Holiday week, but the news didn’t stop. SoftBank is closing a $22.5 billion investment in OpenAI. Disney licensed its characters for OpenAI’s Sora video generator. The TikTok US situation got more interesting. And OpenAI launched its own version of Spotify Wrapped. Plus year-end stats are in - AI repos on GitHub doubled. Quiet week? Not really.
Top Stories This Week
SoftBank Finalizing $22.5 Billion Investment in OpenAI
SoftBank is putting serious money behind OpenAI.
The deal:
The Japanese conglomerate is working to close a $22.5 billion investment in OpenAI by year-end. To make this happen, SoftBank is:
- Selling public-market holdings
- Using additional credit capacity
- Exploring other financing options
Why it matters:
This is one of the largest single investments in an AI company ever. SoftBank has a history of big bets - some worked (early Alibaba), some didn’t (WeWork). This one signals that SoftBank sees OpenAI as a generational opportunity.
The bigger picture:
OpenAI is raising money at a pace we haven’t seen before. Between this deal and the ongoing funding rounds, OpenAI’s war chest is getting massive. That money goes straight into compute, research, and talent. The AI race is expensive, and OpenAI is making sure they can keep up.
Disney Licenses Characters for OpenAI’s Sora
Disney and OpenAI announced a partnership this week.
The deal:
- Three-year licensing agreement
- Disney characters integrated into Sora (OpenAI’s video generator)
- One year exclusivity - after that, Disney can partner with other AI companies
What this means:
You’ll be able to generate videos featuring Disney characters using Sora. The one-year exclusivity clause is interesting - Disney is testing the waters, not committing fully. They want to see how this plays out before locking in.
Why Disney is doing this:
AI-generated content is coming whether Disney likes it or not. People are already creating unofficial Disney content with AI tools. By partnering with OpenAI, Disney gets some control over how their characters appear in AI-generated media.
The risk:
Brand control. Disney is famously protective of their characters. Letting AI generate Mickey Mouse videos is a big step. The partnership probably has guardrails we don’t know about yet.
TikTok US Joint Venture Takes Shape
The TikTok situation in the US is evolving.
What’s happening:
TikTok has finalized a deal with investors including Oracle, Silver Lake, and MGX to establish a US-based joint venture. The new structure will have:
- A majority-American board
- US user data managed under Oracle’s oversight
- Ongoing discussions about algorithm control
The algorithm question:
This is where it gets complicated. TikTok’s recommendation algorithm is the secret sauce. China’s export-control rules make it unclear whether ByteDance can actually transfer algorithm control to a US entity. This isn’t resolved yet.
Why developers should care:
Platform governance and algorithm control are now front-and-center issues. How companies manage recommendation systems, who has access to them, and how they’re regulated - these questions aren’t going away.
NIST Invests $20 Million in AI Research Centers
The US government is putting more money into AI safety and standards.
The investment:
NIST (National Institute of Standards and Technology) is partnering with MITRE Corporation to establish two new centers focused on AI research and development. Total investment: $20 million.
What these centers will do:
- Advance AI technologies and their applications
- Develop standards and best practices
- Focus on responsible AI use
Why it matters:
NIST sets standards that become industry benchmarks. Their work on AI safety and reliability will likely influence how AI systems are built and deployed across sectors. If you’re building AI tools, keep an eye on what comes out of these centers.
AI News
OpenAI Launches “Your Year with ChatGPT”
OpenAI’s answer to Spotify Wrapped is here.
What it is:
An annual review feature that shows users how they’ve interacted with ChatGPT over the past year. Think conversation stats, topics explored, and usage patterns.
Availability:
Currently rolling out to eligible users in:
- United States
- Canada
- United Kingdom
- Australia
- New Zealand
Why OpenAI built this:
Engagement. Features like this get shared on social media. They remind users how much value they’re getting from the product. And they make ChatGPT feel more personal.
Meta Working on “Mango” and “Avocado” AI Models
Meta has new AI models in development.
Mango:
A generative AI model for images and videos. This is Meta’s play to compete with OpenAI’s DALL-E and Sora.
Avocado:
A large language model that will launch alongside Mango. Meta is building a full AI stack - language, image, and video all together.
The strategy:
Meta has been investing heavily in AI, but their consumer products have lagged behind OpenAI and Google. These new models are meant to close the gap.
Microsoft Releases Florence-2 Vision Model
Microsoft Research dropped Florence-2 this week.
What it is:
An open-source vision foundation model that Microsoft is calling a “Swiss Army knife for pixels.”
