Podcast Alpha
@PodcastAlphaX • 15,429 subscribers
Signal. Not noise. Forensic breakdowns of the world’s most influential podcasts. Substack - https://t.co/1GNcZMJIWU DM to remove clips
Videos
Gavin Baker Gavin Baker: HBM DRAM will be 30-40% of all hyperscaler capex in 2027. Baker manages Atreides - his 2025 Micron call is now 14x. This is his next call. HBM requires stacking 12-16 DRAM dies in a single package. Only three companies can do it: Micron, SK Hynix, Samsung. No fourth supplier arrives in 2027. Hundreds of billions annually to three firms. Micron's new supply chain agreements lock in floor pricing above prior cycle gross margin peaks. If you still hold $MU at a commodity discount to $ASML and $LRCX, Baker says that discount is no longer earned. Baker's full case at The All-In Podcast: Source: All-In Podcast -
Podcast Alpha503,026 views • 1 month ago
Ken Griffin just priced a Taiwan invasion: US GDP down 8% in six months. A great depression, that fast. Griffin runs Citadel, one of the largest trading firms in the world, and he does not deal in hyperbole. Lose access to Taiwanese chips and Boeing stops making planes, most new cars stop, consumer electronics stop. $TSM sits in every high-end product, so the whole economy freezes at once. Dylan Patel Dylan Patel of SemiAnalysis makes the same case: fab concentration in Taiwan is the choke point of the entire compute supply chain. If you hold semis or any global supply chain, this is the tail to size, and it is not the same loss for a US book and a European one. The full note maps where the sanctions coalition actually cracks: Source: Goldman Sachs Exchanges -
Podcast Alpha307,831 views • 1 month ago
Levie now uses Salesforce 5x more than at any point before. The Box CEO Aaron Levie connected Salesforce's MCP server to Claude Code. Now he runs customer and market intelligence queries he would never have bothered pulling up by hand. The agent removes the friction. The underlying system gets queried more, not replaced. If you hold $CRM, the agentic era is an engagement tailwind, not a disruption risk - gated on whether the data platform handles the query load. Why incumbents gain from agents: Source: CXOTalk -
Podcast Alpha263,785 views • 2 months ago
Dan Ives Dan Ives just admitted he was "dead wrong" on $ORCL. Gil Luria called the top. Now Luria is the one saying look again. Oracle ran 200 to 330 on a single OpenAI backlog deal, then round-tripped to 140. The market now values the entire ~$630B compute backlog at zero. But OpenAI raised $122B, the largest raise ever, and narrowed to compute, so those bills are payable. Zero starts to look wrong the other way. Jim Chanos James Chanos, on Risk Reversal back on Jul 17, called the AI-compute backlog a mirage. This is the direct bet against that. The read - at a zero backlog, any conversion at all is upside. The risk sits with OpenAI's commitments, not Oracle's execution. Full exchange:
Podcast Alpha97,866 views • 26 days ago
Gavin Baker Gavin Baker: the Anthropic S1 is going to break a lot of people's brains. The entire AI bear case rests on one assumption, that tokens are sold below cost. What the filing settles, and who has to cover: Baker says that assumption is wrong. Open-source tokens are profitable. Anthropic is profitable and generating cash. OpenAI is close. Most tokens in the chain make money today. The S1 is the first audited document that settles it either way. Jim Chanos James Chanos has spent the year on the other side, calling the AI buildout a circular financing story. Plain version for a position holder: a lot of macro and value money is bearish on AI because it believes tokens are subsidized. If the filing shows real margins, that money covers into a re-rating and $NVDA, $TSM, $MU and the hyperscalers all move on it. If the filing shows subsidized tokens, every valuation downstream of it is sitting on sand. One document, one binary, a date on the calendar. David Sacks David Sacks and Jason Calacanis @jason worked him over on it for a full segment. Source: All-In Podcast -
