AI News – February 8: Amazon Bets 50 Billion on OpenAI, Google Goes Shopping, and MRI Becomes Intelligent

The first week of February 2026 is explosive: rumors of a $50 billion Amazon investment in OpenAI shake Silicon Valley, while Google DeepMind responds with a se

If last week we focused on earnings and the GDP that isn't growing fast enough, the first week of February 2026 marks the return of big capital and aggressive strategy. While Wall Street is still digesting quarterly reports, Amazon seems ready for a move that could break the Microsoft-OpenAI duopoly, putting staggering figures on the table. In parallel, scientific research takes a leap forward: AI no longer just generates text or images, but reads the human brain with the "BrainIAC" model.

Here is the reasoned chronicle of a week where computing power meets venture capital.

1. The Financial Earthquake: Amazon Courts OpenAI with 50 Billion?

The loudest rumor of the week could redraw the power map in Silicon Valley.

🔍 What happened:

  • According to digests from LinkedIn (PA Media) and AI Weekly, there are ongoing negotiations for a monster $50 billion investment by Amazon in OpenAI.
  • In parallel, OpenAI has reached an operational milestone: its APIs now generate $1 billion in ARR (Annual Recurring Revenue), confirming massive enterprise adoption.
  • Context: This happens while rival xAI by Elon Musk navigates troubled waters, with reports of an internal crisis despite recent funding rounds.

💡 Why it matters: If confirmed, Amazon's investment would dilute Microsoft's exclusive influence over OpenAI. It means Sam Altman is diversifying his alliances to secure the necessary computing power (AWS data centers) for the race towards AGI.

🎯 Our take: We are witnessing the "cloud war." AI is the fuel, but data centers are the engine. Amazon doesn't want to stand by while Microsoft takes the entire generative AI market.

Sources: LinkedIn Digest, La Cassetta degli AI Trezzi Read also: Dynamic Pricing Algorithms: Strategic Implications

2. Google DeepMind: Shopping Spree for Innovation

While Amazon spends, Google acquires. Mountain View's strategy shifts from "building everything in-house" to "buying the best."

🔍 What happened:

  • Google DeepMind closed three strategic deals in rapid succession, acquiring or forming deep partnerships with:
    1. Common Sense Machines: For 3D model training and physical simulations.
    2. Hume AI: Specialized in "Empathic AI" and vocal emotion analysis.
    3. Sakana AI: A Japanese startup known for its lightweight, nature-inspired evolutionary models.
  • Concurrently, Gemini 3 Flash was released, a version optimized for speed and low cost, designed to compete with open-source models.

💡 Why it matters: Google is trying to fill specific gaps. By acquiring Hume AI, it aims to make its voice assistants not just intelligent, but "empathetic." With Sakana AI, it looks at the Asian market and energy efficiency.

🎯 Our take: Integrating emotional intelligence (Hume) into Google systems could be the real game changer for consumer adoption, making human-machine interaction less robotic and more natural.

Sources: Fladgate, AI-Weekly Read also: AI and Psychology of the Mind: Diagnosis and Algorithms

3. BrainIAC: AI Enters the MRI

We leave Silicon Valley and enter the hospital. This is perhaps the week's news with the greatest social impact.

🔍 What happened:

  • Mass General Brigham (one of the world's most prestigious hospital networks) presented BrainIAC, a foundation model specifically for interpreting Magnetic Resonance Imaging (MRI).
  • Unlike human radiologists who analyze the visual image, BrainIAC analyzes the raw magnetic signal data, managing to reconstruct sharper images with reduced scan times and detecting microscopic anomalies invisible to the human eye.
  • In Catalonia, the healthcare system announced the massive adoption of similar tools for primary screening.

💡 Why it matters: This is the shift from AI that "writes poems" to AI that "saves lives." Reducing MRI times means reducing waiting lists and costs, democratizing access to high-level diagnostics.

🎯 Our take: As we have often discussed, AI in medicine will not replace the doctor, but will give them sensory "superpowers." BrainIAC is the perfect example of high-value-added human-machine collaboration.

