AI & Dev Digest: Google's New Chips, OpenAI's Enterprise Push, and Firefox's AI-Powered Bug Hunt
Stay ahead with today's top tech news: Google Cloud unveils dual-purpose AI chips, OpenAI partners with Infosys to scale AI, and Mozilla leverages Anthropic's Mythos to squash 271 Firefox bugs. Plus, a look at exciting startups from Google Cloud Next and OpenAI's enhanced image generator.
The AI and software development landscape continues its relentless pace! Today, we're diving into Google Cloud's strategic move to tackle Nvidia with specialized AI chips, OpenAI's new alliance to bring its powerful tools to more businesses, and a groundbreaking use of AI by Mozilla to fortify Firefox's security. We'll also spotlight some innovative startups from Google Cloud Next and explore the enhanced capabilities of OpenAI's image generator.
TL;DR
- Google Cloud has launched TPU 8t and TPU 8i, two new AI chips designed for training and inference, respectively, aiming to outperform prior generations.
- OpenAI has partnered with Infosys to integrate its AI tools, including Codex, into Infosys's Topaz AI platform for enterprise clients.
- Google Cloud Next 2026 revealed a $750 million budget to accelerate AI agent development among partners and highlighted several innovative startups.
- OpenAI's ChatGPT Images 2.0 now boasts "thinking capabilities," allowing it to pull information from the web and generate multiple sophisticated images from a single prompt.
- Mozilla utilized Anthropic's Mythos Preview to identify and fix 271 vulnerabilities in Firefox, marking a significant milestone in AI-assisted cybersecurity.
Google Cloud Launches Two New AI Chips to Compete with Nvidia
Google Cloud has unveiled its eighth generation of custom-built AI chips, known as Tensor Processing Units (TPUs), strategically divided into two specialized offerings. The TPU 8t is engineered for AI model training, while the TPU 8i is specifically designed for inference, which encompasses the ongoing usage of AI models post-deployment.
The company asserts that these new TPUs deliver substantial performance enhancements over their predecessors. Key improvements include up to 3x faster AI model training, an 80% better performance-per-dollar ratio, and the capability to link over 1 million TPUs into a unified cluster. This development aims to provide customers with significantly more compute power at reduced energy consumption and cost.
Google Cloud's dual TPU 8t and TPU 8i chips offer up to 3x faster AI model training and 80% better performance per dollar, signaling a direct challenge to competitors like Nvidia in the specialized AI hardware market.
OpenAI Teams Up with Infosys to Bring AI Tools to More Businesses
OpenAI has forged a partnership with Infosys, an Indian IT services giant, to integrate its artificial intelligence tools, including the notable coding assistant Codex, into Infosys's Topaz AI platform. This collaboration aims to assist Infosys's clientele in modernizing software development, automating workflows, and deploying AI systems at scale.
The initial focus of this integration will be on critical areas such as software engineering, legacy modernization, and DevOps. This strategic move by OpenAI to team up with global IT service providers like Infosys underscores a broader industry trend of scaling AI adoption across various businesses, especially as IT firms face pressures from slowing client spending and the rapid advancements in generative AI.
The partnership between OpenAI and Infosys will embed OpenAI's AI tools, including Codex, into the Topaz AI platform, enabling Infosys's clients to accelerate software development and AI deployment at scale.
The Most Interesting Startups Showcased at Google Cloud Next 2026
At Google Cloud Next 2026 in Las Vegas, Google Cloud made its intention clear: it aims to attract and support AI startups on its platform. To facilitate this, the tech giant announced a significant $750 million budget dedicated to helping its Cloud partners, ranging from startups to major consulting firms, boost sales of AI agents to enterprises.
This substantial funding can be utilized for various initiatives, including Gemini proof-of-concept projects, engagement with Google forward-deployed engineers, cloud credits, and deployment rebates. Google Cloud also highlighted several innovative startups leveraging its infrastructure, such as Lovable, which is expanding its use of Google Cloud for launches, and Synthesia, which uses the platform for its AI avatar video generation platform.
Google Cloud has committed $750 million to support AI agent development among its partners, emphasizing its strategy to become the preferred cloud provider for innovative AI startups.
OpenAI’s Updated Image Generator Can Now Pull Information from the Web
OpenAI has rolled out a significant update to its AI-powered image generator, now branded as ChatGPT Images 2.0, which introduces new "thinking capabilities." This enhancement allows the generator to search the web for information, enabling it to create multiple, more sophisticated images from a single prompt. The update is powered by OpenAI's new GPT Image 2 model.
Subscribers to ChatGPT Plus, Pro, Business, and Enterprise can access these advanced capabilities, which include creating visual explainers from uploaded files and reasoning through image structure before generation. ChatGPT Images 2.0 can now produce up to eight images simultaneously while maintaining consistent characters, objects, and styles, making it ideal for tasks like generating manga pages or social graphics. Furthermore, it supports resolutions up to 2K and various aspect ratios, with significant improvements in generating text in non-Latin languages like Japanese, Korean, Chinese, Hindi, and Bengali.
ChatGPT Images 2.0 now features "thinking capabilities" powered by GPT Image 2, allowing it to leverage web information to generate up to eight consistent and sophisticated images at once, improving text generation and offering 2K resolution.
Mozilla Used Anthropic’s Mythos to Find and Fix 271 Bugs in Firefox
Mozilla has announced that its latest Firefox 150 browser release incorporates protections for 271 vulnerabilities that were identified using early access to Anthropic's Mythos Preview. This initiative highlights the profound impact that new AI models can have on cybersecurity and vulnerability hunting.
According to Bobby Holley, Firefox's Chief Technology Officer, these AI tools have dramatically changed the landscape by providing automated techniques capable of covering the full spectrum of vulnerability-inducing bugs. While acknowledging the significant resources and discipline required to manage the influx of bugs uncovered by AI, Mozilla deems this effort essential for user security, especially given that these powerful capabilities will inevitably be accessible to attackers.
Mozilla leveraged Anthropic's Mythos Preview to uncover and fix 271 vulnerabilities in Firefox, demonstrating the critical role AI can play in rapidly enhancing software security against evolving threats.