BigTech Blog Posts

Image with multiple TVs in the background. The TV on the left says “Monopoly Go!, the TV in the center says “Four days of Fantastic Rewards”, and the TV on the right says “Marvel Studios’ “The Fantastic Four: First Steps,” only in theaters”.

Abstract image of Web Guide experiment on Google Search

AI-powered search demands high-performance indexing, low-latency retrieval, and seamless scalability. NVIDIA cuVS brings GPU-accelerated vector search and...

How can service leaders adopt AI without losing the human touch? Start with empathy, trust, and a people-first approach.

Three phone screens showing different screens of using virtual clothing try-on

Xinxing Xu is helping shape the work of Microsoft Research Asia – Singapore by turning advanced AI research into real-world solutions. Learn how he collaborates across sectors and disciplines to drive responsible innovation throughout Southeast Asia. The post Xinxing Xu bridges AI research and real-world impact at Microsoft Research Asia – Singapore appeared first on Microsoft Research.

University teams battle to harden and hack AI coding assistants in head-to-head tournament

Read Google and Alphabet CEO Sundar Pichai’s remarks from the Q2 2025 earnings call.

Logan Kilpatrick sits down with Robby Stein, VP of Product to discuss what it takes to enable billions of people to truly ask anything in Search

Aeneas, a new AI model, helps historians interpret, attribute and restore ancient inscriptions.

Animation showing various tools for creation in the Google Photos app including Photo to video and Remix.

The Bad Guys 2 appears in orange, red and black letters. An image of Wolfie is on the right with a speech bubble saying, "Police reported ahead. So don’t mess this up, it’s taken me years to get on their good side."

Semantic Telemetry helps LLMs run efficiently, reliably, and in near real-time. Learn about the engineering behind that system, including the trade-offs and lessons learned along the way—from batching strategies to token optimization and orchestration. The post Technical approach for classifying human-AI interactions at scale appeared first on Microsoft Research.

We’re publishing a paper in Nature introducing Aeneas, the first AI model for contextualizing ancient inscriptions.

The PDF is among the most common file formats for sharing information such as financial reports, research papers, technical documents, and marketing materials....

The process of converting vast libraries of text into numerical representations known as embeddings is essential for generative AI. Various technologies—from...

We researched the top challenges small businesses like yours face — and found solutions to help you overcome them faster.

Three people engaged in conversation sit in chairs at a Google for Startups Campus.

Today, in partnership with the Earth Fire Alliance and Muon Space, we shared the first images of wildfires detected by FireSat, a revolutionary new satellite constellati…

Today the CMA has announced its proposed decision to designate Google’s ‘Mobile Platform’ (Android, Play, Chrome, Blink) with ‘strategic market status’, along with a roa…

Have you ever wanted to build your own reasoning model but thought it was too complicated or required massive resources? Think again. With NVIDIA’s powerful...

Try the new 1T-parameter open source MoE LLM today.

Gemini 2.5 Flash-Lite, previously in preview, is now stable and generally available. This cost-efficient model provides high quality in a small size, and includes 2.5 family features like a 1 million-token context window and multimodality.

The NVIDIA Collective Communications Library (NCCL) is essential for fast GPU-to-GPU communication in AI workloads, using various optimizations and tuning to...

Try the new 1T-parameter open source MoE LLM today.

This post explores a promising direction for building dynamic digital representations of the physical world, a topic gaining increasing attention in recent...

Generating diverse synthetic prior distributions leads to a tabular foundation model that outperforms task-specific baselines.

Earn Salesforce foundational certifications to validate your knowledge in Agentforce, marketing, or Salesforce fundamentals.

Our bulletin covering coordinated influence operation campaigns terminated on our platforms in Q2 2025.

In the series finale, Amanda Craig Deckard returns to examine what Microsoft has learned about testing as a governance tool. She also explores the roles of rigor, standardization, and interpretability in testing and what’s next for Microsoft’s AI governance work. The post AI Testing and Evaluation: Reflections appeared first on Microsoft Research.