The Big Money: Chips, Startups, and Supply Chains

The investment frenzy in AI infrastructure shows no signs of slowing. AI chip maker SambaNova has raised another $1 billion at an $11 billion valuation, just months after rumors of a potential Intel buyout for a fraction of that price. This mega-round underscores the intense demand for alternatives to Nvidia’s dominance. Source

On the software side, a new analysis suggests that top AI startups are seeing their revenue growth rates accelerate, not just remain high, indicating the market is still expanding rapidly for early movers. Source

However, building this hardware requires raw materials. A report notes that despite a U.S. push for a domestic rare earths supply chain, mined materials are still flowing to Asia where manufacturing demand is currently stronger, highlighting a geopolitical tension in the foundation of AI tech. Source

Cutting Costs & Speeding Things Up

As spending soars, a cost-cutting trend is emerging. Microsoft is joining other giants in relying more on its own, cheaper AI models to reduce massive API expenses to partners like OpenAI. Source

Meanwhile, the hot French startup ZML, backed by Yann LeCun, has released free software designed to optimize and speed up AI inference across diverse chips, potentially lowering operational costs for many companies. Source

Product Launches & Platform Expansions

Meta launched Muse Image, a new AI image generator from its Superintelligence Labs, now powering tools in Instagram and WhatsApp. However, users are already pushing back over how it uses their photos, and reports indicate it can surprisingly incorporate other users’ likenesses. (Source 1, Source 2)

Anthropic’s Claude Cowork agent is expanding from desktop to mobile and web, allowing tasks to continue across devices, signaling a move towards persistent, background AI assistance. (Source 1, Source 2)

In hardware, Solos debuted a lighter, camera-less version of its AirGo smart glasses, prioritizing voice-based AI assistance and privacy over visual capture. Source

The Risks: Security, Ethics, and Bugs

The dark side of proliferation is becoming clearer. Security researchers detail “HalluSquatting,” a method where hackers can use popular AI tools to assemble massive botnets, exploiting LLMs’ tendency to hallucinate code and domains. Source

On platforms, Discord admitted an AI moderation bug wrongfully banned users over harmless images, a blunt reminder of the flaws in automated trust and safety systems. Source

In response to risks, startup Savi raised $7 million for an app designed to protect consumers from realistic AI scams, like AI-generated kidnapping ransom calls. Source

In Research & Applied AI

• Insilico Medicine advanced an AI-discovered drug for a lung disease to Phase III trials, a significant milestone for AI in biotech. Source
• Major CPG companies like L’Oréal and Nestlé are using AI to speed up product development, finding new uses for existing ingredients. Source
• The first American autonomous ground vehicles are seeing combat, with over 100 self-driving ATVs deployed in Ukraine by Forterra. Source
• An MIT project showed how AI chatbots can help novice coders, like military personnel, develop software for niche applications. Source

Editorial Take: Today’s stories paint a picture of an industry hitting escape velocity, where the metrics of growth and investment are becoming astronomical. Yet, sandwiched between the billion-dollar rounds and shiny product launches are critical stories about supply chain fragility, security vulnerabilities, and real-world harms from buggy systems. The theme is one of simultaneous ascent and integration: AI is moving from lab to battlefield, from desktop to mobile, and from code to physical product, all while the financial and ethical foundations are being stress-tested in real time. The gold rush is on, but not everyone is playing by the same rules, and the infrastructure is still being built—sometimes shakily—beneath our feet.