Amazon Considers Global Product Graph With 2027 Target Date

Key Takeaways

• Amazon is considering a global World Product Graph that would include products it does not sell; draft documents give October 1, 2027, as an aspirational launch date, but Amazon says the documents are outdated.
• Coca-Cola and Bayer agreed to contribute product data to the proposed Amazon database, while Procter & Gamble discussed a trial.
• Amazon’s Alexa for Shopping, formerly Rufus, has faced reported accuracy issues; a draft document puts its volume at about 83 million searches weekly, while Amazon says more than 400 million customers used its AI shopping help over the past year.
• New York City lawmakers want Google to explain default restrictions on Search and AI products for K-12 accounts, including access without generative AI, data protections and misinformation safeguards; Google says schools control whether to enable Gemini, Gemini Notebook and Search.
• Researchers’ Probability-Ranked Valid State Search method uses a compact, confidence-based set of likely states to help AI agents plan around uncertain visual predictions.
• JD.com reported more than 50 million AI products sold during its 2026 Singles’ Day event.
• Paessler recommends AI spending caps and cross-functional reviews after a $5,000 development prize led to 29 submissions and unexpectedly high AI costs.
• Cloud native developers exceed 21 million globally; about 75% use AI, while maintainers report a growing volume of low-quality AI-generated contributions.
• A report says Jensen Huang’s exit as Nvidia CEO removes a prominent Washington opponent of tighter AI rules and advocate for AI chip sales to China and open-source AI.
• SAG-AFTRA’s Duncan Crabtree-Ireland says more than 90% of performers affected by misuse of AI likenesses mainly want the content removed; Keysight’s electronica 2026 launches include 1,500-volt, 90-kilowatt test equipment and the XR3 oscilloscope.

Amazon proposes a global product database to compete with AI shopping agents

Amazon is considering a World Product Graph that would gather information on products sold worldwide, including items Amazon does not carry. An internal document says brands and retailers could provide data, while Amazon could also collect information from other retail websites. Coca-Cola and Bayer have agreed to contribute data, and Procter & Gamble took part in discussions about a trial. The document lists October 1, 2027, as an aspirational launch date and says Amazon’s shopping assistant has given many incorrect or incomplete answers. Amazon called the document an outdated brainstorming memo and said its reported usage figures were significantly outdated.

Amazon explores a worldwide product database to challenge Google and AI shopping agents

Amazon is exploring a World Product Graph that would collect information on nearly every product sold worldwide, whether or not Amazon sells it. Internal documents list October 1, 2027, as a planned launch date, though Amazon said the old draft documents may not reflect its current plans. The project aims to help shoppers research and compare products without leaving Amazon for Google, brand websites, and review sites. The documents say Coca-Cola and Bayer agreed to provide product data and Procter & Gamble discussed a trial. They also acknowledge accuracy problems with Amazon’s Alexa for Shopping, previously called Rufus, which handles about 83 million searches a week, according to the documents.

Amazon considers listing products it does not sell in AI shopping push

Amazon is considering a World Product Graph to collect details about products it does not sell, including specifications, release dates, and color options. Reuters reported that the draft proposal gives October 2027 as a target date, while Amazon said the document was outdated and may not reflect current plans. The project could help Amazon compete for product research, which Morgan Stanley’s 2025 research found began on Google for 57% of online shoppers and on Amazon for 16%. When shoppers already knew what they wanted, 41% started on Amazon and 23% on Google. Amazon says more than 400 million customers worldwide used its AI shopping help over the past year.

Amazon explores a wider product catalog to compete with shopping rivals

Amazon is exploring a World Product Graph to collect information about nearly every product for sale, including products it does not carry. An internal document gives October 1, 2027, as a possible release date, but Amazon said draft documents may not reflect its current plans. The project aims to reduce the need for shoppers to use Google, manufacturer websites, and forums to compare products. A larger product data set could also help Amazon improve its AI shopping answers about models, colors, and which products may suit a shopper. The effort is intended to help Amazon compete with search engines, online retailers, and AI shopping assistants.

Amazon plans global product database to counter AI shopping agents

Amazon is considering a World Product Graph to gather information on nearly every product for sale, including products it does not sell, according to an internal document reviewed by Reuters. The plan aims to help shoppers research products on Amazon instead of turning to Google, brand websites, and review sites. It could also help Amazon compete with Walmart and other online retailers, as well as AI agents that shop for users. The document says Amazon’s own AI search agent often gives inaccurate information. Reuters could not confirm whether or when Amazon will put the plan into action, though the document lists October 1, 2027, as an aspirational release date.

NYC Council demands answers from Google over AI access in schools

New York City Council Speaker Julie Menin and Council Member Shekar Krishnan sent Google a letter on Nov. 12 asking about changes to AI access in schools. Google plans to turn off Search and several AI products by default for K-12 student accounts early next year, until school leaders enable them. The lawmakers question whether students can use Search without generative AI and ask about data collection, misinformation safeguards, and pressure to adopt Google AI. The dispute comes as Mayor Zohran Mamdani’s policy limits most student-facing generative AI through eighth grade. Google says the changes give educators more control and plans to test Guided Learning in Search at selected U.S. schools.

NYC leaders challenge Google over school AI restrictions

New York City Council Speaker Julie Menin and Council Member Shekar Krishnan are asking Google how its AI plans could affect students’ access to Search. Google said schools can choose whether to enable Gemini, Gemini Notebook, and Search, but a company leader’s comments raised concerns that tighter AI rules could limit Search access. Menin and Krishnan say Google’s products, including Classroom, Docs, Search, and Chromebooks, are widely used in city schools. They want to know whether students can use Search without AI and what protections cover accuracy and student data. The lawmakers argue that Google should respect the city’s technology decisions rather than pressure schools to adopt AI.

