it ultimately prevails. On Dimon’s separate criticism of Brian Armstrong personally, VanGrack pushed back firmly, crediting Armstrong and Coinbase with fighting longer and harder than any other company for crypto regulatory clarity — including suing the FDIC to push for clearer rules — and drawing on his own background as a former senior SEC regulatory official to back that claim.
On equal regulation vs. bank-style oversight
Bartiromo pushed a specific critique: if Coinbase’s products “sound like a bank” and “walk like a bank,” shouldn’t they be regulated like one, including FDIC coverage? VanGrack’s response was that applying nearly century-old banking rules unchanged to modern financial technology is a fiction, and that the right path is purpose-built regulatory scaffolding through the SEC and CFTC rather than forcing crypto into legacy bank regulatory categories.
Business priorities: the “everything exchange”
VanGrack pointed to Coinbase’s diversification beyond crypto trading — stocks, commodity futures, options, prediction markets, perpetuals, and AI-based advisory tools — as central to a broader thesis that traditional finance and crypto converge into a single “modern financial institution” over the next decade. He cited recent UK regulatory approval to offer derivatives and equities trading to both institutional and retail investors on one platform as an example of where Coinbase is expanding, alongside continued interest in prediction markets as CFTC guidance in that space develops.
On the stock’s decline
Asked about Coinbase shares being down roughly 30% year-to-date alongside weaker crypto trading volumes, VanGrack attributed the move to short-term uncertainty around the CLARITY Act’s progress and bitcoin price action rather than any deterioration in the underlying business, reiterating that he’s “never been more bullish” on Coinbase’s and the industry’s prospects. Bitcoin was trading near $64,000 at the time of the interview, with some traders eyeing a breakout toward $67,000.
Pelgo CEO Chieh Huang Discusses AI, Jobs, and the New American Dream on Bloomberg Radio
On the American dream and global opportunity
Chieh frames American entrepreneurship as “still a beacon for entrepreneurs all over the world,” citing a recent $29 billion raise as proof the dream is “alive and well.” But he’s careful to note this opportunity is no longer geographically contained: “that diaspora of opportunity is all over the world,” driven in part by how much AI “has democratized access to these markets.” His takeaway for investors is not to overlook opportunity abroad simply due to unfamiliarity with a given country.
On AI and jobs — the short-term/long-term split
Chieh takes a firmly near-term pessimistic stance: “I do think many, many jobs, many white-collar jobs are going to go away.” He pushes back on the framing that he’s being too short-term in his thinking: “longer term is tomorrow for me.” His reasoning centers on velocity rather than current capability: “oftentimes we look at what’s the state of the art, but we forget the velocity of the art and how much it’s come... in three years, let alone six, nine, ten years in the future.” He does concede a longer-term growth case exists, but distinguishes that from the disruption he expects in the near term.
The steel belt analogy and the safety net gap
Chieh’s clearest articulation of the risk draws on U.S. industrial history: comparing AI-driven displacement to deindustrialization, he notes that while economists can say offshoring was “a net benefit for the U.S.” in aggregate, “you go to the steel belt and get on in a town square there and tell them that — I don’t think you’re going to make a lot of friends.” His core structural concern is that “we don’t have great safety nets for folks,” unlike other countries where displaced workers “can figure things out, you can retrain.” This gap is the business rationale behind his company: working with employers to fund outplacement and support services for workers being displaced, including those leaving voluntarily.
On China and the AI race
Asked about the technological threat posed by China, Chieh is measured but direct: most entrepreneurs build technology “because they like the technology and they see opportunity,” but AI is powerful enough that “governments invariably need to get involved.” He characterizes the competitive dynamic bluntly from a U.S. vantage point: China’s government is “playing with a stacked deck” and “crushing us short term on price,” flooding markets with units regardless of cost. His prescription is that the U.S. government needs to actively support domestic entrepreneurs to avoid losing the race, drawing a direct parallel to internet-era leadership: “if we were not the leaders of the internet over the last 30 years... take a look at what our economy would have become.”
On where AI development stands today
Chieh offers a specific historical comparison for AI’s current market stage: “I actually think we’re probably in the equivalent of ‘99” — the frothy, skepticism-laden period just before the dot-com crash, when people said the same thing in ‘01 about “the death of the internet,” only for “almost everything that was promised” to become true two decades later. He frames current AI skepticism as a similar, likely temporary, sentiment cycle.
