💬 Live voice translation has historically struggled with high latency. Grab, Southeast Asia’s leading super app, connects millions of people every day with driver- and merchant-partners across the region. With users from around the world, language barriers are common. To address this, Grab is partnering with Google DeepMind to test Gemini 3.5 Live Translate during pickups between travelers and driver-partners. Integrated directly into in-app calls, this speech-to-speech model provides low-latency, two-way translation in over 70 languages. Crucially, it preserves the speakers' natural rhythm, pitch, and intonation. Travellers and drivers make over 10 million monthly voice calls to coordinate pickups through the Grab app. Overcoming language barriers is crucial for seamless journeys! 🗣️ We’re thrilled to explore solutions that could enable better, real-time communication for diverse users across Southeast Asia. Read more about Gemini 3.5 Live Translate here → https://lnkd.in/eQ8ji72n
Google Cloud
Software Development
Mountain View, California 3,278,584 followers
The new way to cloud.
About us
Google Cloud accelerates every organization’s ability to digitally transform its business and industry. We deliver enterprise-grade solutions that leverage Google’s cutting-edge technology, and tools that help developers build more sustainably. Customers in more than 200 countries and territories turn to Google Cloud as their trusted partner to enable growth and solve their most critical business problems.
- Website
-
https://cloud.google.com/
External link for Google Cloud
- Industry
- Software Development
- Company size
- 10,001+ employees
- Headquarters
- Mountain View, California
Updates
-
Claude Fable 5, Anthropic's latest frontier model, is now generally available to all customers in Agent Platform. Discover how enterprise teams can start building at scale with Claude Fable 5 on Google Cloud today → https://goo.gle/4usenhN
-
-
Chip-level price-performance metrics look great on paper, but ML teams have to measure the real-world productivity of their systems, or ‘goodput’. Imagine a manufacturing line. Productivity isn't measured by how fast the conveyor belt spins. If you’re constantly waiting for raw materials, stopping due to machine malfunctions, or your staff are making mistakes, you produce fewer products. Goodput works the same way. 🗓️ Scheduling goodput: The fraction of time all required accelerators are fully available to run the job. 🔄 Runtime goodput: Given those resources, the fraction of time spent making active forward progress rather than dealing with node failures, compilation delays, or checkpoint restarts. 💻 Program goodput: The fraction of peak hardware performance (Model Flops Utilization) that your training code actually extracts from the silicon. Our TPU 8 chips can achieve up to 97% goodput, so nearly every compute cycle is spent productively. Watch the full interview with Amin Vahdat → https://goo.gle/442BOmU #8daysofTPU8
-
On this week’s livestream, see how GPUs are revolutionizing tabular ML and virtual drug screening, featuring experts from Google Cloud and NVIDIA AI.
GPU-accelerated virtual drug screening with cuML and Agent Platform
www.linkedin.com
-
Get ready for #8daysofTPU8! A daily digest where we'll share the technical details and engineering choices powering the next frontier of AI infrastructure—plus resources, reference architectures, and more! Today's topic: Predictions for the Future of AI Infrastructure. Hear from Amin Vahdat, SVP and Chief Technologist, AI and Infrastructure at Google, in this Next '26 session on demand → https://goo.gle/4uyR6Lm
-
-
Announcing our new strategic partnership with IBM to scale AI with human expertise and AI‑powered delivery. Learn how we're expanding IBM Consulting Advantage, an AI‑powered delivery platform, with industry‑specific agents for Gemini Enterprise → https://goo.gle/4xwpo4y
-
-
At BASF Agricultural Solutions, a single product can take up to two years to create. To map this massive 180-site network, it deployed AlphaEvolve on Google Cloud. This isn’t standard AI. It’s an evolutionary coding agent that autonomously writes and refines its own code. By running thousands of experiments, it successfully learned to replicate the intuitive, real-world decisions of expert human planners. The results of initial experiments with this self-evolving digital twin: 🔥 80% higher accuracy over baseline models 🔥 Automated discovery of human-readable supply chain rules 🔥 Global visibility to optimize inventory and confidently prevent stockouts Evolutionary AI is driving real enterprise operations today. Read the story → https://goo.gle/4ohcZ06
-
-
Build smarter agents, together. This demo shows how projects in Gemini Enterprise provide a persistent space for your team and agents to work together on a project, with specific data, shared conversations, and a collaborative canvas → https://goo.gle/4dSEc5D
-
Papa Johns is the first quick-service restaurant to adopt our Food Ordering AI Agent to provide personalized recommendations and voice-powered ordering for 150 million customers. By deploying this multimodal agent on top of its existing Google Cloud data infrastructure, including BigQuery and Cloud SQL, Papa Johns can now process voice to build shopping carts across mobile devices. Papa Johns uses BigQuery to power micro-segmentation for tailored interactions, and the new agent to automatically apply relevant coupons to every order. Plus, Gemini Enterprise Agent Platform helps track key metrics such as conversion rates and cart abandonment. The result is a more efficient ordering process that provides the fast, simple digital transactions that modern pizza consumers expect. Hear from Kevin Vasconi ↓
-
With no coding skills required, build an agent in the Gemini Enterprise app that can research the latest trends and write marketing content that sounds exactly like your brand every time. Now you can generate innovative content, while taking the manual work out of staying on-brand. Learn more → https://goo.gle/4uOC3xT