It is 2026 and enterprise customers are looking at how to effectively leverage both AI and cloud to advance their company’s position and provide greater value to customers. This week is Google’s annual Google Cloud Next conference in Las Vegas where a whopping 37,000 attendees are converging to talk about cloud, data, AI and infrastructure. Google also expected another two million to attend virtually!
As enterprises evaluate which partners and tools make sense to build their business upon, Google is vying to provide an integrated approach to their solutions…and moving up the stack. Let’s take a look at some of the standouts.
Key Takeaways
Google unloaded a truckload of announcements. However, for CIOs and executives, there are a number of standouts.
The Unified Stack
In the era of AI, Google is moving to emphasize the value of an integrated stack that is purpose built for AI. This does not take away from their existing cloud infrastructure, but does show awareness that AI has different needs up and down the stack as compared with classic application architectures.
Instead of a simple layered cake approach, Google is considering how each of the layers in the stack impact the next. From underlying hardware such as their new 8th gen TPUs (more on this in a minute) to AI infrastructure to AI software to platforms to intelligence, customers have the ability to leverage this integrated approach. The infrastructure, which Google calls the AI Hypercomputer, is specifically optimized for AI workloads.
Using a unified stack with awareness up and down the stack also opens the door to greater awareness of complex security threats and mitigations. This is increasing important as AI presents new threats from a myriad of vectors.
The value of this unified approach provides both context and awareness, but also optimization which impacts both performance and cost. As the AI workloads that customers build get more sophisticated, so will the demands on the underlying architecture. Google is projecting ahead with their unified stack approach. At the same time, they are using business outcomes as a North Star to their decisions on the architecture layers all the way down to the chips they’re designing.
8th Gen TPUs
Speaking of chips, Google announced two new TPUs. Yes, TPUs…plural. Since 2013, Google has been creating their own TPU as purpose-built custom silicon to meet new workloads. Their first generation TPU was really just an ASIC. However, two things have happened since. First, in response to the increase in demand and innovation cycles, Google has moved from a 2-year cycle of new versions to annual updates. Second, one solution was not enough to meet to changing demand which led to specialization. In response, Google announced *two* unique TPUs. For the first time, Google has produced two, unique silicon solutions specifically optimized for training (TPU 8t) and inference (TPU8i). Unlike taking a derivative approach, these new TPUs are uniquely designed solutions optimized for their respective purpose.
One thing about TPUs is that they change the price/ performance ratio to provide optimized solutions at a fraction of the cost. Google believes that, over time, we will see an increase in specialized, purpose-built silicon. At the same time, they also see a forthcoming resurgence of CPUs as agentic workflow start to take action.
Gemini Enterprise Agent Platform
Probably one of the most interesting announcements at Google Cloud Next is Google’s new Gemini Enterprise Agent Platform. The platform provide the environment to build, scale, govern and optimize agentic workflows. Enterprises across the spectrum are starting to experiment with building their own agentic workflows. However, self-built agents are not the only agents that will be in use. Google sees three types of agents: Google-produced agents, first-party agents (that the enterprise creates) and third-party agents (that SIs and others publish into the marketplace).
The platform’s ability to provide build and deploy agents, scale agents, govern and secure agents and then monitor and optimize agents provides a cohesive solution for enterprises an end-to-end solution.
Part of Google’s announcements include how their agentic workflows integrate seamlessly with their Workspace products. But it is more than just integration, it’s also context aware across tools and teams.
Gemini Enterprise for Customer Experiences
This one is big for a couple of reasons. First, the solution provides a workflow framework to provide direct interactions with customers through both voice and digital experiences. Think AI workflows meet customer engagement.
Second, not only does it provide a way to build customer engagement workflows, but also demonstrates Google’s move into the customer experience space beyond their existing Workplace applications. This should be a wakeup call for others also in the customer experience space.
By leveraging their insights from data and search, this opens the door to new possibilities in how Google provides tooling to get them closer to engaging with customers. Add in their AI and agentic tooling and it really provides a strong option for customer engagement.
CIO Perspective
This is not the Google of the past and CIOs should take note. Google has long-since been known as a solid cloud platform and tool builder. Today, Google is demonstrating both the depth of their technology chops and pulling it closer than ever to engaging with customers.
For enterprises, the focus on transformation of the business is top of mind for executive teams. AI and technology is the great equalizer and providing new opportunities for enterprises. The challenge is how easy it is to adopt and leverage the technology.
Unlike the Google of the past that built incredibly powerful tools, they are adding new solutions that embrace and accelerate AI adoption. With the announcement of Gemini Enterprise for Customer Experiences, Google is also telegraphing that they are truly thinking about the outcomes as much as the underlying integrated tooling needed to deliver those outcomes.
This is the closest we have seen Google get to understanding your business and providing solutions that up-level the conversation. Right now customer experience and agentic AI workflows are out there. It will be interesting to see what comes next.
If you are interested in the full list of announcements, start with Google Cloud CEO Thomas Kurian’s blog post.
Discover more from AVOA
Subscribe to get the latest posts sent to your email.