Arveum Capital PartnersCapital Partners
🤖

AI Newsletter

May 23, 2026 · 04:45 Uhr

1

OpenAI releases ChatGPT plugin that works directly in PowerPoint

THE DECODER

OpenAI integrates ChatGPT directly into Microsoft PowerPoint, automating a core workflow for millions of office users. The plugin significantly lowers the barrier to AI adoption in the enterprise sector and strengthens OpenAI's position against competitors like Google. At the same time, the strategic alliance between OpenAI and Microsoft deepens, with Office integration providing a competitive advantage.

2

Deepseek prioritizes AGI research over quick profits despite multi-billion financing

THE DECODER

Deepseek secures $10 billion in funding at a $45 billion valuation and signals strategic focus on AGI development rather than short-term profitability – typical of a long-term oriented deep-tech competitor against OpenAI/Google. This intensifies competitive pressure in the AI market and demonstrates Chinese capital strength in strategic tech sectors.

3

OpenAI Codex gets Appshots: send app windows to AI with a keystroke

THE DECODER

OpenAI expands its Codex assistant with Appshots, enabling developers to integrate app content directly into the AI context – a UX upgrade that boosts productivity and simplifies AI integration into existing workflows. This strengthens OpenAI's market position in the coding assistant segment against GitHub Copilot and similar competitors through improved usability and context processing.

4

Despite billion-dollar revenue: OpenAI continues to burn far more money than it makes

THE DECODER

While OpenAI generates billion-dollar revenues ($5.7 billion USD Q1 2026), it burns $1.22 for every dollar earned – a massive profitability problem despite scaling. The negative operating margin of -122% indicates unprofitable business models driven by high infrastructure costs (GPU/computing power).

CRITICALRead article
5

Google tests websites for llms.txt and agent compatibility

THE DECODER

Google develops standardization mechanisms (llms.txt, Agentic Browsing) for AI agent-website compatibility and establishes new technical requirements for websites. This could become the de-facto standard and force websites to adapt their infrastructure – similar to Mobile-First Indexing back then. Companies need to proactively optimize their digital presence for autonomous AI systems.

Tokens: 1,412(889 in · 523 out)

This website uses cookies. Strictly necessary cookies are always active. By clicking "Accept all" you additionally consent to analytics cookies (Google Analytics). Privacy Policy →