Last Week’s AI News #15

Last week was full of holiday cheer, laughter, and time with family during this magical Christmas season. Yet even amid the festivities, AI didn’t take a break, and neither did we. From record-breaking app revenues to groundbreaking AI models, here’s a snapshot of the most important developments in the AI world.

Here’s a brief overview:

• ChatGPT Hits $3 Billion in Global Mobile Spending

• How China built its ‘Manhattan Project’ to rival the West in AI chips

• Mozilla’s new CEO says AI is coming to Firefox, but will remain a choice

• Meta Launches SAM Audio for AI-Powered Sound Editing

• Anthropic Updates Skills, Making AI Agent Training an Open Standard

• Nvidia licenses Groq tech in $20B deal

• Chinese AI Startup Z.ai Launches GLM-4.7, Topping Coding Benchmarks

• Everything else that happened in AI last week

CHATGPT HITS $3 BILLION IN GLOBAL MOBILE SPENDING

ChatGPT has reached a major milestone, generating $3 billion in global consumer spending on mobile devices since its launch in May 2023. Most of this growth occurred in 2025, with spending soaring to $2.48 billion, a 408% increase compared to 2024. The app achieved this milestone faster than other top apps, including TikTok, Disney+, and HBO Max. Competitor AI apps like xAI’s Grok are seeing similar revenue trajectories, signaling a booming market for monetized AI tools.

Why it matters to business:
Rapid consumer adoption and high spending highlight AI apps as a lucrative market. Companies can explore subscription models, in-app purchases, and future ad-based monetization, offering significant opportunities for growth and competitive advantage.

HOW CHINA BUILT ITS ‘MANHATTAN PROJECT’ TO RIVAL THE WEST IN AI CHIPS

In a high-security Shenzhen lab, China has built a prototype machine capable of producing cutting-edge semiconductor chips, challenging Western dominance. Developed with former ASML engineers, the prototype successfully generates extreme ultraviolet light, though working chips are expected by 2028–2030. The project, likened to a “Manhattan Project,” is central to China’s drive for semiconductor self-sufficiency and involves Huawei coordinating thousands of engineers nationwide.

Why it matters to business:
China’s accelerated chip development signals potential shifts in global supply chains and tech competitiveness. Companies in semiconductors, AI, and electronics should anticipate increased competition, alternative sourcing options, and possible geopolitical impacts on technology access.

MOZILLA’S NEW CEO SAYS AI IS COMING TO FIREFOX, BUT WILL REMAIN A CHOICE

Mozilla has appointed Anthony Enzor-DeMeo as its new CEO as the company seeks to adapt in a rapidly evolving browser market. With AI reshaping how users interact with the web, Mozilla plans to add optional AI features to Firefox while diversifying its revenue beyond search. The company is also expanding Firefox into a broader ecosystem, including email, VPN, and AI-powered website creation for small businesses.

Why it matters to business:
Mozilla’s strategy reflects the broader trend of AI integration in browsers, signaling opportunities for companies in software, AI, and digital services. Optional AI features allow Mozilla to retain loyal users while exploring new revenue streams, highlighting the importance of balancing innovation with user trust.

META LAUNCHES SAM AUDIO FOR AI-POWERED SOUND EDITING

Meta has introduced SAM Audio, an AI model that separates and edits sounds using text, visual, and temporal prompts. Users can isolate instruments, remove background noise, or edit audio more efficiently than with traditional tools. SAM Audio is positioned as the first unified multimodal model for audio segmentation, available on Meta’s Segment Anything Playground and for download.

Why it matters to business:
SAM Audio demonstrates how AI can streamline creative workflows, reduce manual audio editing, and expand opportunities in music, film, accessibility, and content production. Companies can leverage AI-driven tools to cut costs, improve efficiency, and develop new services for creators.

ANTHROPIC UPDATES SKILLS, MAKING AI AGENT TRAINING AN OPEN STANDARD

Anthropic has updated its Skills feature for Claude, allowing companies to teach AI agents specific work-related tasks using files and folders. The update makes Agent Skills an open standard, enabling easier sharing and deployment, while also adding centralized management for enterprises. A new partner-built skills directory includes tools from Atlassian, Canva, Notion, Figma, Stripe, and others, expanding capabilities for AI agents in tasks like marketing, project management, and automation.

Why it matters to business:
Open standards for AI agent skills simplify enterprise adoption and collaboration, enabling companies to customize AI for workflows, increase efficiency, and integrate AI across tools and platforms. This can accelerate automation, reduce operational costs, and create competitive advantages in digital workspaces.

NVIDIA LICENSES GROQ TECH IN $20B DEAL

Nvidia has struck a $20B licensing deal with AI chip startup Groq, acquiring rights to Groq’s LPU chips known for high-speed, energy-efficient AI performance. Groq CEO Jonathan Ross and President Sunny Madra will join Nvidia to help integrate and scale the technology, while Groq continues independently under its CFO. This deal is Nvidia’s largest ever and brings back the creator of Google’s TPU chips to compete in the AI silicon race.

Why it matters to business:
By securing Groq’s LPU technology and talent, Nvidia strengthens its position in AI hardware amid growing competition from Google and Amazon custom silicon. Companies and investors in AI chips should monitor how this strategic partnership could shift market dynamics and accelerate AI adoption.

CHINESE AI STARTUP Z.AI LAUNCHES GLM-4.7, TOPPING CODING BENCHMARKS

Chinese AI startup Z.ai has released GLM-4.7, a coding-focused model that achieves 73.8% on the SWE-bench, surpassing open-source rivals and matching top Western AI models. The model is open-sourced on Hugging Face and compatible with coding agents like Claude Code. The launch comes just before Z.ai’s expected $300M Hong Kong IPO, signaling China’s growing influence in competitive AI model development.

Why it matters to business:
GLM-4.7 highlights China’s ability to rapidly develop competitive AI tools and attract funding, suggesting that companies and investors should closely monitor Chinese AI advancements, especially in open-source and coding-focused solutions, which could reshape global AI competition.

EVERYTHING ELSE THAT HAPPENED IN AI LAST WEEK

• Honeycomb examines how accurate AI Agents really need to be. Read the blog to learn why speed, iteration, and self-correction matter more than perfection.

• AI evaluation firm METR posted a new analysis of Claude Opus 4.5, finding it capable of tasks requiring nearly 5 hours of work, the longest duration of any model to date.

• Cursor announced the acquisition of code review platform Graphite, with plans to integrate the tool into its AI-powered code editor.

• OpenAI CEO Sam Altman posted a job listing for a ‘Head of Preparedness’ to plan for and secure increasingly advanced models, including “systems that can self-improve”.

• Alibaba-backed MiniMax released M2.1, a model with powerful capabilities across a variety of programming languages and for mobile and web app development.

• Poetiq published an analysis of its system on the ARC-AGI-2 benchmark running GPT 5.2 X-High, surpassing 70% and scoring the highest of any model by around 15%.

The AI world never sleeps, and neither should your insights! From record-breaking apps to cutting-edge models and game-changing deals, staying ahead means staying informed. As we step into the New Year, may it be filled with innovation, inspiration, and a touch of AI magic. ✨

Don’t miss a single breakthrough, subscribe to our newsletter and get the freshest AI news and trends delivered straight to your inbox every week!

Don't miss news from us

Or join AI Mastery Program for business leaders >