Signal + Noise: GEN AI WEEKLY RUN DOWN, 02.19.25
By Isaiah Steinfeld, Founder & CEO, Neue Alchemy
February 19, 2025
The OpenAI exodus is reshaping the AI landscape—and three competing visions are emerging. From trillion-dollar AGI infrastructure plays to real-time decentralized ecosystems and personalized assistants, this week’s Rundown tracks the people, power shifts, and platforms defining AI’s future.
Plus: Microsoft enters the quantum race, open-source models surge, and the next M&A wave takes shape.
OpenAI and Anthropic are still leading the big-model race, but they’re no longer the only forces shaping AI. Ex-OpenAI leaders are now building competing ventures, pushing the industry in three radically different directions:
🔹 Safe Superintelligence (SSI) (Sutskever) → A $30B AGI safety-first moonshot, raising $1B+ with zero revenue or product—just a vision for “safe superintelligence.”
🔹 Thinking Machines Lab (Murati) → A direct OpenAI challenger, focused on highly personalized AI assistants and human-AI collaboration over general-purpose LLMs.
🔹 xAI’s Grok 3 (Musk) → A real-time, decentralized AI, optimized for speed, scale, and direct social platform integration—built to challenge OpenAI and Meta.
But there’s another front emerging—the open-source AI fight. While these AI giants fight for dominance, a new class of open models is making a real push for mainstream adoption:
🔹 Meta (LLama 3) → Leading the charge on open-weight AI, giving developers more flexibility and challenging closed models like GPT-4 and Claude.
🔹 Mistral → Europe’s biggest AI play, building compact, high-performance open models that are gaining traction in enterprise and developer circles.
🔹 DeepSeek AI → China’s most advanced LLM project, pushing open-source AI at a fraction of the cost of Western models.
Meanwhile, Microsoft just dropped its first quantum chip, Majorana 1, signaling a shift toward Quantum + AI fusion that could reshape the entire compute landscape.
🚨 Why This Matters: This isn’t just competition—it’s a philosophical battle over AI’s future.
• SSI bets on a slow, safety-driven AGI race with heavy regulation.
• Thinking Machines Lab argues AI should be customized and collaborative, not one-size-fits-all.
• xAI believes in mass accessibility and decentralized knowledge over centralized control.
• Meta, Mistral, and DeepSeek are pushing open-source as the real future of AI.
📌 Key AI Moves This Week
🛑 Safe Superintelligence Hits $30B Valuation
Ilya Sutskever’s Safe Superintelligence (SSI) is raising $1B+, led by Greenoaks Capital, with a $500M investment—despite having no product, no revenue, and no commercialization plan.
🔹 🧐 So What? – SSI is now one of the most valuable AI startups ever without a working product. Investors are betting on safety-first AGI as the ultimate moonshot.
🔹 🔥 Hot Take – AGI investment is moving toward trillion-dollar infrastructure plays over short-term ROI.
🔹 Key Implications:
• Expect a massive talent war as SSI poaches researchers from OpenAI, DeepMind, and Anthropic.
• AGI regulation will become a high-stakes issue as policymakers scramble to keep up.
• Investors are signaling they’ll fund long-term AI bets over quick applications—game-changing for startup funding.
💰 Biggest Open Question: Can SSI actually build safe superintelligence before OpenAI?
🛑 Mira Murati’s Thinking Machines Lab Challenges OpenAI
Former OpenAI CTO Mira Murati has officially broken away, launching Thinking Machines Lab to develop customizable, human-centric AI assistants.
🔹 🧐 So What? – This directly challenges OpenAI’s “one model fits all” strategy, shifting focus toward adaptive, collaborative AI that integrates seamlessly into different industries.
🔹 🔥 Hot Take – The future of AI isn’t just bigger models—it’s models designed for individuals.
🔹 Key Implications:
• Enterprise AI will move toward bespoke, fine-tuned models rather than generic LLMs.
• Expect OpenAI, Meta, and Google to pivot toward more customizable AI offerings.
• Murati’s leadership could trigger an OpenAI talent exodus, especially among researchers frustrated with its closed-source approach.
💡 Watch This Space: Murati is bringing in ex-DeepMind, OpenAI, and Google AI talent—when she locks in funding, Thinking Machines Lab could become OpenAI’s biggest competitor.
🛑 xAI Releases Grok 3 – Musk’s AI Challenger Steps Up
Elon Musk’s xAI unveils Grok 3, a real-time learning AI, rolling out on iOS, web, and X (Twitter).
🔹 🧐 So What? – xAI is directly competing with OpenAI, Google Gemini, and Meta’s Llama—but with a focus on speed, real-time learning, and accessibility.
🔹 🔥 Hot Take – Musk is betting on an AI ecosystem that learns from the world in real time, rather than relying on massive pre-trained models.
🔹 Key Implications:
• Grok 3’s integration into X (Twitter) could change AI-driven social media.
• Expect Musk to position xAI as an AI-first alternative to Google Search.
• Real-time AI could be a major differentiator, giving xAI a unique advantage over OpenAI’s periodic model updates.
🚀 Wild Card Move: Musk has hinted at building the world’s largest AI supercomputer—if he follows through, xAI could become a true AI contender.
🛑 Microsoft’s First Quantum Chip – Majorana 1
Microsoft unveils Majorana 1, its first quantum computing chip, featuring eight topological qubits—a breakthrough 20 years in the making.
🔹 🧐 So What? – Quantum computing is finally transitioning from theory to reality—with huge implications for AI training, cryptography, and enterprise computing.
🔹 🔥 Hot Take – AI will hit a compute wall without quantum breakthroughs. Microsoft is positioning itself to own the future of AI scaling.
🔹 Key Implications:
• Microsoft’s quantum-AI fusion strategy could redefine cloud computing.
• This sets up a quantum arms race with Google, IBM, and Rigetti.
• Expect Microsoft to commercialize quantum compute via Azure, reshaping AI hardware competition.
🧐 The Big Question: Will quantum AI be a game-changer or a hype bubble?
💰 AI Investment & M&A Moves
✅ HP Acquires Humane’s AI Pin Assets for $116M – A cautionary tale in AI hardware—Humane failed due to pricing, battery life, and unclear market fit.
✅ Luminance Raises $75M for AI-Powered Legal Contracts – Legal AI is now a core enterprise category, with over 700 enterprise clients.
✅ Hightouch Raises $80M on a $1.2B Valuation – AI-driven data syncing & marketing automation are surging.
💡 Final Thought: AI’s Future is Splitting into Three Paths
1️⃣ Massive AGI bets (Safe Superintelligence, OpenAI)
2️⃣ Personalized AI assistants (Thinking Machines Lab)
3️⃣ AI-first real-time ecosystems (xAI, Meta, Microsoft)
The AI world is no longer a single race—it’s three competing visions.
Building in this space? We’re tracking signals, shaping strategy, and working with companies navigating what’s next.
Let’s connect—whether you’re scaling something ambitious, exploring vertical AI, or rethinking your infrastructure.
📩 hello@neuealchemy.com
