AI Week in Review: Apple, Google, OpenAI, Nvidia Innovations & Funding
The past week has seen a flurry of significant advancements and strategic moves across the artificial intelligence landscape, with major players like Apple, Google, OpenAI, YouTube, and Nvidia driving innovation, alongside substantial funding injections into promising AI startups.
Apple is poised for a transformative leap with its voice assistant, Siri, as it trials a more capable version ahead of a comprehensive overhaul slated for spring 2026. This revamp aims to evolve Siri from a passive information-retrieval tool into an active agent capable of executing multi-step tasks across various third-party applications. Leveraging an upgraded App Intents framework, developers can integrate deep app functionalities, allowing users to control popular services like Uber, Threads, YouTube, Amazon, and WhatsApp purely through natural voice commands. While this promises unprecedented hands-free control, Apple is meticulously addressing privacy concerns and voice recognition accuracy, potentially limiting initial functionality in sensitive areas such as banking and health applications. This strategic move positions Apple to intensify its competition with rivals like Google Gemini and Amazon Alexa in the burgeoning voice-driven AI market.
Meanwhile, OpenAI’s highly anticipated GPT-5 model began its rollout to users around August 7, 2025. The launch, however, was not without its initial hitches, as OpenAI acknowledged issues with a broken routing functionality that made the model appear “dumber” than intended, though this has reportedly been fixed. Despite OpenAI’s claims of improved accuracy and reliability, early user feedback has been mixed, with some reporting an “underwhelming” experience and a “loss of personality” compared to previous iterations like GPT-4o. Concerns have also been raised about the model’s performance in certain benchmarks and its context window, which remains smaller than some competitors, including Google’s Gemini. Nevertheless, OpenAI reports that nearly 700 million people are now using ChatGPT weekly, underscoring the widespread adoption of AI in daily life and work.
In a move to enhance online safety for minors, YouTube commenced rolling out an AI-powered age verification system in the U.S. on August 13, 2025. This system utilizes machine learning to estimate a user’s age by analyzing various signals, including video watch history, search patterns, and account longevity, rather than solely relying on self-declared birthdates. If the AI determines a user is under 18, standard protections for teen accounts are automatically enforced, such as displaying only non-personalized advertisements, activating digital wellbeing tools like “take a break” reminders, and showing privacy prompts. Users who believe they have been misidentified have the option to verify their age through a government-issued ID, a selfie, or a credit card. This initiative aligns with a broader global push to fortify online safeguards for young people, preempting potentially more stringent regulations.
Nvidia, a pivotal force in AI infrastructure, unveiled a comprehensive suite of new AI models, libraries, and computing tools designed to accelerate robotics development at SIGGRAPH 2025. These innovations, primarily focused on “physical AI” applications, include Cosmos Reason, a 7-billion-parameter vision-language model enabling robots to reason, understand physics, and predict actions in real-world environments. The company also introduced updated Omniverse libraries, featuring neural reconstruction capabilities for simulating real-world environments in 3D, and enhancements to its Isaac Sim and Isaac Lab robot simulation frameworks. Supported by new RTX Pro Blackwell Servers and expanded DGX Cloud availability, these tools are already being adopted by leading robotics firms and autonomous vehicle developers, signifying Nvidia’s commitment to pushing the boundaries of embodied intelligence.
Google also marked a significant week with several AI-centric announcements. The tech giant expanded its partnership with Oracle, making its advanced Gemini models, including Gemini 2.5, accessible to Oracle Cloud Infrastructure (OCI) customers. This collaboration will extend Gemini’s multimodal capabilities, from video and image generation to specialized industry models, into Oracle’s enterprise applications for finance, HR, and supply chain. Furthermore, Google announced a substantial $9 billion investment in Oklahoma for cloud, AI, and data center expansions. The company is also preparing for a Pixel event on August 20, where it is expected to unveil new AI features, potentially including an upgraded Gemini Image Generation (GEMPIX) and a lightweight “Nano Banana” model. Google DeepMind also introduced Genie 3, a general-purpose world model capable of generating diverse interactive environments.
Beyond the tech giants, the AI startup ecosystem continued to thrive with notable funding rounds. Cohere, a Canadian AI firm specializing in enterprise-focused agentic AI products, secured a substantial US$500 million, elevating its valuation to US$6.8 billion. This oversubscribed round, led by Radical Ventures and Inovia Capital with participation from Nvidia and Salesforce Ventures, is set to fuel Cohere’s global expansion and its development of secure enterprise AI solutions, following its recent launch of the “North” platform. Dubai-based B2B fintech company Alaan raised $48 million in Series A funding, led by Peak XV Partners. Alaan plans to use this capital to expand its AI-powered spend management platform across the MENA region, enhancing features like AI-driven receipt capture and real-time expense tracking. Lastly, Chai Discovery, a San Francisco-based AI company focused on molecular design for therapeutics, announced a $70 million Series A round led by Menlo Ventures. This funding, bringing their total to $100 million, follows their breakthrough with the Chai-2 model, which achieves a near-20% hit rate in de novo antibody design, a significant leap in accelerating drug discovery.