Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More<\/em><\/p>\n\n\n\n Google Cloud unveiled its seventh-generation Tensor Processing Unit (TPU) called Ironwood on Wednesday, a custom AI accelerator that the company claims delivers more than 24 times the computing power of the world\u2019s fastest supercomputer when deployed at scale.<\/p>\n\n\n\n The new chip, announced at Google Cloud Next \u201925, represents a significant pivot in Google\u2019s decade-long AI chip development strategy. While previous generations of TPUs were designed primarily for both training and inference workloads, Ironwood is the first purpose-built specifically for inference \u2014 the process of deploying trained AI models to make predictions or generate responses.<\/p>\n\n\n\n \u201cIronwood is built to support this next phase of generative AI and its tremendous computational and communication requirements,\u201d said Amin Vahdat, Google\u2019s Vice President and General Manager of ML, Systems, and Cloud AI, in a virtual press conference ahead of the event. \u201cThis is what we call the \u2018age of inference\u2019 where AI agents will proactively retrieve and generate data to collaboratively deliver insights and answers, not just data.\u201d<\/p>\n\n\n\n The technical specifications of Ironwood are striking. When scaled to 9,216 chips per pod, Ironwood delivers 42.5 exaflops of computing power \u2014 dwarfing El Capitan\u2018s 1.7 exaflops, currently the world\u2019s fastest supercomputer. Each individual Ironwood chip delivers peak compute of 4,614 teraflops.<\/p>\n\n\n\n Ironwood also features significant memory and bandwidth improvements. Each chip comes with 192GB of High Bandwidth Memory (HBM), six times more than Trillium, Google\u2019s previous-generation TPU announced last year. Memory bandwidth reaches 7.2 terabits per second per chip, a 4.5x improvement over Trillium.<\/p>\n\n\n\n Perhaps most importantly in an era of power-constrained data centers, Ironwood delivers twice the performance per watt compared to Trillium, and is nearly 30 times more power efficient than Google\u2019s first Cloud TPU from 2018.<\/p>\n\n\n\n \u201cAt a time when available power is one of the constraints for delivering AI capabilities, we deliver significantly more capacity per watt for customer workloads,\u201d Vahdat explained.<\/p>\n\n\n\n The emphasis on inference rather than training represents a significant inflection point in the AI timeline. For years, the industry has been fixated on building increasingly massive foundation models, with companies competing primarily on parameter size and training capabilities. Google\u2019s pivot to inference optimization suggests we\u2019re entering a new phase where deployment efficiency and reasoning capabilities take center stage.<\/p>\n\n\n\n This transition makes sense. Training happens once, but inference operations occur billions of times daily as users interact with AI systems. The economics of AI are increasingly tied to inference costs, especially as models grow more complex and computationally intensive.<\/p>\n\n\n\n During the press conference, Vahdat revealed that Google has observed a 10x year-over-year increase in demand for AI compute over the past eight years \u2014 a staggering factor of 100 million overall. No amount of Moore\u2019s Law progression could satisfy this growth curve without specialized architectures like Ironwood.<\/p>\n\n\n\n What\u2019s particularly notable is the focus on \u201cthinking models\u201d that perform complex reasoning tasks rather than simple pattern recognition. This suggests Google sees the future of AI not just in larger models, but in models that can break down problems, reason through multiple steps, and essentially simulate human-like thought processes.<\/p>\n\n\n\n Google is positioning Ironwood as the foundation for its most advanced AI models, including Gemini 2.5, which the company describes as having \u201cthinking capabilities natively built in.\u201d<\/p>\n\n\n\n At the conference, Google also announced Gemini 2.5 Flash, a more cost-effective version of its flagship model that \u201cadjusts the depth of reasoning based on a prompt\u2019s complexity.\u201d While Gemini 2.5 Pro is designed for complex use cases like drug discovery and financial modeling, Gemini 2.5 Flash is positioned for everyday applications where responsiveness is critical.<\/p>\n\n\n\n The company also demonstrated its full suite of generative media models, including text-to-image, text-to-video, and a newly announced text-to-music capability called Lyria. A demonstration showed how these tools could be used together to create a complete promotional video for a concert.<\/p>\n\n\n\n Ironwood is just one part of Google\u2019s broader AI infrastructure strategy. The company also announced Cloud WAN, a managed wide-area network service that gives businesses access to Google\u2019s planet-scale private network infrastructure.<\/p>\n\n\n\n \u201cCloud WAN is a fully managed, viable and secure enterprise networking backbone that provides up to 40% improved network performance, while also reducing total cost of ownership by that same 40%,\u201d Vahdat said.<\/p>\n\n\n\n Google is also expanding its software offerings for AI workloads, including Pathways, its machine learning runtime developed by Google DeepMind. Pathways on Google Cloud allows customers to scale out model serving across hundreds of TPUs.<\/p>\n\n\n\n These hardware and software announcements come at a crucial time for Google Cloud, which reported $12 billion in Q4 2024 revenue, up 30% year over year, in its latest earnings report.<\/p>\n\n\n\n The economics of AI deployment are increasingly becoming a differentiating factor in the cloud wars. Google faces intense competition from Microsoft Azure, which has leveraged its OpenAI partnership into a formidable market position, and Amazon Web Services, which continues to expand its Trainium and Inferentia chip offerings.