On corporate domicile
Asked whether he’s moving his company to Texas or Florida, Chieh says the question has come up, notes his company is incorporated in Delaware, and allows that “we should have taken a closer look at Texas and maybe we will” — framing state-level competition for corporate domicile as a healthy dynamic.
Investing and Business podcasts
This week in AI: Grok is Back, ChatGPT 5.6, Meta’s new models, China’s Chip Gambit on Limitless Podcast
Grok 4.5: SpaceX AI’s comeback
Ejaaz and Josh frame Grok 4.5 as a major turnaround for SpaceX AI, which they say had fallen visibly behind OpenAI and Anthropic before this release. The model ranks roughly third on general benchmarks behind Fable 5 and GPT 5.5, but stands out on coding tasks and cost — priced at around half of Opus 4.8 and, per one benchmark cited, roughly 17 times cheaper on a cost-to-solve basis, solving problems in about 16,000 tokens versus Opus’s 67,000. The hosts attribute much of the jump to data from Cursor, which SpaceX AI is reported to have acquired for $60 billion; they argue the value came less from raw compute and more from Cursor’s routing data teaching the model how to reason across coding tasks. Elon Musk reportedly said SpaceX AI plans to ship a new foundation model every month through the end of 2026, a claim the hosts treat with some skepticism despite the company’s reported 500,000-GPU cluster.
GPT 5.6 and the new voice mode
OpenAI’s GPT 5.6 launched alongside two smaller models (Sol, Terra, and Luna, per the hosts) after clearing a government review process the hosts don’t detail further. Pricing is cited at $5 per million input tokens and $30 per million output tokens. The bigger story, in the hosts’ view, is a new full-duplex voice interface built on what they call GPT Live 1 — allowing users to talk and listen simultaneously and interrupt the model mid-response. Ejaaz calls it “OpenAI’s biggest release so far this year,” arguing voice will become the primary way people interact with AI models going forward. Both hosts note that voice features have historically run on older, less capable models, so this update effectively brings voice interaction up to frontier-model intelligence for the first time. GPT 6, described as OpenAI’s next flagship model, is reportedly expected within about a month.
China’s push toward custom AI chips
The hosts discuss reports that DeepSeek and a company referred to as Jipu are each developing proprietary AI chips to reduce dependence on Nvidia and Huawei hardware, following a reported Chinese government directive for domestic AI labs to use Chinese-made chips. Josh frames this as part of a broader shift from open-source to more closed, vertically integrated AI development in China, comparing it to OpenAI’s own in-house chip effort. Ejaaz argues export controls may have inadvertently accelerated this shift by forcing Chinese labs to build independent chip capacity rather than remaining reliant on U.S. hardware, and suggests the performance gap between U.S. and Chinese AI chips has narrowed from roughly three years to about six months.
Meta’s new image and video models
Meta released updated image and video generation models (referred to as Muse Image and Muse Spark), which the hosts say represent a meaningful quality jump from Meta’s earlier attempts but still fall short of frontier standalone image tools. The notable technical feature is an “agentic” reasoning process — the model searches for reference images, generates a draft, evaluates its own output, and revises before returning a result. Both hosts say they’re unlikely to personally switch to it given stronger existing options, but view it as a sign Meta is making real technical progress.
ByteDance’s image model and Cloudflare’s micropayments gateway
ByteDance released a new image-editing model the hosts describe as highly competitive on cost and precision editing, comparing it favorably to GPT Image tools for design-specific use cases. Separately, Cloudflare launched a payments gateway enabling small, per-request micropayments — potentially allowing AI agents to pay cents for individual pieces of content instead of full subscriptions. The hosts cite a Cloudflare figure that AI crawler traffic now accounts for 52% of all crawler requests, up from 22% in early 2025, and note AI crawlers pull far more content than they refer traffic back to sites.
David Senra Breaks Down Zero to One: Peter Thiel's Blueprint for Building a Monopoly on Founders Podcast
The real lesson from Steve Jobs
Senra opens with Thiel’s framing that Jobs’s greatest achievement wasn’t product design — it was designing Apple’s business itself, through “definitive multi-year plans” rather than focus groups or short-term thinking. Thiel notes that when the iPod launched in 2001, analysts dismissed it as “a nice feature for Macintosh users” — missing that Jobs had planned it as the first in a new generation of post-PC devices. The broader point: “a business with a good definitive plan will always be underrated in a world where people see the future as random.”
