Jul 29, 2026 · 31m · tbpn
RSI Is Closer Than People Think, Per Tae Kim
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
Financial analyst Tae Kim joins the program to break down semiconductor fundamentals, debunk prevailing media skepticism around AI capital expenditures, and explain why upcoming catalysts like Recursive Self-Improvement will drive exponential compute demand.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. How this is scored →
The hosts as informed peer 5.1 Guest teaching 4.0 Guest disagreement 2.9 The hosts pushing back 2.4
speaking balance: gold is the hosts, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 16:50 Static view counterargument
Kim forcefully rejects the host's framing that capex must plateau, calling it a static perspective that ignores 40% to 80% hyperscaler revenue compounding.
Hardest push from the hosts ▶ 16:25 Host challenges capex infinity thesisThe host cuts in with 'Hold on, hold on, hold on,' pushing back on Kim's bullish cash flow assumptions by arguing revenues must catch up before capital runs out.
Biggest teaching moment ▶ 10:40 Kim dismantles the Kimi efficiency mythKim corrects market assumptions comparing Kimi to efficient small models, explaining its 2.8T parameters and 64-GPU cluster requirements require massive compute.
The host holds their own ▶ 2:39 Host maps Meta's internal AI reorganizationThe host synthesizes multiple sources—from SemiAnalysis and Alexandr Wang to town hall leaks—demonstrating deep industry tracking on Meta's compute strategy.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | The hosts as informed peer | Guest teaching | Guest disagreement | The hosts pushing back | Why |
|---|---|---|---|---|---|---|
| Assessing Prevailing Market Sentiment and Media FUD | 6 | 2 | 2 | 2 | The host demonstrates strong context on Meta's executive messaging, referencing Boz's podcast appearance, Alexandr Wang's statements, and SemiAnalysis reports. Tae Kim largely agrees, explaining how media headlines misrepresented an internal town hall leak as an official strategy shift. | |
| NVIDIA's Open Source Coalition and AI Regulatory Pushback | 6 | 2 | 2 | 2 | The host and Jordy track corporate coalition dynamics and regulatory debates surrounding open-weight distillation. Kim highlights how Jensen Huang and major tech players moved quickly to isolate Anthropic's regulatory push. | |
| Kimi Architecture, Agentic AI, and Recursive Self-Improvement | 4 | 6 | 3 | 1 | Kim educates the hosts on the architecture of Kimi, noting its 2.8 trillion parameters require immense compute rather than causing a compute glut. He argues that recursive self-improvement (RSI) is imminent and will exponentially drive GPU consumption. | |
| Hyperscaler Cloud Economics and the Enterprise Adoption Curve | 6 | 5 | 5 | 6 | The host directly intervenes to challenge Kim's free cash flow thesis, arguing that capex cannot compound infinitely without hitting a plateau. Kim counters vigorously that the host's view is static and ignores 40-80% cloud revenue expansion. | |
| Debunking Data Center Lease Headlines and Financing Rumors | 5 | 5 | 4 | 2 | Kim breaks down the sensationalized Financial Times headline regarding NVIDIA's $50 billion data center lease, showing it equates to a routine annualized expense. The hosts assist in contextualizing the Hut 8 deal structure and SoftBank financing rumors. | |
| NVIDIA's Co-Design Advantage, Supply Dominance, and Power Limits | 5 | 5 | 3 | 3 | The host raises AMD's thesis that agentic coding could erode CUDA's software moat. Kim responds by highlighting NVIDIA's hardware-software co-design, reliability, and aggressive prepayment lockups on HBM and optical components. | |
| NVIDIA's Strategic Ecosystem Investments and Market Outlook | 4 | 3 | 1 | 1 | The hosts and guest discuss NVIDIA's strategic investments in ecosystem companies like CoreWeave and optical suppliers, before wrapping up with friendly banter about Jensen Huang's leather jacket. |
Statements from this episode (14)
Kim: Negative AI Stock Sentiment Driven by Retail and Hedge Fund Unwinds
“So, I think sentiment's very negative. I've kind of had this Huge up, parabolic up move the last few months, and likely a lot of retail and hedge funds piled in, and we were seeing this unwind now.”
Kim: Media Misconstrued Mark Zuckerberg's AI Remarks to Falsely Claim CapEx Cuts
“I think we saw earlier this month I think Reuters quoted, like Zuckerberg's about agentic AI. They took it out of context, and then every Media person was with the hot take that this Meta was seeing bad returns, and they're going to cut CapEx, and then we had …”
Kim: Meta Will Need Six to Twelve Months for Further AI Progress
“It's going to take time. It's going to take six to 12 months before we see any more progress.”
Tae Kim: NVIDIA United the Tech Industry Against Anthropic
“I think it's been very impressive what they've been, they've basically united the entire tech industry against Anthropic in the last, like, three, four days.”
Tae Kim: Apple Would Benefit Most From Open-Weight AI Models
“Apple's the one, which is kind of strange, because they're the one that would most benefit from open source, open weight models being, you know, more available, I would think, but I don't know what Apple.”
Tae Kim: Kimi is a 2.8T parameter model requiring massive compute
“This is not a tiny efficient model. This is 2.8 trillion parameters. It's going to require a ton of compute to serve.”
Tae Kim: Agentic AI growth in next 6-9 months will exceed expectations
“Right now Agentsic AI is taking off, and the next six, nine months are going to be bigger than anyone believes”
Kim: AI Recursive Self-Improvement Could Arrive Within Nine Months, Massively Draining Compute
“RSI, I think, is a lot closer than people think. And if RSI actually happens in the next three, six, nine months, That's gonna soak up insane amount of compute.”
Kim: AMD CEO Lisa Su Raised Data Center AI Forecast to $220B
“Lisa Su raised her CPU, a genetic CPU forecast. Just three months ago, it was a hundred twenty billion for 20 to 30. Three months later, they raised it to two hundred twenty billion.”
Tae Kim: Morgan Stanley estimates AI inference yields 60-80% margins
“Like Morgan Stanley says, if you do inference, it's 60 to 80% profit margins, right?”
Kim: $200B in Current Hyperscaler AI CapEx Will Generate Massive Free Cash Flow
“So that's why you're investing two hundred billion dollars now, and in a year or two, we're gonna see insane amounts of free cash flow.”
Kim: NVIDIA's $50B data center lease requires a 30-year renewal
“It's like a 15 year lease and it's only fifty billion dollars if they renew the lease after 15 years. So it's like over 30 years if they renew it.”