Google Cloud Next 2025 has been an event full of announcements and innovations in the field of artificial intelligence and cloud technology, where the company aimed to demonstrate its continuous progress to make AI more powerful, accessible, and easy to use. Let us review the most relevant announcements.
Among the main novelties is the arrival of Gemini 2.5 Flash to Vertex AI, which aims to prioritize efficiency, allowing developers to adjust the speed, precision, and cost according to their needs. Google presents it as ideal for high-volume and cost-sensitive applications, offering an economical alternative to premium models by sacrificing some precision.
As a “reasoning” model, it verifies facts, thus increasing response time. Google focuses it on real-time applications such as customer service and document analysis. Optimized for low latency and reduced cost, it can be key for virtual assistants and real-time summaries.
Google announced a series of revolutionary innovations for its Vertex AI platform at its Cloud Next event, aimed at solidifying it as a leader in the generation of multimedia content.
The company has introduced artificial intelligence tools capable of creating video, images, audio, and music from simple text prompts:
Chirp 3 presents the ability to create custom voices with just 10 seconds of audio, opening new possibilities in voice content generation. Google highlights its utility for companies seeking to personalize call centers, create accessible content, and establish unique brand voices while maintaining a consistent identity. Furthermore, the new transcription functionality allows separating and accurately identifying different speakers in recordings with multiple participants. Google underlines its utility for improving the clarity and usability of transcriptions in applications such as meeting summaries, podcast analysis, and recordings of calls with multiple interlocutors.
Veo 2, on the other hand, revolutionizes video editing with more precise camera and editing controls, aimed at reducing post-production time and costs. The new features are available in preview with a waiting list.
Lyria, the text-to-music model, allows the generation of complete, production-ready musical pieces, marking a significant advancement in sound content creation. As the company explains, users will be able to “rapidly create soundtracks for marketing campaigns, product launches, or immersive in-store experiences, tailored to the unique identity of your brand. Lyria lets you craft a sound brand that deeply connects with your target audience, fostering emotional connections and enhancing brand recall.”
Additionally, Vertex includes Imagen 3, with enhanced editing capabilities, now including a new way to customize any image. “We have significantly improved the retouching features of Imagen 3 to reconstruct the missing or damaged parts of an image. Our latest update significantly enhances the quality of object removal, offering a more natural and fluid editing experience. Here is an example of how you can quickly remove unwanted objects, imperfections, or distractions from your photos,” explains Google.
These updates make Vertex AI the only platform that integrates generative models for video, images, audio, and music, offering companies and creators an unprecedented set of tools for multimedia content production. Google reaffirms its commitment to AI innovation, facilitating the creation of high-quality content more efficiently and accessibly.
Moreover, Google Workspace is receiving AI-driven updates to enhance productivity in Docs, Sheets, Meet, and Chat.
Gemini 2.5 Pro Experimental, which Google refers to as “the world’s most capable AI model” according to Chatbot Arena, is now available for Gemini Advanced subscribers. Gemini Deep Research operates as a personal AI research assistant, enhanced by the latest smart model to improve analytical reasoning, information synthesis, and research report generation. According to Google, tests have shown that users prefer reports generated by Gemini Deep Research with 2.5 Pro to those from other leading providers in the field of deep research.
Source: Google
Users of Gemini Advanced can access this feature via the web, Android, and iOS to create detailed and easily readable reports on various research topics, potentially saving them considerable time. Additionally, the Audio Overviews feature allows conversion of reports into podcast format for on-the-go listening.
Starting today, Google Cloud customers can grant employees access to unified capabilities of search, analysis, and synthesis of company information, directly from Chrome’s search bar. To facilitate the creation of agents, Google adds the new No-Code Agent Designer, which allows employees without technical training to create custom agents for their workflows and individual needs. Additionally, Agentspace now offers two expert agents developed by Google, Deep Research and Idea Generation, joining previously available tools like NotebookLM for enterprises.
Another of the most notable novelties is the introduction of Ironwood, Google’s seventh-generation TPU, designed for AI inference and offering significant improvements in performance. A specialized chip that Google has designed to speed up the calculations necessary for machine learning. Compared to the previous generation, Ironwood offers five times more maximum computational capacity and six times more high-bandwidth memory capacity. It will be available later this year.
Overall, Google Cloud Next 2025 has demonstrated Google’s commitment to driving innovation in AI and the cloud. With a wide range of announcements and updates, the company seeks to position Google Cloud as a benchmark, offering powerful solutions for businesses of all sizes.
Image: Flux Schnell
Your email address will not be published. Required fields are marked *
Δ