The contrarian question
Thiel’s starting exercise for founders is: “What important truth do very few people agree with you on?” Good answers, he writes, take the form “most people believe X, but the truth is the opposite of X.” Thiel’s line on why so few people attempt this: “brilliant thinking is rare, but courage is even in shorter supply than genius.”
Rejecting the dot-com playbook
Thiel describes running PayPal during the late-’90s bubble, watching founders plan IPOs from their living rooms before even incorporating — “everyone else in the valley was ready to believe anything at all.” Post-crash, the industry converged on four defensive rules: make incremental advances, stay lean and flexible, improve on the competition, focus on product not sales. Thiel inverts every one of them: it’s better to risk boldness than triviality; a bad plan is better than no plan; competitive markets destroy profits; and sales matters as much as product.
What a “creative monopoly” actually means
Thiel is explicit that he doesn’t mean monopoly in the predatory, robber-baron sense. Instead: “creative monopolists give customers more choices by adding entirely new categories of abundance to the world.” His sharpest line on the alternative: “all happy companies are different. Each one earns a monopoly by solving a unique problem. All failed companies are the same: they failed to escape competition.”
Start small, then dominate
Thiel’s sequencing advice is to deliberately target a tiny, underserved niche first, then expand outward — citing Amazon’s internal code name, “the everything store,” and Bezos’s decision to start with books alone. “Sequencing markets correctly is underrated,” Thiel writes, “and it takes discipline to expand gradually.” Senra pairs this with the detail that Apple’s first sale was 50 computers for $25,000 — sold, per Apple lore, with Jobs walking into the buyer’s shop barefoot.
Last mover, not first mover
Thiel downgrades the value of being first: “moving first is a tactic, not a goal... it is much better to be the last mover” — the company that makes the final major advance in a market and then holds it for years or decades. He compares business to chess, quoting a grandmaster: “to succeed, you must study the endgame before everything else.”
Durability over growth
One of Senra’s favorite lines from the book: “growth is easy to measure. Durability isn’t.” Thiel argues founders who chase weekly active-user stats and quarterly targets can hit every number while missing the only question that matters — “will this business still be around a decade from now?”
The power law and the case for secrets
Thiel argues nearly everything in venture outcomes, careers, and decision-making follows a power law rather than a normal distribution — a small number of things (one market, one distribution channel, one decision) matter disproportionately more than all the rest combined. Because the power law isn’t obvious, the most valuable insights are often literal secrets: “every one of today’s most famous and familiar ideas was once unknown and unsuspected. A conventional truth can be important, but it won’t give you an edge.” Thiel’s warning on why people avoid this path: “people are scared of secrets because they are scared of being wrong.”
Sales is hidden, not absent
Thiel devotes a full chapter to distribution, arguing Silicon Valley chronically underrates it: “superior sales and distribution by itself can create a monopoly even with no product differentiation. The converse is not true.” He points out that sales roles are almost always disguised under other titles — “account executives,” “business development,” “investment bankers” — because “sales works best when it’s hidden.” His prescription: treat distribution as part of product design, not something bolted on afterward.
Founders: irreplaceable and dangerous
Thiel’s closing argument is that founder-led companies outperform precisely because founders carry extreme, contradictory traits — traits that would be mutually exclusive in most people. He uses Howard Hughes as a cautionary tale of what happens when those traits go unchecked over decades, contrasted with Steve Jobs’s return to Apple in 1997 as proof of a founder’s irreplaceable value: “Apple’s value crucially depended on the singular vision of a particular person... we need unusual individuals to lead companies beyond mere incrementalism.”
Historian Dr. Joseph Moore on 300 Years of Financial Advice That Worked (and Didn't) on The Meb Faber Show
Why “never save money” was once the best advice in America
Moore explains that before the Civil War, the U.S. federal government didn’t print money — currency was issued by individual banks and companies, redeemable only as long as the issuer stayed solvent. This meant holding cash was dangerous, and the standard advice, as he quotes a grandfather telling his grandson in the book, was: “never save money... get it out of your hands as fast as you can.” Immigrant families who didn’t understand this rule often lost everything when the bank or individual who issued their currency went under or vanished. Moore illustrates this with the story of William Wells Brown, a formerly enslaved man who opened a barbershop in Michigan and paid customers with his own printed scrip — money that circulated for a year before becoming worthless once he left town. Moore argues the “greenback dollar” introduced during the Civil War was revolutionary precisely because it let ordinary Americans hold onto money without fear of it going to zero.
