TickerTain
TickerTain
NewsroomShortsPortfolioConvergence
NewsroomShortsPortfolioConvergence
$ANTHROPIC·$MA····$INTC····$BLUE-ORIGIN·$SPCX····$CRWV····$CRM····$MSFT····$NVDA····$ORCL····$CURSOR·$AAPL····$OPENAI·$AMZN····$UBER····$GOOGL····$META····$TSLA····$DATABRICKS·$PERPLEXITY·$LYFT····$NBIS····$TSM····$LITE····$ANDURIL·
$ANTHROPIC·$MA····$INTC····$BLUE-ORIGIN·$SPCX····$CRWV····$CRM····$MSFT····$NVDA····$ORCL····$CURSOR·$AAPL····$OPENAI·$AMZN····$UBER····$GOOGL····$META····$TSLA····$DATABRICKS·$PERPLEXITY·$LYFT····$NBIS····$TSM····$LITE····$ANDURIL·
←

tony kim

T3 · host / generalist

Tony Kim leads BlackRock's global technology investing across public and private markets. He has been investing in AI infrastructure since before the generative AI wave (e.g., SambaNova in 2019) and focuses on the physical layer of AI: compute, memory, power, optics, and data center redesign. He frames capital allocation around a three-year 'now' window and a 2030 frontier horizon (quantum, orbital compute, SMRs).

5 calls·5 names·100% bull·last heard 2 months ago·Sourcery VC
track record

no scored calls yet — needs a stated position or a categorical verdict, with a matured window vs SPY

top calls

highest conviction · one per company
1sthigh conviction
$SAMBANOVASambaNova Systemsposition

Kim reveals prior investment in AI chip startup SambaNova

SambaNova's next-generation computer architecture represents an early bet on AI-specific compute infrastructure before the generative AI wave.

Sourcery VC2026-07episode →
2ndmedium conviction
$LITELumentum Holdings

Kim focuses on optical interconnects as critical for next-gen data center design

Optical interconnects are essential to solve the power density and bandwidth constraints of AI data centers as copper reaches physical limits at higher compute densities.

Sourcery VC2026-07episode →
3rdmedium conviction
$AVGOBroadcom

Kim highlights Broadcom's chip-model co-design as new AI paradigm

Broadcom's co-design of custom XPU chips tightly integrated with foundation model architectures represents the new path for AI compute, moving beyond generic GPUs to model-specific silicon.

