At last week’s Google Cloud Next ’26 conference in Las Vegas, Google’s parade of announcements reinforced its momentum as an integrated, rising AI giant with its own proprietary stack of hardware, models, and enterprise AI services.

Google launched new line of its TPU chips as well as diverse services to help enterprises build agentic AI services. The new positioning has resonated with investors, who have piled into Alphabet shares, which are up 12% year-to-date and 120% in the past twelve months (Alphabet is the parent company of Google).

It’s a remarkable change in sentiment since the summer of 2025, when investors were selling Alphabet shares on the fears that AI services would eliminate Google’s advantage in search and disrupt its online advertising gravy train. It shows how challenging it has become to pick winners in AI services, where competitive differentiation can shift in a matter of months.

These new developments emphasize the pace of change in the AI services and infrastructure market, where trillion-dollar fortunes can be made and lost overnight. And they show why securing our own supply of AI chips is key, with shortage in the industry driving up pricing of components.

Unwilling to be solely tied to outside AI chip vendors such as Nvidia and AMD, both Amazon and Google have developed their own line of both AI training and inference chips.

Amazon is also on the upswing with its release of Trainium 3 chips, announced in December. Amazon has a signed a deal with Meta to selling millions of its chips to the social media giant, and Amazon Andy Jassy said that Amazon’s Trainium and Graviton chips are a legitimate alternative to Nvidia and AMD .

Google and Amazon’s hardware positioning has offset the strengths of Microsoft, which is now looks increasingly beholden to outside chip providers.

As part of its own chip expansion, Google Cloud last week announced its AI Hypercomputer architecture for datacenter AI, which incorporates new processors, storage, and networking elements, including the following:

  • TPU 8t, optimized for training, which can scale up to 9,600 TPUs and 2 PB of shared, high-bandwidth memory in a single superpod. Google says it achieves 3x the processing power of its Ironwood predecessor and delivers up to 2x more performance/watt.
  • TPU 8i, optimized for inference, directly connects 1,152 TPUs in a single pod. Google says it delivers 80% better performance per dollar for inference than the prior generation and supports millions of concurrent agents.
  • Google Axion N4A instances of Google’s custom Arm-based Axion CPUs, which Google says provide twice the price/performance compared to current-generation x86-based VMs.
  • Google’s Virgo Network, an AI-optimized fabric for training that links either NVIDIA Vera Rubin NVL72 systems or TPU 8t superpods into massive supercomputers packing hundreds of thousands of accelerators.

As the AI infrastructure buildout shifts from training, which builds the key foundational AI models, to inference, which delivers AI services and results to customers, Google is now well-positioned with its own inferencing infrastructure, with it’s own supply of chips.

Carving Out a Position in Enterprise AI

As Futuriom research shows , Google is one of the leading providers of enterprise AI services, with Google’s Vertex AI placing number two among more than 200 enterprise case studies we have examined.

Google has rolled Vertex into a group of products and services built around what it calls the Agentic Enterprise, a term gaining traction around the market for building and managing AI agents. Armed with a growing presence in hybrid cloud infrastructure, Google is sending a message that because it has fully integrated all of the infrastructure for both cloud and AI, that can lend itself to more security and control for enterprises.

Google Cloud CEO Thomas Kurian described Google's Gemini Enterprise Agent Platform as a "single command center for pollicy ... you protect your models, you protect your proprietary enterprise data from threats such as sensitive data leakage," he said in the Cloud Next keynote. "This integrated approach, from secure sandboxes to a single management console, provides the visibility and isolation to run your most sensitive workloads with a high degree of confidence."

The Gemini Enterprise Agent Platform, which incorporates and replaces Vertex AI, is a proven favorite among enterprises, according to Futuriom research, so that will prove an important foundation for its growth. In addition, Google’s messaging follows along the lines of what enterprises have been demanding for AI: proprietary platforms that they can control, along with industrial-strength security.

Enterprise AI Software Threat Grows

With Gemini Enterprise Agent Platform, Google has taken Vertex AI and added an army of services for creating, orchestrating, securing, simulating, tracking, and monitoring agents in Google Cloud. They have also added Model Context Protocol support and specialized agents from an Agent Marketplace that work with Atlassian, Box, Oracle, ServiceNow, Workday, and other enterprise systems and applications.

On the AI services front, these integrations demonstrate a key competitive advantage that Google will have against a vast array of enterprise AI startups, whether it's Amazon, Microsoft, OpenAI, or Anthropic. That is, Google controls its own cloud infrastructure, AI models, and chips. This has the benefit of reducing costs, because Google owns both its own proprietary models (Gemini and others), as well as giving it control over integrations and security of its infrastructure. Most competitors must buy one or the other from somebody else.

For the vendors in the enterprise software category, these announcements should be terrifying. The stocks of enterprise software leaders such as Microsoft, Oracle, Salesforce, and ServiceNow have been dumped as investors question the threat that agentic AI will be to the future of enterprise software. With these announcements, Google is clearly increasing its presence in these markets.

For now at least, Google’s new releases give the company renewed momentum both in the cloud services and enterprise AI services markets.