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 Infrastructure · Gigawatt-scale AI data centers redefine infrastructure investing
now playing · AI Infrastructure
AI Infrastructuretailwindscore 9/10dylan patel
Gigawatt-scale AI data centers redefine infrastructure investing
Training and inference demand drives 1-10 GW campus builds costing $50-500B each, creating a new asset class where power, land, and permitting — not just GPUs — are the binding constraints.
AI Coding Agentstailwindscore 9/10dylan patel
Software development automation drives fastest AI revenue ramp ever
Anthropic grew from $1B to $7-8B ARR in months via code generation (Cursor, Copilot, Windsurf) — the $2T software engineer wage pool creates infinite demand for models that match senior eng…
Semiconductorstailwindscore 9/10dylan patel
Nvidia captures majority of AI gross profit via full-stack control
Nvidia's 75% GPU margins, NVLink/NVSwitch interconnect monopoly, and new demand-guarantee financing let it extract value across the stack while competitors (AMD, custom ASICs) fight for scr…
Grid & Power Infrastructuretailwindscore 9/10dylan patel
US power generation bottlenecked by 40-year supply chain atrophy
AI data center demand growth (doubling every 2 months) collides with dormant turbine, transformer, and skilled labor supply chains, creating multi-year lead times and forcing creative solut…
AI Geopolitics & Export Controlsriskscore 9/10dylan patel
US-China AI race hinges on Taiwan semiconductor choke point
China's insular supply chain builds slower but more resiliently; US leads in chip design but depends entirely on TSMC. A Taiwan blockade collapses US AI infrastructure while China could smu…
AI Economics & Business Modelsmixedscore 8/10dylan patel
Tokconomics reveals inference cost curves determine model serving strategy
Log-log scaling means 10x compute for next intelligence tier; firms must choose between serving smaller models at massive scale (low cost, high adoption) or larger models at low throughput…
Robotics & Physical AItailwindscore 8/10dylan patel
Embodiment data flywheel for robotics is in inning zero
Human-level dexterity requires tactile feedback loops (hand-in-mouth calibration) that no robot has; but basic pick-place-sort tasks in logistics/manufacturing are high-value near-term targ…
Networking & Optical Infrastructuretailwindscore 8/10dylan patel
Chip-to-chip optics at 1.8 TB/s create new semiconductor bottleneck
Blackwell's all-to-all rack connectivity at terabyte-per-second speeds pushes electrical signaling to physics limits, making co-packaged optics and optical circuit switching the next critic…
World Models & Simulationtailwindscore 8/10dylan patel
AI-driven physics simulation accelerates hard-tech R&D cycles
Replacing slow CFD/experimental loops with learned world models (molecules, chemical reactions, semiconductor processes) compresses iteration from months to hours, with Periodic Labs and TS…
Enterprise AI Adoptionheadwindscore 8/10dylan patel
AI-native COGS destroys traditional SaaS unit economics
High token inference costs (COGS) plus unchanged customer acquisition costs prevent new AI SaaS from reaching escape velocity; incumbents with distribution (Google, Microsoft) and low COGS…