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 Hardware & Chip Architecture · Cross-layer co-design turns incremental gains into hundredfold improvements
now playing · AI Hardware & Chip Architecture
AI Economics & Business Modelstailwindscore 10/10dylan patel
AI inference grows into a market larger than oil
Inference could absorb multiple percentage points of global GDP as token consumption becomes embedded across economic activity. Continued software and model optimization is simultaneously c…
AI Hardware & Chip Architecturetailwindscore 10/10dylan patel
Cross-layer co-design turns incremental gains into hundredfold improvements
Optimizing hardware, systems software and model architecture independently can produce incremental gains, but designing them together can turn several twofold improvements into a hundredfol…
AI Infrastructuretailwindscore 10/10dylan patel
Model capability expands demand faster than compute supply grows
Twenty gigawatts of compute should come online this year and more than thirty next year, yet model improvements are expanding the pool of economically valuable tasks even faster. The comput…
Memory & Storagetailwindscore 9/10dylan patel
Memory bandwidth remains AI's slowest-moving architectural bottleneck
DRAM and NAND cell architectures have seen no fundamental breakthrough for decades, leaving memory capacity and bandwidth behind compute growth. Stacking memory directly onto logic could un…
Neoclouds & Cloud Computingmixedscore 9/10dylan patel
AI workloads erase hyperscalers' traditional cloud advantages
GPU customers rent entire clusters under long contracts, reducing the value of tenant isolation and CPU-era cloud architecture. Neoclouds can win through superior GPU performance and faster…
Space-Based Computemixedscore 8/10dylan patel
Orbital compute remains distant but could dominate incremental capacity
Space data centers should contribute less than 1% of compute by 2030, making near-term expectations excessive. By 2040, terrestrial power constraints and terawatt-scale inference demand cou…