1.

Though nearly every organization is experimenting with artificial intelligence, 95% are seeing no return on investment, according to The GenAI Divide: State of AI in Business 2025, a report released this summer by the MIT Media Lab’s Project NANDA (Networked Agents and Decentralized AI). The report, which draws on a review of more than 300 AI initiatives, interviews with 52 organizations, and a survey of 153 senior business leaders, estimates companies have spent $30–40 billion on AI platforms over the past two years, with just 5% reporting any ROI. “The core barrier to scaling is not infrastructure, regulation, or talent,” the authors note. “It is learning. Most GenAI systems do not retain feedback, adapt to context, or improve over time.” Not surprisingly, the firms finding success are those solving for learning and memory and aligning AI with their workflows and operations. (See the box on page 7, where the authors debunk five myths about AI in the enterprise.) So is AI coming for your job? Layoffs have been minimal so far, concentrated mostly in outsourced back-office and customer support.

Read here.

2.

Investors have for a while now examined the environmentally sound use of AI given the demands of energy-hungry data centers that power it. Hence, we liked reading Measuring the environmental impact of delivering AI at Google Scale, a technical paper by a team of scientists there that details how much energy its Gemini apps use for each query, as well as the greenhouse gas emissions and water use from such prompts. Median energy consumption for a Gemini prompt is 0.24 watt-hours, as of May 2025, or roughly the same amount of energy as watching TV for nine seconds, the paper notes. A prompt generates 0.03 grams of carbon dioxide-equivalent emissions across all emissions scopes (which is like running a light bulb for a blink) and consumes 0.26 milliliters of water, or roughly five drops. The challenge of measuring these environmental impacts, and the effort by the tech giant to assess those impacts holistically, stands out. (Table 2 in the paper compares measurements from Google’s framework with existing approaches.) As measurers ourselves, we appreciate the authors’ calling for standard approaches to quantifying AI impacts. As they note, “while the impact of a single prompt is low compared [with] many daily activities, the immense scale of user adoption globally means that continued focus on reducing the environmental cost of AI is imperative.”

Read here.

3.

What happened with plastics? You might have heard that in the dog days of summer, negotiations to reach a global treaty on plastic pollution ended without an agreement. Need an explainer? We found two good sources to catch you up. “The biggest issue of the talks has been whether the treaty should impose caps on producing new plastic or focus instead on things like better design, recycling and reuse,” Jennifer McDermott reported for the Associated Press in No end in sight to plastic pollution crisis as treaty negotiations in Geneva fail.  You’ll get a concise round up of all sides in the negotiations from this piece of reporting. If you need to brush up on the scale of the problem with plastics pollution, we recommend Can a New Global Treaty Tackle the Plastic Pollution Crisis?, a reference-rich curtain raiser by Alex Lee-Emery, Rocky Pairunan and Amy Swift of the World Resources Institute published on the eve of the talks. The world produces roughly one million tons of plastic daily (353 million metric tons a year), with only 9% recycled, over half in landfill and the balance in mismanaged waste or incinerated, creating further problems for human health and the environment. Plastics also generate about 1.7 gigatons of greenhouse gas emissions annually, they note, on par with aviation. The WRI authors detail the differences among macro-, micro- and nanoplastics, and outline both upstream (restrictions, labeling) and downstream (deposit schemes, recognition of informal recyclers) solutions. They also break down the geopolitics of plastic and summarize the background on negotiations since the U.N. Environment Assembly called for a treaty in March 2022.

Read here and here.

4.

Here is a fan favorite among MSCI’s recent research: A company’s sustainability profile carries a risk premia not explained by traditional credit risk factors. Using a decade of data history ending in December 2024, my colleagues Xinxin Wang, Jakub Malich and Anett Husi show in The Financial Materiality of Sustainability Risk in Credit Markets: A Decade of Evidence that bonds from high-ESG-rated issuers had, on average, lower credit spreads than bonds from low-rated issuers. The research builds on MSCI’s 2021 analysis, and confirms that over the extended period, issuers with high ESG ratings exhibited higher gross margin, greater debt-servicing capacity supported by cash flows from operations, and higher credit quality. The authors take the analysis further by isolating the incremental effect of sustainability after controlling for traditional risk factors (such as credit quality, duration and liquidity), indicating that sustainability-related risks were not fully captured in common credit quality measures.

Read here.

5.

The views of the American public on climate change and the latest scientific consensus align in some respects and not in others. But both differ in important ways from the narrative that dominates the media and public discourse on climate. That’s the conclusion of The Science vs. the Narrative vs. the Voters: Clarifying the Public Debate Around Energy and Climate, a report by Roger Pielke Jr. and Ruy Teixeira of the American Enterprise Institute. Their 2024 survey of 3,000 registered U.S. voters found broad support for an “all-of-the-above” approach to energy policy rather than a rapid elimination of fossil fuels. Seventy-one percent of voters backed a mix of energy sources, including both fossil fuels and renewables. Nearly three-quarters (73%) ranked either the cost of energy or the availability of power as most important, compared with just 19% who named climate impacts as their top concern. That contrasts with a media narrative that, while increasingly open to nuclear and other clean-energy technologies, treats “a rapid transition to clean energy, especially renewables, as an article of faith,” the authors write. You might agree or disagree. In Eating Humble Pie in the Climate Change Debate, Robert G. Eccles, a visiting professor at Oxford University and longtime sustainability leader, offers one place to join the debate.

Read here and here.