It's been another busy week at the frontier of artificial intelligence. We're tracking massive funding rounds, strategic acquisitions that are reshaping the competitive landscape, and research that forces us to think deeply about the future of autonomous systems.
This week, we're covering:
- A Historic $2 Billion Seed Round: Mira Murati's new venture is making waves.
- Meta's $14.8 Billion Play: A major acquisition shakes up the data-labeling market.
- Advancements in Embodied AI: A new open-source model for humanoid robots.
- Midjourney Enters Video: The creative AI space gets a powerful new player.
Let's dive in.
Mira Murati's Thinking Machine Lands $2B in Landmark Seed Round
Former OpenAI CTO Mira Murati has secured a monumental $2 billion seed round for her new startup, Thinking Machine, rocketing its valuation to $10 billion. This historic funding signals deep investor confidence in the race to build the next generation of advanced, agentic AI systems.
Key Points:- The company aims to advance AI through a combination of open science principles and practical, real-world applications.
- Murati has already attracted top-tier talent, poaching key researchers and engineers from competitors like Meta AI and Mistral.
- The startup's mission focuses on developing autonomous AI agents capable of complex reasoning, planning, and long-term task execution.
This is one of the largest seed rounds in tech history, underscoring that the AI arms race is intensifying and capital is flowing towards proven leaders. For entrepreneurs and investors, this validates the enormous market potential for agentic AI that moves beyond simple generation. Murati's ability to attract elite talent so quickly also highlights that human capital is the most critical and competitive resource in the AI landscape today.
Meta's $14.8B Power Play for Scale AI Stake
In a strategic move that has sent shockwaves through the industry, Meta has acquired a 49% non-voting stake in Scale AI for a staggering $14.8 billion. Scale AI is a cornerstone of the AI ecosystem, providing the critical data labeling and curation services required to train large models.
As part of the deal, Scale AI’s highly-regarded CEO, Alexandr Wang, will take on a senior leadership role at Meta while remaining on Scale's board. The acquisition gives Meta significant influence over a crucial part of the AI supply chain and brings one of Silicon Valley's most connected founders into its fold.
Why it matters: This deal is about more than just data; it's a strategic talent and infrastructure play. By securing a major stake in Scale, Meta vertically integrates a key dependency and disrupts its rivals, with competitors like Google already severing ties with the company. For the broader market, this raises serious concerns about competition and the potential for antitrust scrutiny, as it concentrates power over a fundamental AI resource within one of the industry's largest players.
The Rundown
Quick Hits: Research, Policy & Products
- OpenAI secures a $200M U.S. Defense Department contract to build tools for cybersecurity and healthcare, launching its "OpenAI for Government" initiative and marking a major step into national security applications.
- BAAI releases RoboBrain 2.0, an open-source AI model for humanoid robots, that excels at interactive reasoning and outperforms some proprietary models on embodied AI benchmarks. (X Post, GitHub)
- Midjourney releases its first video model, V1, allowing users to generate short, high-coherence video clips from text and image prompts, intensifying competition in the creative AI space.
Actionable Insights
Pro Tip of the Week: Few-Shot Prompting
To dramatically improve AI model accuracy, use few-shot prompting. Instead of just asking a question (zero-shot), provide the model with 2-3 examples of the input and desired output first. This gives the AI a clear pattern to follow.
Example Application: For classifying customer support tickets, structure your prompt like this: Input: "My login isn't working." -> Output: "Technical Issue"Input: "How do I upgrade my plan?" -> Output: "Billing Question"Input: "The new feature is amazing!" -> Output: "Positive Feedback"Now, classify this: [Paste New Ticket Here]
This announcement from OpenAI's CEO, along with a follow-up revealing an 80% price cut for the previous `o3` model, dominated the AI conversation this week. It reflects OpenAI's dual strategy: aggressively pushing the performance frontier while simultaneously making last-generation's cutting-edge tech more accessible to broaden market adoption.
That's a wrap for this week. See you soon.