Jul 24, 2026 · 1h 8m · sourcery
BlackRock's Tony Kim on AI's Next Winners? Chips, Memory, Robotics & Quantum · Sourcery with Molly O'Shea
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
Tony Kim, Head of BlackRock's Global Technology Team, joins host Molly O'Shea on Sourcery to discuss the macro pivot from software to compute-centric AI infrastructure, market capital reallocation, hardware bottlenecks, and the future of robotics and frontier tech.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Molly holds 15.3% of the talking time here. How this is scored →
Molly as informed peer 4.6 Guest teaching 4.9 Guest disagreement 1.8 Molly pushing back 0.7
speaking balance: gold is Molly, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 39:46 Tony rejects the commodity chip thesis
Tony passionately dismisses the conventional Silicon Valley belief that chips were merely commodities, pointing out they hold the highest profit margins in the entire economy.
Hardest push from Molly ▶ 57:56 Molly challenges PE AI rollupsMolly pushes back on the premise of PE rollups by arguing they are primarily acquiring customer bases rather than fundamentally new technology products, citing General Catalyst's Creation Fund.
Biggest teaching moment ▶ 21:20 Tony details the 10/20/30 market cap reallocationTony walks through the dramatic redistribution of global market caps from software to hardware, detailing the structural transformation from the BCE to AD era of tech.
Molly holds their own ▶ 54:27 Molly connects enterprise context layer to Palantir ontologyMolly demonstrates strong domain understanding by instantly identifying Tony's enterprise context layer description as Palantir's ontology framework.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Molly as informed peer | Guest teaching | Guest disagreement | Molly pushing back | Why |
|---|---|---|---|---|---|---|
| RAISE Summit Paris Overview & Panel Topics | 4 | 2 | 1 | 0 | Molly opens the interview in Paris at RAISE Summit, asking Tony to overview his panels and presence at the conference. The tone is casual, collaborative, and conversational. | |
| The Shift to Compute-Centricity & Data Center Redesign | 3 | 6 | 2 | 0 | Tony breaks down the macro shift from software-centric to compute-centric architecture and the physics-driven redesign of data centers. Molly lets him unpack his detailed thesis without interruption. | |
| Power Architecture & Grid Infrastructure Revolution | 4 | 6 | 2 | 1 | Tony details the energy architecture transformation and memory bottlenecks ('RAMpocalypse'). Molly playfully corrects who originated the term RAMpocalypse before letting Tony educate on duration mismatches. | |
| Sponsor Segment - Brex | 5 | 7 | 2 | 1 | Following the Brex ad, Molly brings in Coatue public market insights and SK Hynix dynamics to set up Tony. Tony delivers an extensive masterclass contrasting cloud economics with the massive capital requirements of AI. | |
| Capital Allocation & The 2030 Horizon | 5 | 5 | 1 | 1 | Molly asks how BlackRock evaluates long-dated technology risk like quantum computing. Tony details portfolio construction across the current 3-year vortex versus long-tail bets converging on 2030. | |
| Sponsor Segment - MongoDB & AssemblyAI | 6 | 6 | 3 | 1 | Molly details her visits to Figure AI and Boston Dynamics to discuss robotics form factors. Tony vigorously refutes the historical view that semiconductors are commodities and analyzes the wave of Chinese robotics IPOs. | |
| Companion Robotics Form Factors | 4 | 4 | 1 | 0 | Molly and Tony discuss companion and eldercare robotics versus humanoids. Tony explains why smaller non-threatening form factors like Star Wars droids will unlock massive demographic utility. | |
| Downstream Data Layer, Token Flow & Enterprise Refactoring | 6 | 5 | 3 | 2 | Molly notes how speech models and data layer companies benefit downstream, offering Palantir's ontology concept as an analog. Tony nuances token flow dynamics, and they debate private equity AI rollups. | |
| 12-Month Outlook & Breakthrough Technologies | 4 | 3 | 1 | 0 | Molly closes by asking about Tony's 12-month outlook and his personal mentors, leading into Tony's reflections on historical leaders and individuals who gave him a chance early in his career. |
Statements from this episode (16)
Tony Kim: Tech sector has shifted from software-centric to compute-centric allocation
“We went from a, let's call it a
A software-centric world to a compute-centric world.”
Tony Kim: AI compute requires $10 trillion in CapEx over five years
“So this is the irony of it all, that we're, the trillion dollars of CapEx this year, and the 10 trillion over the next five years that are coming,
Is to move data centimeters and millimeters.
That's AI.”
Tony Kim: Data centers are shifting from copper to optical interconnects
“Another will be around the chip architectures, and then another one will be around the movement of data, and so we're going from a regime of copper to a regime of light.”
800-volt power architectures and solid-state transformers will transform AI data centers
“The rise of 800 volt. It's gonna have a transformative effect. And then ultimately a solid state transformers.”
Tony Kim: Memory primacy will surpass pure compute in upcoming AI architectures
“Today we're all talking about compute, compute, compute. I think the primacy of memory will become even more important.”
Tech revenues are shifting from high-margin cloud to lower-margin AI compute
“Like, these, you must build these new data centers, because at some point, all of your revenue will become here, and this revenue over here in the old data center, that was asset light, high margin, is now moving to asset heavy, lower margin, big dollars, but …”
Global tech market cap is $10T software, $20T Mag-7, $30T hardware
“10 trillion in software service internet, 20 trillion in Mag-Seven, 30 trillion in non-Mag-Seven compute chips hardware. That's a, I don't think people would realize that we are that compute hardware centric now. 10, 20, 30. Before AI, BCE era, it was probably…”
Tony Kim: AI foundation models have drained market cap from traditional software
“The models themselves, like it or not, have consumed the market cap out of software and services. It is just, you know, it's like the Borg. It is consumed.”
Tony Kim targets 90% of his investments within a three-year AI horizon
“In this kind of three year window, That is probably where 90 plus percent of my investment, well, ninety-ish percent, something like that, you know, the majority is going in.”
Tony Kim: Utility-scale, million-qubit quantum computers will arrive around 2030
“It's very interesting when you look at these long-dated technologies. All roads converge to 2030. It's like quantum computing, utility scale, logically error-corrected, million-qubit quantum computer, twenty-thirty.”
Tony Kim: Chip companies have higher profit margins than software or pharmaceuticals
“They have the highest profitability of any company in the world of the sector is chip companies. They're higher margins than software, pharmaceuticals, industrials, telecom, anything.”
Tony Kim: China has 30 to 40 potential robotics IPOs this year
“I'm looking at the current pipeline. I see 30, 40 potential IPOs this year in China.”
Future robotics will pair Chinese-manufactured physical bodies with Western AI brains
“And then what you might have is ultimately you can mix and match Chinese physical robot with a Western brain. And I know that's happening.”
Companion robots do not require perfect human-scale motor articulation for market success
“Do you need to have A fully figured, perfect motor function with, you know, all the hand articulation, or could you get something that can appeal to that? And I think that's possible. I think that will be a fascinating market to see. So yeah, I think that'd be…”
Tony Kim: Enterprises must position themselves within the AI token flow
“So, you must be in this token flow to either resell or repackage the tokens with your context and your very specific application. You serving the token with your intelligence? Is it the proprietary closed source token? Is it the open source token? Sitting on a…”