Introduction
The week of October 13–17, 2025 was packed with milestone product launches, record-setting partnerships, and fresh signals of how fast artificial intelligence is maturing. From OpenAI transforming ChatGPT into a full app ecosystem to new breakthroughs in video generation and enterprise adoption, here’s what defined the week.
1. OpenAI Turns ChatGPT into a Platform
At DevDay 2025, OpenAI officially launched the Apps SDK, enabling developers to build, publish, and monetise third-party apps inside ChatGPT. Partners such as Canva, Coursera, and Zillow joined the launch wave. ChatGPT now reaches 800 million weekly active users — up 100 million in a single month — cementing its place as the most-used AI platform globally.
2. OpenAI × AMD: A Historic Chip Supply Deal
OpenAI announced a multi-year chip partnership with AMD, giving it access to the next-generation MI450 GPUs and up to 6 GW of compute capacity starting 2026. The agreement includes an option for OpenAI to acquire a minority equity stake in AMD — a move widely viewed as the most significant hardware partnership in AI to date.
3. Anthropic Expands Claude Deployment with Deloitte
Anthropic announced its largest enterprise deployment yet — rolling out Claude Sonnet 4.5 to 470,000 Deloitte employees across 150 countries. The partnership aims to scale AI-assisted consulting, risk analysis, and client delivery while maintaining enterprise-grade compliance.
4. Sora 2 Pro and Google Veo 3.1 Set a New Bar for AI Video
Two major video-generation releases redefined creative AI this week:
- OpenAI Sora 2 Pro introduced sharper detail, consistent motion across cuts, and integrated sound + voice generation, enabling fully synchronised cinematic clips from text prompts.
- Google Veo 3.1 rolled out comparable multimodal capabilities — text-to-video with realistic camera movement, ambient sound, and voice-over synthesis — now available to a limited set of YouTube and DeepMind partners.
Together, they mark the transition from silent AI video to true audiovisual storytelling, closing the gap between script and production.
5. DeepMind Unveils AlphaEvolve for Theorem Discovery
Google DeepMind released AlphaEvolve, an AI system capable of automating parts of theorem discovery — an advance demonstrating how reinforcement learning can accelerate abstract reasoning in mathematics. The system reportedly matched human-expert performance in international benchmarks.
6. GITEX Global 2025 Highlights Sovereign AI
At GITEX Global 2025 in Dubai, governments and enterprises showcased advances in sovereign AI — systems trained and hosted within national borders to enhance privacy and regulatory control. Regional players such as UAE’s Falcon AI and Saudi Data Authority announced new models focused on energy, finance, and education.
7. Healthcare and Science Adoption Accelerates
The U.S. Department of Health and Human Services increased funding for AI-driven paediatric cancer research, while joint projects between Google and Oxford advanced cosmic-event detection models. Across sectors, AI’s role in healthcare, physics, and pharma continues to deepen.
The second week of October 2025 underscored AI’s rapid evolution from experimental tools to platform infrastructure. OpenAI’s Apps SDK opened new economic layers, AMD’s partnership reshaped the hardware race, Anthropic scaled enterprise adoption, and Sora 2 Pro + Veo 3.1 introduced sound-aware generative video. The AI ecosystem is no longer just innovating — it’s consolidating.
