TickerTain
TickerTain
NewsroomShortsPortfolioConvergence
NewsroomShortsPortfolioConvergence
←
▶ 93:00 · $AMD · Lisa Su details AMD's heterogeneous compute strategy and Anthropic partnership at Advancing AI
episode briefing
TBPN

AMD CEO Lisa Su Live on TBPN

2026-07-23 · 17 company · 17 thematic
sentiment
15 bull0 bear2 neu
speakers
lisa su

CEO of Advanced Micro Devices since 2014; led AMD's transformation into AI and high-performance computing leader with MI300/MI355/MI450 accelerator roadmap.

anne

Founder of AMP venture firm and infrastructure arm building 2GW US AI compute capacity; advocates 'output maxing' efficiency; investing in nuclear (Maver) to solve energy bottleneck for 2030 AI scaling.

seth cohen

Co-founder of Maver (AI infrastructure/nuclear); former DOE Chief Counsel for Nuclear Policy and NRC conservator; led nuclear waste narrative shift resulting in 27 governors competing for fuel cycle campuses.

oliver cameron

Founder of Odyssey building foundational world models; previously at Cruise working on self-driving cars; raised $310M for world model development across robotics, gaming, and autonomy.

muhammad

Founder of Ideogram building AI foundation for enterprise design and content; previously started Google's Imagen; released open-weight frontier image model for on-prem enterprise deployment.

aj

Founder of AMP, a venture firm (AMP Foundry) and infrastructure platform (AMP Grid) building ~2 GW of US AI compute capacity; investor in Maver and Periodic Labs; focused on energy, compute utilization, and frontier AI applications.

quote
“We fundamentally believe that there is no one-size-fits-all. Like the world is a heterogeneous world... The portfolio that we have, CPUs, GPUs, we acquired Xilinx so we brought the physical AI component in there... allows you to truly pick…”
—lisa su
now playing · $AMD
$ANTHROPICbullish· highlisa su
Anthropic deploys 2 GW of AMD MI450s in $5B deal and uses Claude to accelerate hardware bring-up
Anthropic's partnership with AMD for up to 2 GW of MI450 capacity signals massive compute demand; using its own models to optimize hardware bring-up (MI355) demonstrates a powerfu…
$CBRS···bullish· mediumjohn coogan
AMD partners with Cerebras on disaggregated inference for agentic AI at massive scale
Cerebras' wafer-scale engine paired with AMD's Instinct GPUs creates a disaggregated inference pipeline optimized for each phase, addressing the latency and throughput demands of…
$ODYSSEYbullish· higholiver cameron
Odyssey builds foundational world models at GPT-2 stage, targeting robotics, gaming, and autonomy with broad data pretraining
World models pretrained on diverse internet-scale data (not narrow driving/robotics data) then fine-tuned for specific tasks will outperform specialized vision-language-action mod…
$IDEOGRAMbullish· highmuhammad
Ideogram releases frontier open-weights image model for enterprise design, enabling on-prem deployment with data sovereignty
Ideogram's compact open-weights model at the quality frontier allows enterprises to self-host for IP protection and cost control, addressing the consistency and data sovereignty b…
$BLACK-FOREST-LABSbullish· mediumaj
Black Forest Labs trains Flux 3 video model on AMP Grid with 12 researcher co-founders
A 12-founder researcher team building frontier video models on dedicated AI infrastructure (AMP Grid) represents a new model for capital-efficient, talent-dense AI labs competing…
$AMP-GRIDbullish· highaj
AMP Grid building 2 GW of AI infrastructure, procuring 2030 capacity, investing in nuclear to solve energy bottleneck
AMP is deploying 2 GW of US AI compute capacity with a 2030 energy procurement horizon, spinning out its infrastructure arm, and investing in nuclear (Maver) to overcome the power…
$AMPbullish· high· posanne
AMP spinning out 2GW infrastructure arm, procuring nuclear energy for 2030 AI capacity
US data centers utilize only 15% of paid flops (85% waste = national security crisis); AMP is building 2GW capacity and investing in nuclear (Maver) to solve the energy bottleneck…
$ANTHROPICbullish· highaj
AMP founder reveals Anthropic at $40B+ run rate with non-circular revenue from coding agents
Anthropic has achieved over $40B annualized revenue driven by real enterprise adoption of coding agents, not circular cloud spending, validating the AI application layer business…
$PERIODIC-LABSbullish· highaj
Periodic Labs uses AI-driven RL loop with robotic synthesis and X-ray verification to discover room-temperature superconductors
Closing the loop between AI materials prediction, robotic synthesis, and physics-based verification (X-ray diffraction) enables rapid RL-driven discovery of superconductors, demon…
$MAVERbullish· highseth cohen
Maver reframes nuclear waste as $4x Saudi oil energy asset, secures 27 governor applications for fuel cycle campuses
Commercial spent fuel retains 97% of its energy potential; Maver's policy work flipped the narrative from liability to asset, triggering state-level competition to host full fuel…
$MAVERbullish· high· posseth cohen
Maver reframes nuclear waste as $4x Saudi oil reserves energy asset, 27 governors competing for sites
100,000 tons of spent fuel retains 97% energy content (4x Saudi proven reserves); narrative shift turned Yucca Mountain liability into asset with 27 governors applying for nuclear…
$GOOGL···bullish· mediumjohn coogan
Google Cloud hits 82% YoY growth as TPU sales become meaningful revenue line
Google Cloud's 82% growth and emerging TPU system sales indicate successful monetization of internal AI infrastructure; search revenue remains resilient at 17-19% growth despite L…
$OPENAIbullish· highjohn coogan
OpenAI raises 2030 compute spend target to $750B and breaks ground on $20B Georgia data center
OpenAI's compute budget increase from $600B to $750B through 2030 and the $20B Project Camila data center signal sustained exponential demand for AI infrastructure driven by codin…
$DEEPSEEKbullish· highjohn coogan
DeepSeek runs 85% inference margins on 20K H100 equivalents, achieves $1B API revenue with 10-month GPU payback
DeepSeek's disclosed unit economics — 85% inference margins, 10-month GPU payback, and $1B API revenue — prove frontier model inference can be highly profitable even with constrai…
$COGNITIONneutral· lowjohn coogan
Cognition acquires Poke team to accelerate AI agent interface development
Cognition's acquisition of the Poke consumer AI hardware team suggests a push toward novel human-agent interaction paradigms beyond chat interfaces.
$AMD···bullish· high· poslisa su
Lisa Su details AMD's heterogeneous compute strategy and Anthropic partnership at Advancing AI
AMD's portfolio of CPU, GPU, FPGA and adaptive compute enables right-sized workloads; deep partnerships with Anthropic and OpenAI validate the MI300/MI355/MI450 roadmap and Helios…
$NVDA···neutral· highlisa su
Lisa Su rejects winner-take-all GPU narrative, advocates heterogeneous compute for AI workloads
The AI compute market will not consolidate around a single GPU architecture; diverse workloads require CPUs, GPUs, FPGAs, and specialized accelerators working together, favoring A…