AI: Weekly Summary. RTZ #948

Groq agreement for Nvidia’s AI Inference startup: from Nvidia $20 billion AI acquisition and hiring licensing deal with Groq, AI inference startup (no relation to Elon Musk’s Grok AI xAI)is just another in a string of similar multibillion-dollar deals this year. These include Microsoft/Inflection, Google/CharacterAI And windsurfing, Meta/scale AI and many more. The deal helps Nvidia consolidate its great software and hardware vs upcoming competition AMDIt is own best customersand a crowd of AI chip startups around the world. Groq founders and key employees expected to join Nvidia, while Groq continues as an independent company. In this way, it particularly resembles Meta’s $14 billion deal with Scale AI. Nvidia’s fJensen Huang, President and CEO has a good track record small and large acquisitions and partnerships to strengthen its formidable suite of AI infrastructure assets. Expect this trend to continue for Nvidia and its peers until 2026. More here.

Google continues its race with Gemini 3 Flash: Google continued its lead with Gemini 3releasing a smaller, more efficient Gemini 3 Flash with plans for broad, global deployment early next year. It’s supposed to deliver a decisive blow against OpenAI and others”, with a “faster and cheaper model in the Google ecosystem”. THE the model is optimized for “multimodal planning and reasoning” abilities, including the ability to watch videos, look at the pictureslisten to audio and read text, while transforming these inference queries into data and results for consumer users. The model also exploits Google’s vertical position AI infrastructure stack with its own TPU chips. This gives Google an advantage over competitors AI computing is further constrained for now through 2025. Next year, of course, will see a few more multi-gigawatt AI data centers come online, which could ease the computing shortage for everyone. competing LLM AI models. More here.

A tough race for AI devices in 2025: It was a eventful year for AI devices with not much to show so far. It’s a growing media consensus for now, despite optimism for 2026 and beyond. Next year’s big event, of course, will be ‘AI device portfolio‘ being launched by OpenAI next year in collaboration with Jony Ive, after the $6.5 billion merger earlier this year. There is also Apple’s AI Powers Siri Devices and platforms, as well as Amazon Alexa Plus. Not to mention similar offerings from Google. Another big area is AI Smart Glasses Since Meta, Google, Apple, Amazon and others. As well as the general growing category of AI wearable devices. Apple would plan cameras on its formidable Airpods platform. And of course a multitude of AI voice chatbots broadcast via hundreds of millions of smartphones around the world. There is therefore a lot to wait despite a slow start in 2025. More here.

AI Race rivals the real space race run by NASA: Great technology Mag 7 and OpenAi are already committed to investing more than $400 billion in 2025 alone, more than the 300 billion dollars forok in 2025 dollars for the USA and NASA to land two men on the moon in 1969 from 1958. It was a geopolitical question The “space race” between the United States and the USSR, a bit like the The “AI space race” between the United States and China Today. The difference being the absence of a own objective: land a human on the moon. AGI is a much more blurry full stop, and therefore a infinite game vs finite game of the real space race. And the United States landed a total of 12 men on the moon in 1972, and none for more than half a century. It is useful to consider that the United States spent more than $2 trillion in current dollars on thousands of NASA projects, manned and unmanned, in 67 years, from 1958 to 2025. This compares to 1.4 trillion dollars committed by OpenAI alone by 2030. With general awards are late THE investments. More here.

Many surprises in AI in 2025: THE NeuroIPS AI Conference 2025 in San Diego this month deals a whole look back at 2025 for the AI technology wave. Five big surprises from my point of view included Meta’s Hiring Dash AI to build beyond Llama, China’s leadership in open source AI, Google’s The return of GeminiOpenAI AI Glove and end of yearCode Red’and of course AI spending offsets GDP collapse due to prices. And those were just the big ones. Others include the Mad race across 50 states to secure power for the Gigawatt AI Data Centers through a welter of local and state regulations, including The federal government opposes it. And of course, the modest reduction in fears of AI catastrophe compared to AI FOMO will invest before demand. Not to mention the growing rush to fund AI data centers in space. More here And here.