What it can do:
- Image processing
- Visual analysis
- Multi-task vision applications
Why it matters:
Open-source vision models give developers tools to build image-understanding features without relying on closed APIs. Florence-2 joins a growing ecosystem of open vision models.
Developer Tools
Microsoft Updates Azure Developer CLI and SharePoint Framework
Microsoft shipped updates to their developer toolchains this week.
Azure Developer CLI:
Improvements focused on:
- Better developer experience
- Streamlined cloud deployment workflows
- Preparation for 2026 cloud integrations
SharePoint Framework:
Updates to modernize SharePoint development and integrate with Microsoft’s broader developer platform.
Why it matters:
If you’re building on Azure or SharePoint, check out what’s new. Microsoft has been quietly improving their developer tools throughout 2025.
Year-End Numbers
GitHub: 4.3 Million AI Repositories Created
GitHub released some year-end statistics, and the AI numbers are striking.
The growth:
- 4.3 million AI-related repositories created
- Nearly doubled since 2023
- AI is now a standard part of development workflows
Language trends:
- TypeScript: 1 million new contributors (66% year-over-year growth)
- Python: 850,000 new contributors (48% year-over-year growth)
Why TypeScript is surging:
Two reasons. First, frameworks are defaulting to TypeScript. Next.js, SvelteKit, and others ship TypeScript configs by default now. Second, AI-assisted development works better with stricter type systems. The AI makes fewer mistakes when it has types to work with.
AI Productivity: 20-50% Improvement
Year-end reports are quantifying what developers have felt all year.
The numbers:
- 20-50% productivity improvement from AI tools
- 14% of engineering orgs saw operational cost reductions of 11-19%
- Some organizations report over 20% cost reduction
What developers use AI for most:
- Code generation
- Answering specific technical questions
- Debugging
The takeaway:
AI is no longer experimental for most development teams. It’s part of the workflow. The organizations that have figured out how to integrate AI effectively are seeing real returns.
Gallup: Workplace AI Usage Rising
It’s not just developers. AI adoption is rising across the workforce.
The trend:
- More employees reporting occasional AI use
- Significant jump in frequent users
- AI moving from “experiment” to “daily tool”
What this means for developers:
You’re not just building for developers anymore. Your users increasingly expect AI features. And they’re more comfortable using them than they were a year ago.
What This Week Teaches Us
Money is pouring into AI: SoftBank’s $22.5B investment in OpenAI is one of the largest single AI investments ever. The biggest players are going all-in.
AI companies are buying developer tools: Earlier this month Anthropic got Bun, Cursor got Graphite. The pattern is clear - AI companies want to own the whole development experience, not just the models.
Disney is testing AI waters carefully: One year of exclusivity, then freedom to work with others. Smart way to experiment without committing.
Platform governance is getting real: TikTok’s US joint venture puts algorithm control front and center. Who owns the recommendation engine matters.
The numbers are in: 20-50% productivity gains. 4.3 million AI repos. These aren’t projections anymore. They’re measurements.
The Numbers That Matter
- $22.5 billion - SoftBank’s investment in OpenAI
- $20 million - NIST investment in AI research centers
- 4.3 million - AI-related repositories on GitHub
- 1 million - New TypeScript contributors (66% YoY growth)
- 850,000 - New Python contributors (48% YoY growth)
- 20-50% - Productivity improvement from AI tools
- 3 years - Disney’s licensing deal with OpenAI (1 year exclusive)
Quick Hits
SoftBank betting big on OpenAI - $22.5B is serious money. They’re selling assets and using credit lines to make it happen.
Disney characters coming to Sora - Three-year deal, but only one year exclusive. Disney is testing, not committing.
TikTok US structure taking shape - Oracle oversight, majority-American board. Algorithm control still unclear.
TypeScript growth is massive - 66% more contributors than last year. AI-assisted coding is part of the reason.
Meta building full AI stack - Mango for images/video, Avocado for language. Competing with OpenAI on all fronts.
Your Year with ChatGPT is here - OpenAI’s Spotify Wrapped equivalent. Check your stats.
NIST investing in AI safety - $20M for two new AI research centers with MITRE. Standards are coming.
Holiday week, but the deals kept coming. SoftBank’s $22.5B OpenAI investment is the headline. Disney testing AI-generated content with Sora is worth watching - if Disney is doing it, others will follow. And the year-end numbers confirm what we’ve all felt: AI changed how developers work in 2025. The question for 2026 isn’t whether AI will matter. It’s how fast things will move.
See you next year. Happy New Year to all developers.
