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·
←

ben beharon

T3 · host / generalist

CEO and principal analyst of Creative Strategies, brought on as the bull in a Nvidia bull/bear debate.

3 calls·3 names·100% bull·last heard 10 months ago·The Information
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
$NVDANvidia

Beharon argues installed base and CUDA moat compound as models scale to 20T parameters

Nvidia's exponentially growing installed base of GPUs and CUDA software ecosystem create a self-reinforcing advantage for training ever-larger models (10T to 20T parameters), and Nvidia will soon fragment its product line to offer inference-specific chips that leverage the same developer lock-in.

The Information2025-11episode →
2ndmedium conviction
$AMDAdvanced Micro Devices

Beharon says AMD will be fully competitive with competitive racks by 2027

AMD now has competitive silicon and will deliver competitive rack-scale systems by 2027, becoming fully competitive with Nvidia in the AI accelerator market.

The Information2025-11episode →
3rdmedium conviction
$GOOGLAlphabet

Beharon argues Google TPU is better than Nvidia for large AI workload classes

Google's TPU architecture makes a compelling case as a superior chip for large classes of AI workloads, representing credible internal competition to Nvidia from a hyperscaler customer.

The Information2025-11episode →

most discussed · click a bar to filter

  • $AMD
  • $NVDA
  • $GOOGL

recurring themes

  • AI Hardware & Chip Architecture1
  • AI Infrastructure1
  • AI Economics & Business Models1
3 total
$AMD
···
Advanced Micro Devices
MEDben beharon·The Information·10 months ago·Can Jensen Huang Lead Nvidia Through its Biggest Challenge Yet?
Beharon says AMD will be fully competitive with competitive racks by 2027
AMD now has competitive silicon and will deliver competitive rack-scale systems by 2027, becoming fully competitive with Nvidia in the AI accelerator market.
"AMD just had an event last week where they outlined their road map. They have now competitive silicon. They're going to have competitive racks by 2027. They're going to be fully c…"
1:01
$NVDA
···
Nvidia
HIGHben beharon·The Information·10 months ago·Can Jensen Huang Lead Nvidia Through its Biggest Challenge Yet?
Beharon argues installed base and CUDA moat compound as models scale to 20T parameters
Nvidia's exponentially growing installed base of GPUs and CUDA software ecosystem create a self-reinforcing advantage for training ever-larger models (10T to 20T parameters), and Nvidia will soon fragment its product line to offer inference-specific chips that leverage the same developer lock-in.
"Nvidia has the single largest installed base of GPUs both at the edge and in the cloud and that number is increasing by an exponential. So, as much as you know, I agree TPUs are g…"
1:36
$GOOGL
···
Alphabet
MEDben beharon·The Information·10 months ago·Can Jensen Huang Lead Nvidia Through its Biggest Challenge Yet?
Beharon argues Google TPU is better than Nvidia for large AI workload classes
Google's TPU architecture makes a compelling case as a superior chip for large classes of AI workloads, representing credible internal competition to Nvidia from a hyperscaler customer.
"On top of that and probably more crucial is Nvidia is competing with all its customers. All the hyperscalers are designing their own silicon. If you look at Google TPU, you can ma…"
1:26
8
AI Hardware & Chip Architectureheadwind
Hyperscaler custom silicon and AMD threaten Nvidia's GPU dominance
Google TPUs are already superior for large AI workloads, AMD will be fully competitive with rack-scale systems by 2027, and all major hyperscalers are designing their own silicon, structurally eroding Nvidia's moat.
7
AI Infrastructuremixed
Nvidia will fragment product line to capture inference market opportunity
Nvidia will leverage its CUDA ecosystem and developer lock-in to launch inference-specific chips, but the inference market is early and workloads undefined, making it anyone's game until silicon footprints clarify.
7
AI Economics & Business Modelstailwind
Hyperscalers can sustain AI capex far longer than frontier labs
Microsoft, Amazon, and Google spend under 30% of cloud revenue on capex and have diversified profit streams to fund decade-long AI buildouts, unlike cash-burning frontier labs such as OpenAI that lack visible funding for trillion-dollar plans.