Bitcoin as a “Zeppelin,” and the future of crypto
Moore’s provocative thesis, laid out in an essay called “Bitcoin is a Zeppelin,” is that Bitcoin proved digital currency could work the way Zeppelins proved flight was possible — but neither turned out to be the practical long-term solution. He argues crypto’s lasting value will come not from Bitcoin’s original ideological goal of undermining fiat currency, but from dollar-backed stablecoins letting people in unstable economies transact in dollar value — ironically extending the life of the fiat system Bitcoin was built to challenge.
Real estate: a path to a modest fortune, not a great one
Moore is blunt that real estate “is a great way to build a modest fortune” but “not the way to build a great fortune,” pointing out that zero of the 100 largest U.S. fortunes were made primarily in real estate. He argues the widely held belief that real estate reliably appreciates is a relatively recent and historically unusual idea, backed by Federal Reserve data showing flat inflation-adjusted home prices in cities like Pittsburgh, Atlanta, and Houston from the 1890s to the 1990s. What made real estate valuable historically, he argues, wasn’t appreciation — it was that families could leverage 95 cents on the dollar into something that held its value, at a time when currency itself was untrustworthy.
His own real estate scars
Moore recounts buying his first rental property, which quickly escalated into owning a fourplex plus a single-family home. Early on, he says “human traffickers moved into the house,” and a separate property had “hoarders with rats” — one of whom “pulled a shotgun on me,” forcing Moore to “leap off the deck” and run for his car. He credits sticking with the business — eventually growing to over 40 properties — with building meaningful wealth, framing real estate investing as something that only works as an active business, not a passive asset class: “if you just want the passive returns from real estate... you want a bond with an HVAC.”
The stocks-vs-bonds myth
Moore challenges the popular assumption that stocks always outperform bonds long-term, citing recent academic work showing bonds actually beat stocks through a lengthy stretch of the 1800s and that the two asset classes were roughly tied until World War II. He notes that in 1912, the near-universal financial advice was to buy bonds over stocks — a view that only flipped in the 1910s-20s as unprecedented inflation (prices doubled in five years) and new income taxes pushed everyday Americans into stocks for the first time, giving rise to the “invest for the long run” philosophy that still dominates today.
Is trying to beat the market worth it?
Moore argues that while it’s possible to beat the market — citing Warren Buffett as proof — it’s rarely worth an average investor’s time. He notes that even elite managers overseeing $600 million-plus tend to outperform by only about half a percent annually, achieved through 70-hour work weeks and elite credentials; applied to a median 401(k) balance, that same effort might net an extra $300-500 a year. He shares his own experiment trading based on the “Cramer bounce” — a documented pattern where stocks recommended on Mad Money tend to revert — noting that while he technically beat the market doing it, “I hear my wife holler, ‘Come quick, she’s doing it,’ and I missed my daughter’s first steps.”
Dividends: the most misunderstood part of investing
Moore says the disconnect between how people think about dividends and how they actually work is severe — citing survey data suggesting a majority of retail investors, and a significant share of professionals, mistakenly believe dividends are “free money” that doesn’t reduce the stock’s value, treating them like a bond coupon. He notes that historically dividends were 96% of stock market returns from the George Washington era until the mid-1980s, but today the S&P 500’s dividend yield sits near an all-time low around 1%, with almost all returns now coming from price appreciation — a structural shift he says most investors haven’t internalized.
Five core lessons from 300 years of financial advice
Moore distills his research into five recurring principles: solve other people’s problems (most financial advice focuses on the individual’s own spending, which “can keep you from going broke” but rarely builds real wealth); take more risks, since modern safety nets like insurance and bankruptcy protection make failure far more recoverable than in most of history; move toward opportunity (noting that Americans change addresses far less often today — one in 13 annually versus historically one in three); marry well, since marriage stability strongly predicts financial outcomes; and stay optimistic, which he argues is “wildly underrewarded” in advice but “hugely overrewarded” in the actual economy.
Economist Jim Paulsen on Excess Returns podcast
The call: a 10-20% correction, concentrated in tech
Jim expects a correction of roughly 10-20% over the next several months, driven primarily by “new era” (tech and communication services) stocks, which he estimates could fall over 20%, while the broader market declines a more modest 10% or less. He’s careful to frame this as tactical, not a recommendation to exit tech entirely: “I’m not suggesting anyone tell sell all their new era stocks... I am thinking they may want to go to an underweight.” He does not expect a recession or a full bear market, and believes the broader multi-year bull market trend remains intact into 2030 or beyond.
