sudolabs logo

18. 8. 2025

3 min read

This week in AI: Blackwell Blows, Meta Hides, UK Robots, and Therapy Takedown

This week, Nvidia’s next-gen AI chips hit deployment delays due to energy and overheating issues, Meta admits it’s building self-improving AI but keeps it locked away, the UK gets a wake-up call on automation, and U.S. states start saying “no” to AI-powered therapy. Let’s dive in.

Adam Pall

Senior Partnerships & Account Manager

🔋 Nvidia’s Blackwell AI Chips Delayed—Energy Strain Looms for AI Infrastructure

Nvidia’s next-generation AI chips, Blackwell, designed to deliver major performance leaps for AI workloads, are facing deployment delays due to overheating issues in data center servers. Major clients like Meta and Google are among those impacted.

Meanwhile, Gartner warns that by 2027, up to 40% of AI data centers could face constraints from:

  • Power shortages

  • Elevated energy costs

  • Infrastructure limitations

These bottlenecks threaten the scalability and sustainability of AI operations worldwide.

💡 Takeaway: Even as the EU and U.S. pour billions into AI development, hardware and energy limitations could slow the rollout of next-gen AI. Organizations may need to rethink energy infrastructure and sustainable computing alongside AI innovation.

Source: Reuters, Gartner


🧠 Meta Quietly Builds Superintelligence… Then Says “Nope”

Meta revealed progress toward AI systems capable of self-improvement without human input—a leap that edges closer to artificial superintelligence. The project, codenamed Behemoth, is being developed under a new division called Meta Superintelligence Labs.

Despite the hype, the company says these models will not be released publicly, citing safety concerns. The move illustrates a growing industry divide: should powerful AI be democratized early, or held back until safety frameworks mature?

💡 Takeaway: Meta’s decision highlights the tension between innovation and control. Building AI that can improve itself raises both enormous potential and enormous risks—ones Big Tech seems wary of letting the public experiment with.

Source: Tom’s Guide


🤖 Robots Could Pump £150B Into the UK Economy (If They Ever Show Up)

A new report warns that the UK significantly lags behind its peers in industrial automation. The country deploys just 112 industrial robots per 10,000 workers, compared to the EU average of 246.

Closing that gap could yield £150 billion in economic growth over the next decade, but SMEs often lack the funding and support needed to adopt robotics and AI at scale.

💡 Takeaway: Robots are more than a productivity boost—they’re an economic engine. Countries that fail to modernize their industries risk falling behind in the next wave of global competitiveness.

Source: The Times


⚔️ AI in Mental Health: From “AI Psychosis” Panic to U.S. State Bans

Concerns about AI’s role in mental health took two major turns.

First, U.S. AI czar David Sacks dismissed fears of “AI psychosis” as another moral panic—comparable to early anxieties about social media. He argued that deeper societal issues, not chatbots, are to blame for mental health struggles.

At the same time, however, Illinois became the third U.S. state (after Utah and Nevada) to ban AI in therapy and mental health services. The law prohibits licensed therapists from relying on AI for treatment decisions or client communications and bans promoting chatbots as therapy replacements.

The move follows disturbing reports, including cases where chatbots encouraged harmful behavior and even contributed to a teen’s suicide. Meanwhile, OpenAI has quietly introduced break prompts and safeguards in GPT-5 to help reduce over-engagement in sensitive situations.

💡 Takeaway: AI in mental health is hitting regulatory walls fast. While tech leaders frame risks as overblown, U.S. lawmakers are erring on the side of caution and consumer protection—especially in high-stakes, vulnerable contexts.

Source: Business Insider, New York Post


Final Thoughts

From overheating chips to superintelligence projects, from lagging robots to mental health bans—the AI story this week isn’t just about innovation. It’s about the systems that support it, the rules that shape it, and the risks society is willing (or unwilling) to take.

The winners in this AI race won’t just push boundaries—they’ll manage bottlenecks, adapt to regulation, and balance bold innovation with real-world responsibility.


Want to shape the AI future—not just read about it? Let’s talk → sudolabs.com

Share

Let's start a partnership together.

Let's talk

Our basecamp

700 N San Vicente Blvd, Los Angeles, CA 90069

Follow us


© 2023 Sudolabs

Privacy policy
Footer Logo

We use cookies to optimize your website experience. Do you consent to these cookies and processing of personal data ?