Aug 10, 2026 · 53m · big-technology
Demis Steps Down, Apple’s Memory Problem, Microsoft’s Clever Trick
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
Alex Kantrowitz and guest MG Siegler analyze major shifts across the tech industry, including Demis Hassabis stepping down as CEO of Google DeepMind, Apple's AI-driven memory supply shortages and hardware pricing strategy, and Microsoft's accounting maneuvers amid soaring AI capital expenditures.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Alex holds 36.7% of the talking time here. How this is scored →
speaking balance: gold is Alex, purple is the guest (3 minute bins)
MG pushes back against Alex's critique of the 25-year schedule by pointing out 'there's levels to this' and separating real estate leases from replaceable GPU hardware.
Hardest push from Alex ▶ 37:01 Challenging Microsoft's 25-Year Depreciation NarrativeAlex refuses to accept the 25-year data center accounting premise, arguing that GPUs represent half the cost and cannot survive past three to seven years.
Biggest teaching moment ▶ 33:17 Explaining CapEx to OpEx ReclassificationMG explains the accounting mechanics of how lengthening lease amortization schedules allows Microsoft to shift $15 billion from CapEx into OpEx.
Alex holds their own ▶ 37:01 Dissecting GPU Cost Fractions in AI Data CentersAlex demonstrates deep subject knowledge by citing hardware replacement metrics and showing why long-term depreciation schedules fail to reflect GPU capital reality.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Alex as informed peer | Guest teaching | Guest disagreement | Alex pushing back | Why |
|---|---|---|---|---|---|---|
| Emergency Analysis: Demis Hassabis Steps Down at Google DeepMind | 0 | 0 | 0 | 0 | Solo emergency monologue by Alex Kantrowitz analyzing Demis Hassabis stepping down as Google DeepMind CEO. Host scores are set to zero per monologue rules. | |
| Big Technology Podcast Friday Edition Introduction with MG Siegler | 6 | 4 | 1 | 2 | Alex and MG discuss Apple's post-earnings drop and memory component supply squeeze. Alex offers an informed hypothesis about Apple using memory shortages to justify raising hardware margins, while MG builds on Apple's pricing history. | |
| Upcoming iPhone Price Hikes, Services Slowdown, and App Store Scrutiny | 6 | 4 | 1 | 2 | The conversation covers upcoming iPhone price increases, the unexpected softening of Services growth, and App Store antitrust scrutiny. Alex probes whether incoming leadership will look weak if they concede to EU regulators, while MG elaborates on App Store gaming dynamics. | |
| Hardware as a Service: Apple's Device Subscription Strategy | 5 | 4 | 1 | 1 | Both analysts examine Apple's hardware subscription initiative via Klarna as a way to soften consumer sticker shock. Alex highlights how a $30 monthly fee changes consumer psychology compared to a $2,000 upfront purchase. | |
| Microsoft's Data Center Lease Accounting and CapEx Reclassification | 5 | 6 | 2 | 2 | Alex introduces MG's reporting on Microsoft extending data center lease accounting from 15 to 25 years. MG explains how shifting expenditures from CapEx to OpEx gave Microsoft an optical advantage on Wall Street. | |
| Debating GPU Lifespans Versus 25-Year Data Center Depreciation | 8 | 5 | 3 | 7 | Alex aggressively challenges Microsoft's 25-year depreciation timeline, citing short GPU lifespans and heavy hardware replacement costs. MG clarifies the accounting distinction between the physical building shell and the internal server hardware. | |
| Cloud Revenue Surges and Circular AI Capital Investment | 7 | 5 | 1 | 3 | Alex questions the authenticity of massive hyperscaler cloud revenue surges, noting the circular nature of funding OpenAI and Anthropic. MG reinforces the point by breaking down Amazon's $50 billion investment tranche and Satya Nadella's hedging strategy. | |
| Google's Internal AI Battles: Large Language Models vs. World Models | 7 | 5 | 2 | 2 | Alex synthesizes an analysis regarding Google's strategic prioritization of world models over LLMs due to research interests. MG agrees while evaluating how recursive self-improvement in LLMs could challenge that strategy. |