Skip to content
podcastTechnologyAIApple

Apple Introduces AirTag 2 - DTH

Apple launches AirTag 2 with enhanced UWB while Microsoft debuts the Maya 200 AI chip. Plus, the EU Commission opens a probe into X and designates WhatsApp a "Very Large Online Platform" amid growing regulatory scrutiny.

Table of Contents

Apple has officially launched the second-generation AirTag, featuring enhanced Ultra Wideband capabilities, while Microsoft unveiled its newest custom silicon, the Maya 200, designed to power next-generation AI models. These major hardware announcements coincided with significant regulatory moves in Europe, where the European Commission intensified scrutiny on both X and WhatsApp under the Digital Services Act.

Key Points

  • Apple AirTag 2: Features a second-generation UWB chip with 50% better precision finding and extended range, maintaining the $29 price point.
  • Microsoft Silicon: The new Maya 200 chip is built on a 3nm process, claiming 3x the performance of Amazon's Trainium.
  • EU Regulation: The European Commission opened a probe into X regarding AI-generated deepfakes and designated WhatsApp a "Very Large Online Platform."
  • NVIDIA Weather AI: Three new Earth-2 models were released, with medium-range forecasting reportedly beating Google’s DeepMind.
  • Federal AI Usage: The U.S. Department of Transportation is reportedly testing Google Gemini to draft federal regulations to accelerate timelines.

Hardware and Silicon Advancements

Apple’s introduction of the AirTag 2 marks a significant iterative update to its popular tracking device. The new model incorporates a second-generation Ultra Wideband (UWB) chip that Apple claims increases precision finding by up to 50% while simultaneously extending Bluetooth range. The device also includes a louder speaker to facilitate easier location and new privacy protocols designed to mitigate unwanted tracking.

Notably, the AirTag 2 supports wrist-based finding on newer Apple Watch models. Despite the hardware upgrades, Apple has maintained the pricing structure at $29 for a single unit and $99 for a four-pack. Pre-orders have opened immediately, with shipping scheduled for later this week.

In the enterprise sector, Microsoft unveiled the Maya 200, an AI accelerator built on TSMC’s 3-nanometer process. The chip is designed to run GPT-5.2 and other large language models.

Microsoft reports that the Maya 200 delivers more than 1 billion transistors and offers 30% better performance per dollar than the company's current offerings.

Microsoft claims the chip achieves three times the FP4 performance of Amazon’s latest Trainium chip and posts stronger FP8 numbers than Google’s newest TPU. The hardware is currently rolling out to Microsoft data centers.

European Regulatory Crackdown

The European Commission (EC) has taken aggressive steps to enforce the Digital Services Act (DSA). The Commission opened a formal investigation into X (formerly Twitter) regarding its Grok AI tool. Regulators are probing the platform's ability—or lack thereof—to prevent the generation of sexualized deepfakes of real individuals. If X fails to implement necessary changes, the EC stated it may impose interim measures and issue fines of up to 6% of the company's global revenue.

Simultaneously, the EC designated Meta’s WhatsApp as a "Very Large Online Platform" (VLOP). This designation makes WhatsApp the first messaging service subject to the EU’s strictest online governance rules. The classification applies specifically to WhatsApp's public channels.

Implications for Meta:

  • WhatsApp must report on how it mitigates systemic risks to children and democratic processes.
  • The first compliance report is due within four months.
  • Non-compliance risks investigations and fines comparable to those threatened against X.

AI Innovation and Public Sector Adoption

NVIDIA continues to expand its footprint in specialized AI applications with the release of three new Earth-2 weather models: medium-range, nowcasting, and global data assimilation. NVIDIA claims its medium-range model outperforms Google’s DeepMind Gencast across more than 70 variables, offering forecasts up to 15 days out. The "nowcasting" model utilizes geostationary satellite data to predict storm impacts within a zero to six-hour window.

In the public sector, a report from ProPublica indicates that the U.S. administration is utilizing Google’s Gemini to draft federal transportation regulations. The Department of Transportation (DOT) is reportedly serving as a test bed for this initiative.

DOT's general counsel reportedly described the strategy as prioritizing speed and volume, noting that regulations only need to be "good enough" as the agency attempts to compress drafting timelines from months to roughly 30 days.

Market Outlook

Looking ahead, Apple’s software strategy appears to be pivoting toward generative AI. According to Bloomberg’s Mark Gurman, Apple plans to preview a Gemini-powered Siri update in February via an iOS 26.4 beta. A complete overhaul, codenamed "Campos," is expected to debut at WWDC in June alongside iOS 27. This update will reportedly feature a rebuilt conversational UI running on a new Apple Foundation model.

Meanwhile, the transition of TikTok to U.S. ownership has faced early technical hurdles. The platform experienced outages over the weekend attributed to a power failure at a U.S. data center. Users also noted new terms of service allowing for more granular data collection, including precise location tracking and the storage of AI interactions.

Latest

The creator of Clawd: "I ship code I don't read"

The creator of Clawd: "I ship code I don't read"

Peter Steinberger, creator of Clawd, merges 600 commits daily using a fleet of AI agents. In this deep dive, discover how he challenges engineering norms by shipping code he doesn't read, treating PRs as "Prompt Requests," and replacing manual review with autonomous loops.

Members Public
The Clawdbot Craze | The Brainstorm EP 117

The Clawdbot Craze | The Brainstorm EP 117

The AI landscape is shifting to autonomous agents, led by the viral "Claudebot." As developers unlock persistent memory, OpenAI refines ad models, and Tesla hits new milestones, software intelligence meets real-world utility. Tune into The Brainstorm EP 117.

Members Public