PLUS: Anthropic posts its first profit, OpenAI solves an 80-year-old problem, and an AI film's $400k compute bill
Happy reading
Nvidia is wasting no time, unveiling its next-generation 'Rubin' AI platform just months after its Blackwell architecture. The new hardware signals a major push beyond the data center and into the world of physical, embodied AI.
This rapid-fire release schedule shows Nvidia is aggressively building the full-stack infrastructure for AI's next chapter. But with new platforms emerging so quickly, is the company creating an unbeatable ecosystem or moving too fast for the industry to keep up?
In today’s Next in AI:
Nvidia unveils its next-gen Rubin AI platform
Anthropic posts its first-ever profit
OpenAI model cracks 80-year-old math problem
An AI film’s $400k compute bill at Cannes
Nvidia's Next Chapter
Next in AI: Just months after announcing its Blackwell architecture, Nvidia has unveiled its next-generation platform at COMPUTEX 2026. Codenamed "Rubin," the new platform signals a major push into AI factories and physical, embodied AI.
Explained:
The centerpiece is the Vera Rubin NVL72, a rack-scale AI supercomputer that combines 72 Rubin GPUs with new Vera CPUs to deliver 10x higher inference performance per watt for massive, trillion-parameter models.
Nvidia is also expanding to the edge with the NVIDIA Jetson Thor, a powerful compute module designed to bring generative AI to autonomous robots, industrial systems, and medical devices.
For autonomous vehicles, the new NVIDIA Alpamayo open platform helps developers train AVs to handle rare and complex long-tail driving scenarios, like interpreting ambiguous hand signals.
Why It Matters: This rapid succession of platforms shows Nvidia’s strategy to build the complete, full-stack infrastructure for AI's next wave, from massive cloud systems to intelligent machines in the physical world. By creating specialized hardware for data centers, robotics, and AVs, the company is laying the foundation for more capable and autonomous systems to become a part of our daily lives.
Anthropic's Profit Engine
Next in AI: Anthropic is projecting a massive $10.9 billion in Q2 revenue and its first-ever operating profit, signaling a new phase of maturity where rapid growth is finally translating into a sustainable business model.
Explained:
The company’s projected Q2 revenue more than doubles its $4.8 billion from Q1 and is on track to exceed all of last year’s revenue in a single quarter.
Strong enterprise adoption is driving this growth, with over 1,000 companies now spending at least $1 million annually on Anthropic’s Claude products and APIs.
The impressive financials are fueling talks of a new funding round that could value Anthropic at nearly $900 billion, positioning it to surpass rival OpenAI.
Why It Matters: This marks a critical turning point, proving that foundational AI models can become highly profitable businesses. Investor focus is now quickly shifting from user growth and hype to sustainable, enterprise-driven revenue.
An 80-Year-Old Problem Solved
Next in AI: OpenAI announced that one of its models made a breakthrough on a famous 80-year-old mathematics problem. This marks a significant milestone in AI's ability to reason through complex scientific challenges.
Explained:
The AI tackled the Erdős planar unit distance problem, which asks how many pairs of points on a flat surface can be the same distance apart. For decades, mathematicians assumed grid-like patterns were optimal, but the model discovered a completely new family of constructions that performs better.
While the AI didn't produce a final answer for the problem's growth rate, it successfully disproved the long-standing conjecture by showing its proposed limit was too low.
Unlike a controversial claim last year, this result was validated by mathematicians, including Thomas Bloom and Tim Gowers, who helped refine the AI's proof and highlighted the power of human-AI collaboration.
Why It Matters: This demonstrates that AI can serve as a powerful tool for scientific discovery by exploring unconventional paths that human researchers might dismiss. These advanced reasoning capabilities are not limited to math and have profound implications for accelerating progress in fields like biology, physics, and engineering.
AI's Cannes Debut
Next in AI: A 95-minute, fully AI-generated film called “Hell Grind” just premiered at the Cannes Film Festival. The real story, however, isn't just the film itself—it's that AI compute power accounted for a stunning $400,000 of its $500,000 budget.
Explained:
The film’s 80% compute cost highlights the immense processing power required for high-quality, long-form video generation. Startup Higgsfield AI spent just two weeks on production but racked up the massive bill prompting models for thousands of video clips.
Creating the film still required deep filmmaking know-how, from camera composition to lighting. Prompts averaged 3,000 words each just to ensure visual consistency and direct the AI to respect the laws of physics.
The premiere signals a shift in Hollywood from existential fear to cautious acceptance of AI. The mood at Cannes reflects a growing sense of artistic uncertainty as the industry figures out how to integrate these new tools.
Why It Matters: This project proves that feature-length AI filmmaking is no longer theoretical, but it also reveals a new economic reality. The primary barrier for AI-native media is quickly becoming the cost of compute, not the capability of the creative tools.
AI Pulse
China banned Nvidia’s RTX 5090D V2, a chip designed specifically to comply with US export rules, marking the first time Beijing has blocked one of the company's custom chips made for the country.
Spotify partnered with Universal Music Group on a deal that allows users to create and share AI-generated covers and remixes using the voices of participating artists.
Intuit announced it will lay off over 3,000 employees, about 17% of its workforce, as part of a restructuring to focus its resources on AI initiatives.
Grok appeared in just three of over 400 publicly identified AI use cases within the U.S. federal government, signaling slow adoption for xAI's chatbot compared to rivals from OpenAI, Google, and Anthropic.