Labor market: metrics that “only occur in the middle of recessions”
Jim highlights that averaged household and payroll employment growth has been essentially flat (net zero) over the trailing year — something he says has never previously been considered acceptable in his career dating to the early 1980s. More concerning to him is that the unemployment rate and part-time employment are both rising simultaneously, a combination he says historically only shows up mid-recession, not during an expansion. He also flags that labor force growth has turned negative and full-time employment is declining, noting that without the drop in labor force participation, the unemployment rate would likely be near 5% rather than in the 3% range.
Policy has quietly tightened since the Iran conflict
Jim argues that since hostilities with Iran began, several economic supports reversed: the 10-year Treasury yield rose roughly 70 basis points to nearly 4.60%, real money supply growth fell back toward roughly a quarter of a percent year-over-year, the yield curve re-flattened, and the U.S. dollar rose 5-6% in nominal terms. He also notes federal deficit spending relative to GDP has contracted roughly 2 percentage points, from about 7% to 5%, over the past 15 months. Using a composite policy index weighted toward Treasury yields, the dollar, and oil prices, he projects — with roughly a 3-month lag — a meaningful slowdown in economic momentum through the balance of the year.
Oil price peaks tend to bite with a lag
Jim notes that in every major oil price spike since 1970, the real damage to the stock market and economy has historically emerged after oil prices peaked and pulled back — not while headlines were focused on the spike itself. He believes oil has likely peaked for this cycle and expects the delayed economic drag to show up in the coming months, along with a similar lagged drag from the flattening yield curve hitting S&P 500 earnings roughly 12 months out.
A historic earnings and spending bifurcation
Jim points to a stark divergence: forward 12-month earnings for the “new era” sectors (information technology and communication services) have spiked dramatically, while the other nine S&P 500 sectors’ combined forward earnings have been nearly flat for two to three years. He shows a similar pattern in real GDP, where new-era investment spending (13% of real GDP) has grown at an 8% annualized pace over six quarters versus roughly 1% for the remaining 87% of the economy. His question: “How long can one-third be on fire and most of the rest of it is barely moving?”
Complacency, not necessarily optimism
Jim distinguishes between optimism and complacency, arguing the data shows the latter. Individual investor stock exposure relative to cash is near its highest level on record, behind only the dot-com peak. He also flags a growing divergence between the Morning Consult consumer sentiment survey and the S&P 500 — the two moved together for most of the bull market but have diverged sharply since the March 30 AI surge began. Corporate and household cash as a percentage of GDP, historically closely tied to market performance, has also decoupled from stock prices since the bull market began.
Valuation: a 60% premium to trend
Using a measure comparing the S&P 500 to its long-term trend line since 1950, Jim notes the index sat at roughly a 20-23% premium in late 2025 — elevated but within historical range. In the past six months, that premium has jumped to 60% above trend, a level he says has only been exceeded once, during the dot-com bubble.
Defensiveness has left the market
Jim highlights that defensive sectors (utilities, consumer staples, healthcare, REITs) now make up only 16-17% of S&P 500 market cap, down to roughly half the weighting seen at the 2009 market bottom — meaning less downside protection and, historically, higher volatility. A separate measure comparing low-beta to high-beta stock performance going back to 1962 shows defensiveness at one of its lowest readings on record, with prior similarly low readings coinciding with major market tops including 1973, 2000, 2008, and 2020.
A quiet leadership rotation already underway
Jim’s most novel data point: since October of last year, the nine “old era” S&P sectors have outperformed the two “new era” sectors (tech and communication services) — and as of early July, “old era” stocks have now beaten “new era” stocks over a full trailing 12-month period for the first time in this bull market. He frames this as the second distinct leadership regime of the current bull run, following roughly two years of tech dominance from October 2022 to October 2025.
Demographics as the deeper five-year concern
Beyond the near-term correction call, Jim’s larger worry is demographic: U.S. labor force growth has slowed to roughly half a percent annually, and he shows a tight historical correlation between five-year trailing labor force growth and both GDP growth and bond yields. Based on consensus demographic forecasts, he projects GDP growth may struggle to exceed 1.5-2% over the next five years, inflation could run closer to 0-1%, and the 10-year Treasury yield could fall back toward the 3% range or lower — a view he says runs counter to widespread expectations that yields are headed structurally higher.
If you found these summaries helpful, please leave a comment or send us a message. We appreciate your feedback!