Did Microsoft Just Kick off an R&D Revolution?

Brought to you by Transformative Studio

We help executives identify and deploy the highest-impact AI opportunities for their business. Learn more at Transformative.Studio.

Hey, it's Clark.

This week, Microsoft's AI autonomously discovered a new battery material in 200 hours by running 32 experiments, treating each failed attempt as valuable data rather than a dead end. This methodical persistence mirrors what Satya Nadella described in his viral interview about using AI agents as his "chiefs of staff"—agents who never get discouraged, never lose focus, and never stop connecting dots across vast information landscapes.

With this, we’re watching grit become programmable.

Let’s dive in…

IN TODAY'S EDITION: 

Top Story: Microsoft's AI autonomously discovers novel battery material in 200 hours

Industry Watch: Volvo integrates Gemini AI into vehicles

Notable Transactions: OpenAI secures $11.6B for AI supercomputing

New Tools to Try: Plansom for AI-powered project management

Power Prompt: Deep-research-led process of naming your product

THIS WEEK’S TOP STORY

Microsoft's AI discovers novel battery material in 200 hours, reducing traditional R&D timelines by years

Working with Pacific Northwest National Laboratory, Microsoft's autonomous AI system performed 32 experiments to identify a new material for sodium-ion batteries. The system independently designed experiments, analyzed results, and overcame failures by adjusting its approach. This breakthrough showcases AI's ability to dramatically accelerate material science innovation. (VentureBeat)

INDUSTRY WATCH

Manufacturing | Volvo First to Embed Google Gemini AI in Vehicles 

Volvo becomes the first automaker to integrate Google's Gemini AI directly into cars, enabling natural voice interactions beyond standard infotainment. The technology, built into Android Automotive OS, allows drivers to ask complex questions and receive information, from navigation to recipes, without internet connectivity. (The Verge)  

Real Estate | Zillow Introduces 'Offer Strategies' Feature to Guide Homebuyers

Zillow is testing a new tool that provides users with local insights, including how long the home has been on the market versus how soon homes go pending in that neighborhood. Prospective buyers can select a "strong," "competitive," "moderate" or "weak" offer to see a price range for each and the estimated likelihood of an offer in that range being accepted. (Real Estate News)

Agriculture | AI-IoT System Detects Plant Disease with 99.8% Accuracy 

Researchers developed an automated system that combines AI with connected sensors to detect and treat plant diseases in real time. Using a modified irrigation setup with smart cameras, the system achieves 99.8% accuracy across 11 plant health categories, enabling targeted treatment while reducing manual monitoring. (AZO)

Tech | Microsoft's AI Agents Collaborate Autonomously on Complex Tasks 

Microsoft unveiled its AutoGen framework, enabling AI agents to communicate and break down multi-step business processes independently. In demonstrations, the agents coordinated to resolve customer support issues and automate DevOps tasks without human input, potentially redefining workplace workflows. (VentureBeat)

Healthcare | MIT's 'Sybil' AI Predicts Lung Cancer Risk Six Years Ahead 

MIT's Jameel Clinic developed 'Sybil,' an AI that forecasts lung cancer risk up to six years using only low-dose CT scans. The system achieved 86% and 94% accuracy in predicting one-year risk for smokers and non-smokers respectively, validated on US and Taiwanese datasets. (Healthcare in Europe)

Banking & Finance | Visa's AR Manager Automates Virtual Card Payments, Reducing Supplier Processing Time

Visa has launched AR Manager, a tool that automates virtual card transactions for suppliers, eliminating manual processing and accelerating payment reconciliation. This innovation aims to enhance cash flow and boost the acceptance of virtual cards across businesses. (PYMNTS)

Retail | Google Unveils AI Shopping Tools at I/O 2025 

Google introduced AI-driven shopping features, including "AI Mode" with visual panels that update product listings based on queries, price tracking with spending limits and notifications, and enhanced virtual try-on using personal photos. These features will roll out in the U.S. over the coming months. (TechCrunch)

Media | Netflix Explores AI-Generated Personalized Advertising 

Netflix is reportedly testing generative AI tools to customize and insert branded messages directly into programming by 2025. The technology could allow advertisers to adapt commercials for individual viewers, blending them seamlessly into shows with unprecedented precision. (PYMNTS)

NOTABLE TRANSACTIONS

OpenAI’s Stargate Project Secures $11.6B to Build AI Supercomputing Hub – Backed by Microsoft, the project will span 1,000 megawatts in power capacity and launch by 2028, supporting OpenAI’s long-term AI infrastructure needs. (The Decoder)

Google Invests $150M in AI Glasses Partnership with Warby Parker – The collaboration aims to create consumer-ready AI-powered glasses with real-time assistance and advanced vision features. (TechCrunch)

Moonvalley Lands $53M for AI Video Creation Platform – The stealth startup is building tools to generate high-quality videos using generative AI, according to an SEC filing. (TechCrunch)

Granola Raises $43M to Power AI-Driven Meeting Transcript Search – The company helps enterprise teams instantly retrieve insights from past calls, with backers including Benchmark and ex-GitHub CEO Nat Friedman. (PYMNTS)

Salesforce Acquires ConvergeIQ to Accelerate AI Agent Development – The acquisition will enhance Salesforce’s capabilities in building autonomous AI agents to automate complex business tasks. (PYMNTS)

NEW TOOLS TO TRY

POWER PROMPT

This Week: Deep-Research-led Process of Naming Your Product

It does research on each name it comes up with, checking for trademark availability, domain cost, etc.

I need a name for the new product I'm building. In the following order, I want you to:

- think deeply about what the product is and what would resonate with the user
- come up with a list of 30 initial candidates (each unique and near-optimal)
- going one by one, do intensive research on each... are there others with the same or similar names? are there negative connotations to the names? is the trademark available? show your work for each of these.
- make a list of 'still viable' names, and a list of 'crossed out' names
- for the 'still viable' list, check if each one has available and acceptable domains that fit the budget I give you
- give me a final set of options in a table with reasoning, domains/cost, etc.

<inputs>
   <product_description>
      {{product_description}}
   </product_description>
   <budget>
      {{budget_for_domain}}
   </budget>
   <names_i_like_with_good_vibes>
      {{names_you_like}}
   </names_i_like_with_good_vibes>
</inputs>

Prompt credit: Matt Shumer

THAT’S A WRAP

Enjoyed this edition? Share the Transformative AI Newsletter with colleagues to help them stay ahead in the AI landscape.

Remember: AI mirrors our intelligence back to us, but consciousness remains our unique signature.

See you next week,
Clark Valberg

Brought to you by
Transformative Studio

We help business leaders identify opportunities to leverage AI for real business impact.

Together, we’ll consider your unique challenges, identify high-impact opportunities and deploy custom AI workflow agents that deliver measurable ROI in a short timeframe.

Request an AI Assessment to see where the highest-impact opportunities are for your business.