The Executive Data & AI Brief - September 2025

September 2025 Edition, By Emmanuel Asimadi

September 2025 by the Numbers

$100B

OpenAI-NVIDIA Partnership Value

10 GW

GPU Systems to Deploy

7

Companies in FTC Child Safety Probe

200K

Claude 4.5 Context Window (tokens)

Key Developments Timeline

September 11

Oracle Stock Surges 36%

$300B OpenAI deal revealed, $455B total contracted revenue

1

September 15-16

Samsung AI Forum 2025

Global AI research convergence with Yoshua Bengio, Stefano Ermon

2

September 2025

Anthropic Claude 4.5 Sonnet

Agentic AI with computer use and 200K context window launched

3

September 2025

EU AI Act Goes Live

First comprehensive AI governance framework takes effect

4

September 24

Microsoft Adds Anthropic

Claude models integrated into Copilot, ending OpenAI exclusivity

5

September 25

Databricks-OpenAI Launch

Agent Bricks brings frontier models to enterprise data lakehouses

6

September 2025

Google AI Mode Expansion

Rolled out to 180+ countries with multimodal Search Live

7

🏗️

1. The AI Infrastructure Arms Race Escalates

• Oracle Emerges as AI Infrastructure Dark Horse with $300B OpenAI Deal

• OpenAI × NVIDIA sign an unprecedented $100 B / 10 GW deal to co-build AI data centers — the largest compute commitment in history.

• CoreWeave expands its OpenAI contract by $6.5 B, while Meta inks a $14 B compute agreement of its own.

• Why it matters: Compute capacity, energy, and supply chains are now strategic assets. Expect power availability and GPU access to appear in board-level risk registers.

👉 Implications: Compute capacity, energy, and supply chains are now strategic assets. Expect power availability and GPU access to appear in board-level risk registers.

👉 What could I do?

⚔️

Platforms & Enterprise AI

• AWS Bedrock adds DeepSeek-V3.1, Anthropic Claude 4.5, and new AgentCore tooling.

• Databricks partners with OpenAI to embed GPT models natively.

• Microsoft Breaks OpenAI Exclusivity, Integrates Anthropic's Claude

• Snowflake, Salesforce, IBM, and Microsoft roll out agent frameworks, semantic layers, and AI copilots

• Google Scales AI Mode to 180+ Countries, Introduces Multimodal Search Live

👉 Implications: The enterprise stack is consolidating around multi-model, governed AI platforms. Vendor strategy now equals AI strategy. Expect 20-30% better task performance by matching workload to specialized models

👉 What could I do?

🤖

Model Innovation & Agentic AI

• Anthropic Claude 4.5 Sonnet: Agentic Work Arrives

• Google’s Gemini 2.5 adds “Computer Use” agents, multimodal visual search, and Chrome integration.

• Microsoft 365 Copilot introduces Agent Mode for workflow automation.

• Meta Connect 2025 unveils Ray-Ban Display smart glasses and new Horizon OS SDKs.

• Apple quietly tests its internal “Veritas” chatbot.

👉 Implications: AI is no longer a feature — it’s the interface. Employee and customer expectations are being reset toward intelligent assistance. Agentic architecture shifts from "copilot" assistance to autonomous task delegation; redefines human-AI collaboration models

👉 What could I do?

⚖️

Governance, Trust & Regulatory

• California Finalizes New Privacy Rules for AI Systems

• EU AI Act: Governance Rules Take Effect

• Federal Judge Rejects Anthropic's Record $1.5B Copyright Settlement

• FTC Launches Major AI Chatbot Child Safety Probe

👉 Implications: Global regulators are shifting from principles to enforcement. Every AI initiative now needs a governance playbook.

👉 What could I do?

Love to know more or work with me?

leave a comment or reach out to emmanuel@re-data.ai