ENFR
8news

Tech • IA • Crypto

TodayMy briefingVideosTop articles 24hArchivesFavoritesMy topics

SpaceX $80B IPO, Gemini AI Search Overhaul, Mistral Buys Emmi

AISaturday, May 23, 2026· 18 videos

Briefing

Audio player
0:00 / 0:00

SpaceX targets $80B IPO

SpaceX has formally launched its IPO process aiming to raise up to $80 billion, potentially surpassing Saudi Aramco’s $26 billion record. The company reported $18.67 billion in revenue and employs over 22,000 staff, spanning launches and Starlink services. Filings reveal a major pivot, with AI infrastructure spending triple that of space operations. It positions SpaceX as a hybrid AI and connectivity giant targeting a $28.5 trillion market.

AI IPO wave reshapes markets

SpaceX, OpenAI, and Anthropic are preparing blockbuster public listings expected to raise tens of billions. Analysts anticipate one of the most intense IPO cycles in years, concentrated around AI leaders. New NASDAQ rules enabling faster index inclusion could drive massive passive inflows immediately after listing. This dynamic may trigger volatility as funds rebalance to gain exposure to AI-heavy equities.

Google Gemini transforms search

Google is overhauling its core product with Gemini-powered AI search, replacing links with direct, conversational answers. Features like AI Overviews and AI Mode reduce the need to click external sites. The system integrates multimodal inputs and can generate synthesized outputs, shifting search into a task-oriented assistant. This marks the biggest change to web navigation in over two decades.

Alphabet revenue surges on AI

Alphabet reported $109.9 billion in Q1 2026 revenue with 11 consecutive quarters of double-digit growth. Google Cloud reached $20 billion, while AI-driven search ads generated about $60 billion. Profits jumped 81%, signaling that AI is already a major profit engine. Gemini now claims 900 million users, fueled by deep integration across Google’s ecosystem.

Mistral acquires Emmi AI

Mistral has acquired Austrian firm Emmi AI to expand into industrial, physics-based AI systems. Unlike generative models, Emmi’s technology simulates real-world constraints for engineering and manufacturing. Tasks like plastic mold design—once requiring 12,000 tests over 18 months—can now be optimized in minutes. The move positions Mistral as a serious European contender in industrial AI.

Princeton unveils Continual Harness

Princeton researchers introduced Continual Harness, a system that improves itself in real time without retraining resets. It modifies its system prompts, sub-agents, skills, and memory during active tasks. The approach enables continuous learning loops, moving beyond static training paradigms. Early results show progression from novice to near-expert performance autonomously.

Enterprise agents gain traction

New workspace AI agents are automating workflows across tools like Slack, CRM systems, and Linear. They convert raw feedback into structured outputs such as PRDs, tickets, and reports. Builders can enforce granular permissions and data access rules, ensuring enterprise control. The shift reflects growing adoption of agentic systems that act, not just respond.

Cast brings real-time voice AI

Cast is a browser-based tool enabling real-time conversations with customizable AI personas using OpenAI and Google Gemini APIs. It runs locally with no installation, emphasizing privacy and accessibility. Unlike traditional pipelines, it uses real-time inference for fluid, interruptible dialogue. The project reflects a broader push toward low-friction, community-driven AI ecosystems.

Videos covered

Previous briefings · AI