Podcast Alpha29,366 views • 8 days ago
Ken Griffin: China now leads in 67 of 74 critical technologies. The consensus that the US pulled ahead is the trade he is fading. The Citadel founder says China leads from solar to EV batteries to quantum, powered by 1.4 billion people and far more STEM graduates. His fix is not tariffs. It is educating US youth to out-innovate. Alexandr Wang Alexandr Wang, Meta's chief AI officer, pushes the other way and calls the US clearly ahead in AI. That is the bull/bear line to watch. The bull case for Chinese leaders in solar, batteries, and quantum is a decade-long trend, not a trade, and tariffs are the wrong lever to short it. Full breakdown of the China-tech scorecard and the trade read: Source: Goldman Sachs Exchanges -
Podcast Alpha155,613 views • 1 month ago
$GOOGL has an anti-aging asset nobody is pricing. Friedberg david friedberg walked through a Calico paper that should be on Alphabet holders' radar. Calico is Google's longevity lab. Working with Revel Pharma, it used AlphaFold plus directed evolution to design a novel enzyme that degrades CML, the glycation end-product that stiffens tissue and drives aging. The result is not a slide. Across five recursive design cycles the enzyme degraded **52-97%** of CML on human proteins, and on donated skin from patients over 70 it eliminated about **55%** of CML - reversing the skin's profile toward that of a 31-year-old. It is AI-designed proteins doing something genuinely new in biology, not a benchmark score. Chamath Chamath Palihapitiya frames the first market as cosmetic: a workable anti-aging cream is a multi-trillion-dollar category. This is the AlphaFold thesis paying out in a product line, and the AlphaFold-to-drugs bridge is exactly what Demis Hassabis Demis Hassabis has argued is the real near-term return on Google's AI - not chatbots, but designed molecules. Same company, same stack, a second revenue engine the market files under "research." Implication: this is deep-optionality inside a name you already hold on search and cloud. It does not move the $GOOGL model this year. But it is a reminder that Alphabet's AI edge shows up in drug and materials design, an option the Street assigns roughly zero value. If a Calico or Isomorphic asset reaches a real market, that is upside no analyst is underwriting. Watch for Calico or Isomorphic Labs to move a designed molecule toward a commercial or clinical path.
Podcast Alpha123,739 views • 1 month ago
PayPal peaked near $322B. It trades around $30-40B now. David Sacks David Sacks walked the fall: a 25-year-old product still growing about 7% a year, cut to a tenth of its peak value. That is not a broken business. That is a cheap one - and cheap plus a strategic acquirer is how you get an M&A event. Chamath Chamath Palihapitiya expects the ~$60-per-share bid to clear **10-to-15% above** the offer, and hinted at a rival bidder with PayPal roots and trillions to spend. That is the arb: the announced price is a floor, not the exit. Friedberg david friedberg framed it as the second instance of a pattern - AI-native operators buying stale, founder-departed businesses and reviving them. The first instance is the cross-podcast anchor. Ryan Cohen Ryan Cohen laid out a $56B plan to take over eBay ($EBAY), the same playbook: a dormant marketplace, a founder long gone, an operator who thinks software and distribution can restart growth. PayPal is that trade with a payments network attached. Implication for $PYPL holders: this is event-driven arb, not a re-rating on fundamentals. The upside is the bidding dynamic - floor bid plus a hinted second buyer - not a bet that PayPal suddenly grows faster. The risk is a deal break on antitrust or financing, which drops it back to a 7%-grower multiple. Size it as a merger-arb position, not a conviction long. Watch for a competing bid to surface. Chamath's "trillions to spend" hint is the catalyst that would reprice this fast.
