PLUS: A post-transformer model, Amazon’s new chatbot, and xAI’s $20B round
Good morning, AI enthusiast.
Nvidia has just unveiled its next-generation ‘Vera Rubin’ AI chip platform, promising a massive leap in performance that will power the next wave of AI development.
The new architecture delivers a 5x jump in inference speed and includes a new CPU specifically designed for agentic reasoning. With this dramatic increase in power, how quickly will this new hardware accelerate the race toward more complex and autonomous AI systems?
In today’s AI recap:
- Nvidia’s new 5x faster Rubin AI chip
- A new ‘post-transformer’ AI architecture
- Amazon’s new Alexa web chatbot
- xAI secures massive $20B funding round
Nvidia’s New AI Chip
The Recap: At CES 2026, Nvidia unveiled its next-generation ‘Vera Rubin’ AI chip platform, the successor to Blackwell. The new architecture promises massive performance gains for developing and running next-generation AI.
Unpacked:
- The Rubin platform delivers a 5x increase in inference speed and a 3.5x boost in training performance, enabling more complex models to run faster and more efficiently.
- It’s more than just a GPU; the architecture includes a new Vera CPU specifically designed for agentic reasoning and tackles growing bottlenecks in memory and data transfer.
- Major cloud providers and AI labs are not waiting, with partners like AWS, OpenAI, and Anthropic already planning to integrate the new Rubin systems.
Bottom line: Nvidia’s relentless hardware development cycle continues to set the pace for the entire AI industry. This significant jump in power will directly enable the creation of larger and more capable AI models.
The Post-Transformer Era?
The Recap: A company named Pathway has published research on a new ‘post-transformer’ AI architecture designed to enable true memory, continual learning, and interpretability—tackling three of the biggest limitations in today’s LLMs.
Unpacked:
- Named ‘Dragon Hatchling’ (BDH), the new architecture is based on a biologically inspired network of locally interacting neuron particles, aiming to mimic the brain’s ability to generalize over time.
- Early results show that BDH rivals GPT-2 performance on language tasks at similar parameter counts, demonstrating its viability without sacrificing the power of attention mechanisms.
- The project is backed by Lukasz Kaiser, a co-inventor of Transformers, adding significant credibility to Pathway’s approach in moving beyond current model limitations.
Bottom line: This represents a fundamental shift in model design, moving away from the brute-force scaling of Transformers toward more efficient and brain-like architectures. If successful, this research could pave the way for AI systems that learn continuously from new data without constant retraining.
Amazon’s Alexa Challenges ChatGPT
The Recap: Amazon is taking the AI chatbot race to the browser, launching Alexa.com to give its Alexa+ assistant a web-based home. This move directly challenges existing players like ChatGPT by making Alexa available beyond its native smart speakers and devices.
Unpacked:
- The biggest differentiator is its deep integration with Amazon’s ecosystem, allowing you to start a task on the web and seamlessly continue it on an Echo speaker or Fire TV.
- Unlike pure chatbots, Alexa+ is positioned as an “agent” that takes real-world action, from managing family calendars and controlling smart home devices to ordering groceries directly from your chat.
- Amazon is aiming to make Alexa+ a central hub for family life, encouraging users to upload documents and schedules to help manage everything from kids’ soccer practice to vet appointments.
Bottom line: Amazon is leveraging its massive hardware and e-commerce ecosystem to carve out a unique space in the crowded chatbot market. The battle is shifting from pure conversational ability to which AI can most effectively manage your daily life and get things done.
xAI Secures $20B in Massive Funding Round
The Recap: Elon Musk’s AI venture, xAI, has officially closed a massive $20 billion funding round, catapulting its valuation to an estimated $230 billion.
Unpacked:
- This $20 billion Series E round represents one of the largest single investments into an AI company, providing significant capital for its ambitious projects.
- The new $230 billion valuation places xAI among the most valuable private technology companies in the world, positioning it as a major competitor.
- The funding signals a clear intent to accelerate development and directly challenge the industry’s established leaders in the fiercely competitive AI space.
Bottom line: This infusion of capital gives xAI the resources to aggressively pursue large-scale model training and recruit top talent. Expect increased competition and faster innovation across the AI landscape as investment in foundational models continues to grow.
The Shortlist
Boston Dynamics demoed its humanoid robot Atlas at CES, which will integrate Google DeepMind’s Gemini Robotics for advanced tasks in Hyundai factories starting in 2028.
Nvidia unveiled Alpamayo at CES 2026, a new family of open-source AI models designed to help autonomous vehicles reason through complex driving scenarios.
OpenAI signed a memorandum of understanding with the U.S. Department of Energy to accelerate scientific discovery by pairing frontier AI models with the DOE’s national laboratories and advanced computing resources.
OpenAI revealed in a new report that over 40 million people globally use ChatGPT for health information daily, with over 5% of all messages now related to healthcare topics.