Sourcery VC2026-07episode →

most discussed · click a bar to filter

  • $LITE
  • $AVGO
  • $SPCX
  • $PLTR
  • $SAMBANOVA

recurring themes

  • Data Center Infrastructure1
  • Memory & Storage1
  • AI Infrastructure1
  • Networking & Optical Infrastructure1
  • AI Hardware & Chip Architecture1
5 total
$LITE
···
Lumentum Holdings
MEDtony kim·Sourcery VC·2 months ago·How BlackRock’s Tony Kim Thinks About Investing in the AI Revolution
Kim focuses on optical interconnects as critical for next-gen data center design
Optical interconnects are essential to solve the power density and bandwidth constraints of AI data centers as copper reaches physical limits at higher compute densities.
"a third with Lumenum and kind of uh bringing optics to kind of nextgen data center design"
1:51
$AVGO
···
Broadcom
MEDtony kim·Sourcery VC·2 months ago·How BlackRock’s Tony Kim Thinks About Investing in the AI Revolution
Kim highlights Broadcom's chip-model co-design as new AI paradigm
Broadcom's co-design of custom XPU chips tightly integrated with foundation model architectures represents the new path for AI compute, moving beyond generic GPUs to model-specific silicon.
"finally with uh with uh Broadcom and uh XPU and AI code design for chips... Obviously they did that with uh with a new uh jalapeno chip that recently came out. So um so those are…"
8:41
$SPCX
···
SpaceX
MEDtony kim·Sourcery VC·2 months ago·How BlackRock’s Tony Kim Thinks About Investing in the AI Revolution
Kim sees orbital data centers as 2030 convergence point for AI infrastructure
Moving compute to space via orbital data centers could unlock a radical rethink of terrestrial data center constraints around power, cooling, and land, with SpaceX's launch capability as key enabler.
"I'm excited for is uh the next forward step toward orbital data centers and because that that also engenders a radical change in in in data centers right if if you keep pushing on…"
62:54
$PLTR
···
Palantir Technologies
MEDtony kim·Sourcery VC·2 months ago·How BlackRock’s Tony Kim Thinks About Investing in the AI Revolution
Kim validates Palantir's ontology layer as blueprint for AI-native enterprise
Palantir's ontology/context layer that embodies cumulative enterprise knowledge is the exact architecture needed for AI agents to orchestrate proprietary data and token flows.
"I think Palunteer calls this ontology. Ontology. Exactly. Yeah. Yeah. Exactly. So you have this ontology layer, this context layer sitting on the the data the data foundation with…"
54:54
$SAMBANOVA
SambaNova Systems
HIGHtony kim·Sourcery VC·2 months ago·How BlackRock’s Tony Kim Thinks About Investing in the AI Revolution· position
Kim reveals prior investment in AI chip startup SambaNova
SambaNova's next-generation computer architecture represents an early bet on AI-specific compute infrastructure before the generative AI wave.
"One of my companies I was involved with Sabanova. Lipu was supposed to be one of the speakers."
2:19
9
Data Center Infrastructuretailwind
Data center redesign from kilometers to millimeters drives $10T capex cycle
AI's need for compute density forces data centers to shrink data movement from kilometers to millimeters, triggering a full-stack redesign across power, cooling, networking, and rack architecture that will consume $10T+ over five years.
9
Memory & Storagetailwind
RAM apocalypse looms as memory intensity surpasses compute in AI architectures
Model architectures are shifting from compute-heavy to memory-heavy (mirroring human brain), creating structural memory shortage as 3-4 year fab lead times clash with exponential demand growth for HBM, DRAM, and custom memory solutions.
9
AI Infrastructuretailwind
Compute primacy drives historic market cap shift from software to hardware
The AI era has flipped tech market leadership from software-centric (10T) to compute/hardware-centric (30T+), as insatiable model demand makes compute the new value anchor and forces complete data center rebuild.
8
Networking & Optical Infrastructuretailwind
Copper-to-light transition essential for intra-rack and intra-chip AI connectivity
As AI compute density increases, electrical interconnects hit physics limits; optical I/O and co-packaged optics become mandatory to move data at millimeter scale with acceptable power and heat.
8
AI Hardware & Chip Architecturetailwind
Model-chip co-design emerges as new moat for foundation labs and custom silicon
Leading AI labs are tightly co-designing model architectures with custom silicon (XPU/ASICs), creating a feedback loop where model specs inform chip design and chip capabilities inform model architecture, disadvantaging generic merchant silicon.
8
Energy & Power Generationtailwind
800V architecture and solid-state transformers to unlock data center power efficiency
New power architectures eliminating voltage step-down losses, combined with behind-the-meter generation (SMRs, fusion), are necessary to support gigawatt-scale AI clusters and represent a parallel infrastructure investment cycle.
8
Semiconductorstailwind
Semiconductor consolidation creates duopolistic pricing power across categories
Decades of VC neglect caused chip company consolidation into few survivors per category, creating non-commodity pricing power with highest sector margins; AI demand now exposes this structural scarcity.
8
Robotics & Physical AItailwind
Chinese robotics manufacturing wave enables low-cost embodiment for Western AI brains
China's 130+ robotics companies and 30-40 near-term IPOs reflect manufacturing scale advantages; the winning paradigm may mix Chinese hardware bodies with Western foundation model brains, accelerating commercial humanoid deployment.