Sources: HumAI Blog, LinkedIn Weekly Digest

4. New Models: Logical Reasoning and Human Voices

Technology moves fast. Two releases this week raise the quality bar.

🔍 What happened:

  • Logical Intelligence unveiled Kona, a model defined as "AGI-like" for its pure logical reasoning capabilities (demonstrated by solving complex puzzles like Sudoku without specific training, but by reasoning about the rules).
  • ElevenLabs launched Version 3 (V3) of its Text-to-Speech engine. The novelty? The management of pauses, breaths, and emotional intonations is so realistic that it makes the synthetic voice indistinguishable from a human one in conversational contexts.

💡 Why it matters: Kona marks a step forward towards "System 2 Thinking" (slow, logical reasoning), overcoming the limits of LLMs that often "guess." ElevenLabs V3, on the other hand, poses new ethical challenges for security (vishing and voice fraud).

🎯 Our take: While logic improves reliability (fewer hallucinations), ultra-realistic voice increases security risks. Companies will soon have to adopt audio detection systems to protect themselves.

Sources: AI Weekly Digest, AI-Weekly Newsletter Read also: Cybersecurity and AI: Low-Cost Hacking and Automatic Defense

5. Regulation: The UK Investigates Grok

There's no week without regulatory intervention. This time it's Elon Musk's turn.

🔍 What happened:

  • The UK authority Ofcom has opened a formal investigation into Grok (the xAI AI integrated into X/Twitter). The accusation is of not having implemented sufficient "guardrails" (safety barriers) against the spread of misinformation and harmful content, especially generated by users.
  • The ICO (Information Commissioner’s Office) released a report on Agentic AI, warning that the use of autonomous agents acting on behalf of users requires new standards of legal responsibility.

💡 Why it matters: It shows that the era of "release and then fix" is over. Regulators are starting to target platforms not only for data privacy, but for the safety of content generated and spread by algorithms.

🎯 Our take: The governance of Autonomous Agents will be the hot topic of 2026. If an AI books a wrong flight or defames someone online, who is to blame? The owner or the developer?

Sources: Fladgate, CGSPAM

📊 What do these developments tell us?

The week of February 2-8, 2026 highlights three clear trends:

  1. Consolidation of Giants: With Amazon entering the game heavily, the space for small generalist players shrinks. It's a war between ecosystems (Microsoft/OpenAI vs Google vs Amazon/OpenAI?).
  2. Verticalization: Generic AI makes way for specific AI (BrainIAC for medicine, Kona for logic). This is where real added value is created now.
  3. Security vs. Realism: The more human-like AI becomes (ElevenLabs), the more critical the need for defense tools (Ofcom/ICO) becomes.

Until next week.

FAQ: Frequently Asked Questions of the Week

1. Is it true that Amazon will buy OpenAI? It's not about a full acquisition, but a strategic investment (rumored at $50 billion). OpenAI would remain independent, but Amazon would become a key infrastructure partner, reducing OpenAI's total dependence on Microsoft Azure.

2. What is a medical "Foundation Model" like BrainIAC? Unlike an AI trained on cat photos and the internet, a medical foundation model is trained from the start on biological and clinical data (e.g., raw MRI signals, genetic sequences). This allows it to "understand" biology much better than a generalist model like GPT-4 adapted after the fact.

3. Why is "logical" AI like Kona different from ChatGPT? ChatGPT (and classic LLMs) work on a probabilistic basis: they guess the next word. Kona (and reasoning models like o1) use search and logical verification techniques to solve problems step by step. They are less creative, but much more reliable for math, coding, and planning.

4. What does Elon Musk risk with the Grok investigation? He risks hefty fines in the UK and potential operational restrictions in Europe if he cannot demonstrate effective mechanisms to moderate content generated or amplified by his AI.

5. What is the "Empathic AI" Google is investing in? It's a branch of AI that seeks to detect and respond to human emotions. By acquiring Hume AI, Google wants its assistants to understand not only what you say, but how you say it (sad, angry, sarcastic tone), to respond more appropriately and humanly.