Podcast Alpha127,062 views • 1 month ago
2,000 Cybercabs in one day. Not a city-by-city rollout. Robert Scoble Robert Scoble, who has tracked autonomous driving since the Stanford lab that became Waymo, says Tesla is staging Cybercabs in lots at Giga Texas to flood Austin in a single media event. A trickle is vulnerable to one bad incident. A mass launch resets public perception overnight. The tell is public: those staging lots show up in drone imagery. Watch for 2,000+ units before any soft cybercab launch. How the batch-launch call gets confirmed or killed, on Cern Basher: Source: Cern Basher -
Podcast Alpha195,412 views • 2 months ago
$MU at 6x earnings. $INTC at 100x. Same cycle. Gil Luria's cleanest dislocation on The Real Eisman Playbook Steve Eisman: the market prices Micron as if the memory cycle ends next year, and Intel as if it runs to 2030. Both cannot be true. His read - memory is now the better market than CPUs, which flips the historical order. So Micron at 6x while Intel sits at 100x is the inconsistency, and the trade. Jordi Visser Jordi Visser told Anthony Pompliano last week the easy money in the AI trade is over. This is the counter - the dispersion inside semis is where it is still mispriced, not gone. Net - a pair setup, not a beta bet. Long the multiple priced for a top, against the one priced for five more years. Full breakdown:
Podcast Alpha69,418 views • 26 days ago
S&P 500 earnings: +30% in 2 years (1998-2000). Then -40% in 12 months in a mild recession. That swing was not from the recession. It was from the telecom build-out collapsing. Companies realized they ordered 10,000 routers and needed 2,000. They pulled order books. Vendor revenues evaporated overnight. Depreciation costs kept running for years. Jim Chanos on iConnections: the AI CapEx boom has identical mechanics. Nvidia's earnings are CapEx boom output. S&P estimates are rising fast because infrastructure spending flows directly into a small number of vendors. If AI infrastructure spending slows - and order books get pulled - the earnings collapse is fast and severe regardless of how mild any recession is. The full dot-com analog and what it means for $NVDA positioning today: Source: iConnections -
Podcast Alpha145,017 views • 2 months ago
Dylan Patel Dylan Patel of SemiAnalysis: Anthropic's margin on an Opus 4.8 API token is north of 80%. It is net-income profitable excluding stock comp in Q2 2026, potentially profitable including it by Q3. Here is why that matters. At 80%-plus, even doubling compute costs leaves Anthropic above 50% gross margin. Every GPU it rents, at any above-market rate, is immediately accretive. It can outbid the whole market for scarce compute and still print money. Lower-margin labs cannot. The compute crunch is an Anthropic tailwind, via Sequoia Capital: Source: Sequoia Capital -
Podcast Alpha100,167 views • 1 month ago
$NVDA sits at its lowest forward P/E in a decade. Gavin Baker says that multiple is missing two moats. The Street prices it like AMD and custom silicon are closing the gap. Baker's case: Nvidia rolled out a financing structure that pairs equity stakes in AI buyers with a GPU-price-floor revenue share, indirect vendor credit since Nvidia is contractually barred from buying its own chips. It bridges buyers through their cash-flow-negative years and could grow into a large, royalty-like cloud business on its own. David Sacks made the parallel case on All-In, that July's selloff was leverage unwinding, not a break in AI fundamentals, the same read that keeps a name like Nvidia's actual moat intact even while the multiple compresses. The financing edge does not show up on a P/E screen. That is exactly why Baker says the market has not priced it.
Podcast Alpha34,734 views • 18 days ago
India's $200B IT services industry will be gone. Not "under pressure." Not "disrupted." Gone. Vinod Khosla - 40-year VC, $50M OpenAI backer in 2018 - says AI agents can do most of what TCS, Infosys, and Wipro charge for. The displacement is not a question of if. Only when. India's only path: pivot engineering talent to AI deployment for healthcare, education, and agriculture. Very few companies anywhere in the world know how to do this. The window is short. Full breakdown: Source: SparX by Mukesh Bansal -
Podcast Alpha122,074 views • 2 months ago
$NVDA is the "safe short." Same framing preceded Amazon's and Google's breakouts. Dan Loeb, Third Point ($30B AUM), on The All-In Podcast: the stock's persistent discount isn't a fundamental ceiling. It's crowded short mechanics. Long/short pods must carry a short book. At $5 trillion, $NVDA is the obvious answer - large enough to absorb notional, with a long run that gives narrative cover. That structural pressure is what suppresses the stock. Loeb's call: undervalued on 2-3 year forward earnings. When fundamentals force a re-rate, the short unwind will be faster than a fundamental re-rate alone. Google was a "safe short." Amazon was a "safe short." Both broke out. Full NVIDIA analysis + homebuilder short: Source: All-In Podcast -
Podcast Alpha119,471 views • 2 months ago
Bill Ackman Bill Ackman: Pershing Square just bought $2B of $MSFT. He funded it by selling part of the fund's $GOOGL stake. Ackman runs Pershing Square on 12-15 concentrated positions, so a swap this size isn't commentary, it's a live trade you can track. He's explicit it's not a quality call on Google. Alphabet's price ran up far enough that its forward return no longer cleared what redeploying the same capital into Microsoft could earn. Business quality first, then price, and price just flipped the math. Dan Nathan on Risk Reversal RiskReversal Media flagged the other side of this four days earlier: Google's chart never touched its 200-day average through the Mag 7 pullback, the clearest technical tell that capital is rotating back into the exact name Ackman just trimmed. If Nathan's rotation call is right, Ackman sold into buying pressure. Watch Pershing Square's next 13F to see whether the swap holds or reverses.
