Scale AI

Meta and Nvidia push open-weight AI models

Meta and Nvidia released open-weight AI models in August 2026 as part of a broader U.S. effort to compete with leading Chinese AI labs. Meta published Muse Glimmer and said it would open weights for Muse Spark 1.2; Nvidia released Nemotron 3.5 Lightning and described its Nemotron family as “truly open source,” publishing related training datasets, techniques, and model weights. More than 20 U.S. tech companies had recently urged policymakers to avoid premature restrictions on open-weight models. Industry figures — including Box CEO Aaron Levie and analysts at Forrester and D.A. Davidson — said the moves restore U.S. presence in the open-source foundation-model ecosystem but noted challenges winning developer trust after prior proprietary shifts.

Meta Reports Q2 Results After Market Close

Meta is scheduled to report second-quarter results after the market close on July 29, 2026, with Street/analyst estimates forecasting $7.22 adjusted EPS and $60.17 billion in revenue (about 26% year-over-year growth). Investors will focus on capital expenditures as Meta pours money into AI infrastructure and data centers while trying to monetize its AI models and services. The company recently debuted Muse Spark 1.1 and released Muse Image, and has pursued large AI investments including a $14.3 billion deal tied to Scale AI and a $14 billion BlackRock data-center venture. Analysts expect Reality Labs to post a loss of about $5.07 billion on $423.4 million in revenue. Meta’s capex guidance and plans to sell/lease AI capacity to third parties (Anthropic in preliminary talks) are central to investor scrutiny.

TIME Migrates to Contentstack to Power AI Content

TIME has partnered with Contentstack and migrated its digital publishing infrastructure onto the Contentstack Agentic Experience Platform (AXP). The migration — reported as a move to an API-first, headless CMS foundation — aimed to improve editorial operations, speed content delivery, and enable AI-driven audience experiences. TIME leaders say the new platform reduced day-to-day engineering dependency while enabling features such as site registration for personalization, a TIME Games product, new newsletters, and enhanced video experiences. Michael Mraz (General Manager of B2C & Chief Product Officer, TIME) praised the platform’s balance between developer flexibility and newsroom workflow; Neha Sampat (Founder & CEO, Contentstack) framed the move as a pragmatic, foundation-first approach to AI adoption. Michael Mraz is scheduled to discuss the strategy at ContentCon later this year.