<\/p>\n\n\n\n What separates Google\u2019s approach is its vertical integration. While rivals have partnerships with chip manufacturers or acquired startups, Google has been developing TPUs in-house for over a decade. This gives the company unparalleled control over its AI stack, from silicon to software to services.<\/p>\n\n\n\n By bringing this technology to enterprise customers, Google is betting that its hard-won experience building chips for Search, Gmail, and YouTube will translate into competitive advantages in the enterprise market. The strategy is clear: offer the same infrastructure that powers Google\u2019s own AI, at scale, to anyone willing to pay for it.<\/p>\n\n\n\n Beyond hardware, Google outlined a vision for AI centered around multi-agent systems. The company announced an Agent Development Kit (ADK) that allows developers to build systems where multiple AI agents can work together.<\/p>\n\n\n\n Perhaps most significantly, Google announced an \u201cagent-to-agent interoperability protocol\u201d (A2A) that enables AI agents built on different frameworks and by different vendors to communicate with each other.<\/p>\n\n\n\n \u201c2025 will be a transition year where generative AI shifts from answering single questions to solving complex problems through agented systems,\u201d Vahdat predicted.<\/p>\n\n\n\n Google is partnering with more than 50 industry leaders, including Salesforce, ServiceNow, and SAP, to advance this interoperability standard.<\/p>\n\n\n\n For enterprises deploying AI, these announcements could significantly reduce the cost and complexity of running sophisticated AI models. Ironwood\u2019s improved efficiency could make running advanced reasoning models more economical, while the agent interoperability protocol could help businesses avoid vendor lock-in.<\/p>\n\n\n\n The real-world impact of these advancements shouldn\u2019t be underestimated. Many organizations have been reluctant to deploy advanced AI models due to prohibitive infrastructure costs and energy consumption. If Google can deliver on its performance-per-watt promises, we could see a new wave of AI adoption in industries that have thus far remained on the sidelines.<\/p>\n\n\n\n The multi-agent approach is equally significant for enterprises overwhelmed by the complexity of deploying AI across different systems and vendors. By standardizing how AI systems communicate, Google is attempting to break down the silos that have limited AI\u2019s enterprise impact.<\/p>\n\n\n\n During the press conference, Google emphasized that over 400 customer stories would be shared at Next \u201925, showcasing real business impact from its AI innovations.<\/p>\n\n\n\n As AI continues to advance, the infrastructure powering it will become increasingly critical. Google\u2019s investments in specialized hardware like Ironwood, combined with its agent interoperability initiatives, suggest the company is positioning itself for a future where AI becomes more distributed, more complex, and more deeply integrated into business operations.<\/p>\n\n\n\n \u201cLeading thinking models like Gemini 2.5 and the Nobel Prize winning AlphaFold all run on TPUs today,\u201d Vahdat noted. \u201cWith Ironwood we can\u2019t wait to see what AI breakthroughs are sparked by our own developers and Google Cloud customers when it becomes available later this year.\u201d<\/p>\n\n\n\n The strategic implications extend beyond Google\u2019s own business. By pushing for open standards in agent communication while maintaining proprietary advantages in hardware, Google is attempting a delicate balancing act. The company wants the broader ecosystem to flourish (with Google infrastructure underneath), while still maintaining competitive differentiation.<\/p>\n\n\n\n How quickly competitors respond to Google\u2019s hardware advancements and whether the industry coalesces around the proposed agent interoperability standards will be key factors to watch in the months ahead. If history is any guide, we can expect Microsoft and Amazon to counter with their own inference optimization strategies, potentially setting up a three-way race to build the most efficient AI infrastructure stack.<\/p>\n Daily insights on business use cases with VB Daily<\/strong><\/p>\n If you want to impress your boss, VB Daily has you covered. We give you the inside scoop on what companies are doing with generative AI, from regulatory shifts to practical deployments, so you can share insights for maximum ROI.<\/p>\n \n\t\t\t\t\tThanks for subscribing. Check out more VB newsletters here.\n\t\t\t\t<\/p>\n An error occured.<\/p>\n<\/p><\/div>\n
\n<\/div>Shattering computational barriers: Inside Ironwood\u2019s 42.5 exaflops of AI muscle<\/h2>\n\n\n\n
From model building to \u2018thinking machines\u2019: Why Google\u2019s inference focus matters now<\/h2>\n\n\n\n
Gemini\u2019s thinking engine: How Google\u2019s next-gen models leverage advanced hardware<\/h2>\n\n\n\n
Beyond silicon: Google\u2019s comprehensive infrastructure strategy includes network and software<\/h2>\n\n\n\n
AI economics: How Google\u2019s $12 billion cloud business plans to win the efficiency war<\/h2>\n\n\n\n
The multi-agent ecosystem: Google\u2019s audacious plan for AI systems that work together<\/h2>\n\n\n\n
Enterprise reality check: What Ironwood\u2019s power and efficiency mean for your AI strategy<\/h2>\n\n\n\n
The silicon arms race: Will Google\u2019s custom chips and open standards reshape AI\u2019s future?<\/h2>\n\n\n\n
![\"\"\/](\"https:\/\/venturebeat.com\/wp-content\/themes\/vb-news\/brand\/img\/vb-daily-phone.png\")
\n\t\t\t\t<\/div>\n<\/p><\/div>\n<\/div>\t\t\t<\/div>\r\n