Westworld For Real

What It Covers

NVIDIA's GTC 2026 conference dominates headlines with Jensen Huang projecting $1 trillion in chip orders through 2027, while Samsung quietly kills its tri-fold phone, OpenAI pivots toward enterprise and coding, Amazon launches one-hour delivery, and Disney deploys its first AI-trained animated robot character at theme parks.

Key Questions Answered

• •NVIDIA Revenue Trajectory: Jensen Huang publicly projected $1 trillion in cumulative orders for Blackwell and Vera Rubin chips through 2027, doubling the $500 billion figure from just one year prior. Rubin chips outperform Blackwell by 3.5x on training and 5x on inference, reaching 50 petaflops, with production ramping in the second half of 2026.
• •AI Graphics Backlash: NVIDIA's DLSS 5 announcement, which uses real-time neural rendering to add photorealistic lighting and materials at up to 4K, triggered immediate negative reactions across Reddit and social media. Gamers labeled it "AI slop," forcing NVIDIA to clarify within hours that developers retain full artistic control over the technology's visual effects.
• •OpenAI Strategy Pivot: OpenAI is deprioritizing broad product bets like Sora video generation and hardware devices to refocus on coding tools and enterprise customers. The shift is a direct response to Anthropic's dominance in business AI via Claude Code. Leadership acknowledged that scattered compute allocation and unclear strategic direction weakened execution throughout 2025.
• •Foldable Phone Market Signal: Samsung discontinued its $2,899 Galaxy Z Tri-Fold in South Korea and the US after just three months, never expanding beyond direct Samsung sales channels. The device's failure reinforces that foldable phones above $1,500 remain niche products, and the format's viability at scale likely depends on whether Apple enters the category.
• •Disney Robot Deployment Method: Disney trained its Olaf robot using 100,000 simultaneous virtual simulations on a single NVIDIA RTX 4090 GPU, completing training in two days via reinforcement learning. The simulation rewarded movement patterns that avoided joint overheating and noise. This simulation-first pipeline reduces robot development time from years to months, enabling Disney to scale characters across entire park lands.