Podcast Alpha49,547 views • 1 month ago
$GOOGL fell 7-10% on its first negative free cash flow quarter since IPO. Chamath Palihapitiya Chamath Palihapitiya says buy the machine. Google has compounded capital at roughly 32% ROIC for 25 years, and it just guided capex to $195-205B. The Street read a red flag. Chamath reads a compounder spending into its lead. The deeper point - frontier models now commoditize within weeks, so value drains from the model layer toward infrastructure. That fragmentation hurts Anthropic and OpenAI and helps Google, which monetizes at silicon and cloud no matter which model wins. Cloud grew 82% on a $100B run rate. Gavin Baker Gavin Baker has argued hyperscaler capex is rational investment, not waste - the same read. The call - the negative-FCF print is an entry into the cleanest way to own the AI buildout without betting on a single model. Why model-layer commoditization is a tailwind for $GOOGL: Source: All-In Podcast -
Podcast Alpha42,627 views • 1 month ago
Fred Thiel Fred Thiel runs the largest public bitcoin miner. His number: a mining site costs about $1M per megawatt, all in. An AI site runs $10-15M per megawatt before a single GPU goes in. Same land, same power hookup, 10-15x the build cost. That's why Mara is converting, not mining. It already controls over 4GW after the Long Ridge and HIF deals, against 1.1GW of live mining today. The catch: Mara had no hyperscaler relationships and no build capability, so it's leaning on Starwood at 80% loan-to-value. The megawatts are real. The tenant demand is borrowed. Chase Lochmiller Chase Lochmiller of Crusoe makes the same energy-first case - the AI constraint isn't chips, it's power you can actually plug in. The tell for $MARA isn't hash rate anymore. It's how many of those gigawatts turn into signed AI leases. Thiel said it to Natalie Brunell Natalie Brunell ⚡️ on Coin Stories. How Mara funded the pivot and what breaks it:
Podcast Alpha35,196 views • 27 days ago
Power is 5-7% of data center revenues. Some geothermal and nuclear stocks are trading at 50-70x earnings to supply it. Jim Chanos on iConnections: the AI halo has reached its furthest extension into alternative energy. The bull case is that data centers need power and there's a shortage. Chanos's counter: the US doesn't have a power shortage. It has a permitting bottleneck and turbine delivery delays. Both resolve in 2-3 years. When the permitting clears, the scarcity narrative collapses. And even if these companies win a power contract - power is 5-7% of a data center's revenues. Companies at 30-40x EBITDA for potentially supplying a fraction of a 5-7% cost line are priced for a constraint that is transient. Full breakdown of the alternative energy overhang: Source: iConnections -
Podcast Alpha75,076 views • 2 months ago
If you hold $CRWV, Dylan Patel just named your real risk, and it is not performance. Dylan Patel of SemiAnalysis: CoreWeave benchmarks above Amazon and Google on GPUs, because hyperscaler multi-tenant security overhead hurts rented-rack performance. The weakness is the balance sheet. Neoclouds pre-sell compute six months before it is live, funded by debt backed by paper contracts. The first wobble already showed up: a Crusoe customer asked to pause construction. Watch the contract book, not the benchmarks. The neocloud thesis in full, from SemiAnalysis on Sequoia Capital: Source: Sequoia Capital -
Podcast Alpha55,590 views • 1 month ago