TickerTain
$SPCX····$NVDA····$MSFT····$GOOGL····$ANTHROPIC·$TSLA····$CRWV····$NBIS····$HIMS····$TSM····$ASML····$FIGURE·$LYFT····$UBER····$SNOW····$META····$ELEVENLABS·$CURSOR·$COGNITION·$OPENAI·$AAPL····$RKLB····$XAI·$AMZN····$AMD····$000660.KS····$INTC····
← newsroom
AI Geopolitics & Export Controls · Chinese open models rely on US frontier distillation; US labs won't open-source competitive weights
now playing · AI Geopolitics & Export Controls
Frontier AI Modelstailwindscore 9/10clay bavor
Unbounded demand for frontier intelligence; open models serve commoditized tasks via distillation
Frontier models will retain value for high-stakes complex domains (coding, science, legal) where demand for intelligence is effectively unbounded. Open weights models will handle commoditiz…
AI Infrastructuretailwindscore 9/10clay bavor
Token costs face structural floor from GPU scarcity; engineer token budgets heading to 20% of salary
Reasoning models (o1) dramatically increase inference compute per token, while GPU supply constraints (power, capacity) create a hard floor on token pricing. Bavor predicts steady-state tok…
AI Coding Agentstailwindscore 9/10clay bavor
AI-native engineers 3-20x more productive; internal agents run company operations
Engineers using Claude Code/Codex achieve 3-20x feature shipping velocity. Sierra built 'Pine Cone' — an internal agent with MCP gateway accessing all company systems — that runs engineerin…
AI Agentstailwindscore 9/10clay bavor
Effectively unbounded enterprise demand for AI agents; forward-deployed model accelerates adoption
Enterprise buyers are in market simultaneously due to board pressure, creating unprecedented pull. Sierra's Palantir-inspired forward-deployed engineering model (deeply embedded with design…
AI Geopolitics & Export Controlsmixedscore 7/10clay bavor
Chinese open models rely on US frontier distillation; US labs won't open-source competitive weights
The performance gap between US and Chinese open models stems from Chinese companies' willingness to do scaled distillation of US frontier models. US labs and hyperscalers have no incentive…
Cybersecuritytailwindscore 7/10clay bavor
Offensive AI capabilities ratcheted up five notches; cybersecurity a 'pretty good bet'
AI-driven offensive cyber capabilities have dramatically increased, making defense more critical than ever. However, the defensive solution may ultimately be frontier models themselves (Myt…
AI Infrastructureheadwindscore 6/10clay bavor
On-device AI cannot alleviate frontier compute constraints; thermal limits bind
While on-device AI will improve consumer applications and may enable 'home appliance' compute boxes, thermal and power limits on phones prevent meaningful contribution to frontier training…