New method helps AI agents plan around uncertain visual predictions

Researchers propose a way for AI agents to handle uncertainty from pretrained perception models such as vision-language models. Instead of relying on one best guess or tracking every possible state, their method builds a minimal set of likely states that meets a chosen confidence threshold. The approach estimates state probabilities from feature predictions and task rules while keeping a compact belief representation. The researchers’ algorithm, Probability-Ranked Valid State Search, has correctness guarantees and polynomial complexity in the unconstrained case. Tests using a vision-language model found that the method scaled with problem size and improved decisions in task planning.

JD.com sells more than 50 million AI products during Singles’ Day

JD.com sold more than 50 million AI products during its 2026 Singles’ Day event, according to the article. The report was published on Nov. 12, 2026. The article also lists company and investor rating information, including a last closing price of 183.51 CNY and an average target price of 259.13 CNY. It gives a mean analyst consensus of Buy based on 35 analysts.

Why companies need firm limits on AI spending

Paessler says companies should set AI spending limits and review whether the costs create real value. The company saw expenses rise after a development event offered a $5,000 prize for new features, leading to 29 submissions and larger-than-expected bills for AI prompts, tokens, and agents. Paessler recommends setting group or individual limits when tools are introduced and matching token limits to spending caps. Caps can trigger a review: managers may approve more spending for critical work or pause it when use is not tied to a business need. The article says a cross-functional AI team can help control costs and prevent unapproved AI use.

AI Could Make Smaller Mortgage Loans Worthwhile for Lenders

Artificial intelligence could help lenders make smaller mortgage loans more profitable and serve borrowers who have often been too costly or difficult to reach. AI can assess a borrower’s ability to repay more quickly and accurately, while helping lenders match people with mortgage products that fit their needs. Automating lending tasks could also lower costs and let staff focus more on customer service and relationships. The article says wider use of AI may create new opportunities for both lenders and borrowers.

Cloud Native Developers Top 21 Million and Lead in AI Use

The global number of cloud native developers has passed 21 million, and about 75% use AI in some form, compared with about 55% of other developers. Research discussed by Killen and Bollmann-Dodd suggests cloud native tools are helping developers and organizations adopt AI. But maintainers say they are struggling with a flood of AI generated pull requests and issues, many of which they consider low quality. Bollmann-Dodd also expects developers to pay closer attention to the cost of models and tokens. The Linux Foundation sponsored the SiliconANGLE and theCUBE segment, but had no editorial control over its content.

Jensen Huang’s Exit Leaves Washington Without a Key AI Policy Voice

Nvidia co-founder Jensen Huang’s departure as chief executive leaves a gap in Washington, where he was a strong opponent of tighter AI rules. He pushed to allow AI chip sales to China, argued that safety regulations could hurt U.S. companies, and supported open source AI technology. Now, bipartisan anger over allegations of his affair with a Chinese spy is putting his policy agenda and its supporters under pressure. Democrats who have promised stronger AI regulation won elections last week, while more industry voices warn that AI could threaten public safety and disrupt the economy.

SAG-AFTRA Leader Says Performers Want Misused AI Images Taken Down

SAG-AFTRA leader Duncan Crabtree-Ireland said more than 90% of performers who face misuse of their likeness through AI mainly want the content removed, not damages. Speaking at Web Summit in Lisbon, he backed the proposed No Fakes Act as a way to give people control over their image, voice, and words. Crabtree-Ireland said many members believe AI can be used alongside performers if basic rules protect them, though he estimated 10% to 15% want the union to block it. Some members want AI tools to support their creativity, but he said they do not want to be replaced. He also stressed the difference between people and AI-generated characters.

Keysight Unveils AI and Power Testing Products at electronica 2026

Keysight Technologies introduced new test and measurement products at electronica 2026 for AI infrastructure, communications, power electronics, and cybersecurity. Its expanded high power automated test equipment supports up to 1,500 volts and configurations reaching 90 kilowatts, for uses including EV charging, aerospace systems, and large batteries. The company also announced a collaboration with Tower Semiconductor to certify photonic design tools and launched the compact XR3 oscilloscope for scalable multi channel testing. Keysight said the products add to its ability to test complex systems as AI data centers, vehicles, and aircraft use denser, higher voltage designs. Investors are also watching software and services revenue, annual recurring revenue, and orders for AI infrastructure, 1.6T interconnect, and EV power testing.

Sources

• Amazon plans a 'World Product Graph' to own the world's product data and fend off AI shopping agents, Reuters reports
• Amazon's "World Product Graph" Ambition: Owning Global Product Data to Counter AI Shopping Agents
• Amazon Eyes Products It Doesn't Sell in AI Shopping Tussle
• Amazon plans product data expansion to counter AI shopping rivals
• Exclusive-Amazon’s plan to own the world’s product data and fight AI shopping agents
• Google’s AI push could put NYC students’ search access at risk, Council leaders warn
• City Leaders Press Google Over AI Schools Plan
• Robust Task-Aware State Estimation from Pre-trained Perception
• JD.com Sells Over 50 Million AI Products During 2026 Singles' Day
• Saying no to AI spend is how we protect AI value
• Can AI Make Smaller Mortgage Loans Profitable?
• Cloud-native developers top 21 million and lead in AI adoption
• With Jensen Huang Gone, the A.I. Industry Loses Its Most Powerful Backer in Washington
• SAG-AFTRA Boss Says AI Takedowns Matter More Than Damages To Performers
• Keysight Technologies (KEYS) Rolls Out AI And Power Test Products At electronica 2026