Summary

Apple refreshed the Mac Studio with M5 Max and M5 Ultra and gave the Mac mini M6 silicon, at higher prices. OpenAI's Jalapeño chip beat Nvidia on efficiency, Perplexity went fully local, WhatsApp toughened logins, and Uber livestreamed teen rides. Links Apple updates the Mac Studio with M5 Max and M5 Ultra, with up to 4.3x faster AI performance, faster gr…

Key takeaways

• OpenAI's Jalapeno ASIC (built with Broadcom) outperformed Nvidia's GB200/GB300 superchips by 1.5–1.9x on energy efficiency and 1.7–3.6x on latency for inference workloads across major frontier models, signaling a credible in-house alternative to GPU dependency.
• Perplexity's Portable Computer brings a full agentic AI stack—model, inference engine, tools, and security sandbox—entirely on-device with zero token costs, reframing local hardware as an economically superior substrate for always-on AI agents.
• Apple's M6 chip, its first on a 2nm process, delivers up to 40% CPU uplift and 2x GPU performance over M4, while the M5 Ultra supports 512GB unified memory—raising the ceiling for on-device AI and professional compute at incrementally higher price points.

Why this matters

The convergence of on-device AI (Apple silicon, Perplexity Portable Computer) and custom inference ASICs (OpenAI Jalapeno) signals a structural shift in enterprise AI economics—reducing cloud token costs, improving data governance, and challenging Nvidia's dominance in the inference market.

Entities

Related reports

• Nvidia On The March

  Tech Brew Ride Home

• Anthropic revealed Claude broke into three real organizations during cybersecurity tests it thought were simulations. Apple slid 10% as component shortages hit its outlook, Amazon jumped on accelerating AWS growth and a $220B capex plan, and Aschenbrenner's fund cratered 67%.

  Tech Brew Ride Home

• Anthropic & OpenAI Have Changed What Moving Fast Means | Krishna M, Elevation Capital

  The Neon Show

Show notes

Apple refreshed the Mac Studio with M5 Max and M5 Ultra and gave the Mac mini M6 silicon, at higher prices. OpenAI's Jalapeño chip beat Nvidia on efficiency, Perplexity went fully local, WhatsApp toughened logins, and Uber livestreamed teen rides. Links Apple updates the Mac Studio with M5 Max and M5 Ultra, with up to 4.3x faster AI performance, faster graphics, and up to 512GB of unified memory for $2,499+ (Apple Newsroom) Apple unveils a Mac mini with M6 and M5 Pro, with up to 4x faster AI per

Themes

• openai