Google Cloud Unveils AI Agent Strategy at Cloud Next Conference

At Google Cloud Next, We're Off to See the AI Agents (And Huge Performance Gains)

Google's AI ambitions were on full display at its Cloud Next conference this week, but for me, two things stood out: It's now firmly promoting AI agents and is laser-focused on efficiency. As usual, there was a lot of AI hype. "It's a unique moment," said Google Cloud CEO Thomas Kurian, with CEO Sundar Pichai adding that it can "enable us to rethink what's possible." Pichai pointed to how Google helped recreate The Wizard of Oz for The Sphere in Las Vegas, converting it from a conventional movie format to one that works on a 160,000-square-foot screen. "Even a few years ago, such an undertaking would have been nearly impossible with conventional CGI," according to Google, with the team needing to "account for all the camera cuts in a traditional film that remove characters from parts of certain scenes, which wouldn't work at the new, theatrical scale that was envisioned." The AI-enhanced Wizard debuts on Aug. 28. But the "chance to improve lives and reimagine things is why Google has been investing in AI for more than a decade," Pichai said this week. A Surprisingly Fast Ascent for Agents AI agents were the real wizards at Cloud Next, however, with Kurian announcing several new interoperability-focused features for Google's AgentSpace platform. Notably, the new Agent Development Kit within Vertex AI supports the Model Control Protocol (MCP) that allows agents to access and interact with various data sources and tools, rather than requiring custom integrations for plugins. MCP was announced by Anthropic just a few months ago, and now it seems that all the major AI software companies are supporting it. In addition, Google announced a new Agent2Agent protocol that allows agents to communicate with each other regardless of the underlying model and framework they were developed with. Google offers its own purpose-built agents and tools for letting you build your own agents, but is now has a multi-cloud platform that "allows you to adopt AI agents while connecting them with your existing IT landscape, including your databases, your document stores, enterprise applications, and interoperating with models and agents from other providers," Kurian says. Salesforce CEO Mark Benioff appeared via video to talk about how Salesforce is working with Google to develop and connect its agents. The idea is that you could use Google's agents, create your own, or integrate with third-party agents. And of course, Kurian talked about how Google helps you create AI systems while addressing concerns about sovereignty, security, privacy, and compliance. Among the agents Google is producing are Customer Agents for call centers with human-like speech and dialog in Google's Customer Engagement Suite; Creative Agents for media production, marketing, advertising, and design teams; Data Agents for Big Query; and a number of Security Agents. Google also is introducing an Agent Gallery, no-code Agent Designer, Idea Generation agent, and a Deep Research agent. Meanwhile, agents in Google Workspace include a "Help Me Analyze" agent for Sheets, Workspace Flows to help automate tasks, and audio overviews, which turns Docs into audio summaries. It's interesting to me how quickly the agents concept has evolved. It was only a year ago that companies started talking about building them, as opposed to chatbots, which just answered questions. To me, it seems like a lot of agents are chatbots connected to robotic processing automation (RPA) tools, but that's fine if it can actually help businesses be more efficient. Now it seems like every major AI company is competing to create platforms that work with agents across software companies. Gemini Goes Pro AI isn't cheap; Google invests around $75 billion in capital expense, mostly for servers and data centers, Pichai says. The two most interesting areas here are the underlying models and the next-generation chips that will power them. A few weeks ago, Google announced Gemini 2.5 Pro, which Pichai describes as "a thinking model that can reason through its thoughts before responding." Gemini 2.5 Pro is now Google's high-end model, available through its AI Studio, Vertex AI, and Gemini app. At Google Cloud Next, Pichai announced that Gemini 2.5 Flash, a thinking model with low latency and the most cost-efficient performance, is coming soon. In addition, Google announced improvements to a variety of other AI models for specific uses. Imagen 3, its image-generating model, now offers better detail, richer lighting, and fewer distracting artifacts, Kurian said. Veo 2, the latest version of its video-generation tool, creates 4K video that is watermarked, but with features such as "inpainting," or removing parts of images. Chirp 3 creates custom voices with just 10 seconds of input. And Lyria transforms text prompts into 30-second music clips. With all these tools, "Google is the only company that offers generative media models across all modalities," Kurian says. All these models are available on Google's Vertex AI platform, which now supports more than 200 models, including those from Google, third parties, and open-source ones. Other changes include Vertex AI Dashboards to help monitor usage, throughput, and latency, new training and tuning capabilities, and a Vertex AI Model Optimizer. Strike While the Ironwood Is Hot In the infrastructure area, the biggest announcements was Ironwood, Google's 7th generation Tensor Processing Unit (TPU). Due later this year, this chip is said to offer twice the performance per watt of the current Trillium chip. Pichai says it has 3,600 times the performance of the first TPU Google introduced in 2013. In that time, Google has become 29 times more energy-efficient. Amin Vahdat, Google's VP & GM for Machine Learning, Systems, and Cloud AI, says demand for AI compute has increased by more than 10x a year for more than eight years, by a factor of 100 million. Google's newest TPU Pods support over 9,000 TPUs per pod and 42.5 exaflops of compute performance. (The pods will be offered in two sizes, one with 256 TPUs and the other with 9,216.) Still, these chips are "just one piece of our overall infrastructure," Vahdat said. Instead, Kurian talked about a building an "AI Hypercomputer" that involves multiple technologies. As part of this, Google also announced new compute instances with Nvidia's GPUs, as well as a cluster director that lets users deploy and manage a large number of accelerator chips; some new storage pools, called "hyperdisk exopools" as well as an "anywhere cache" that keeps data close to the accelerators, and a zonal storage solution, which offers five times lower latency for random reads and writes compared with the fastest comparable cloud alternative. In addition, the company announced new inference capabilities for the Google Kubernetes Engine and Deepmind Pathways for multi-host inferencing with dynamic scaling. Overall, Kurian claimed that putting all these things together means that Gemini 2.0 Flash powered by Google's AI Hypercomputer achieves 24 times higher intelligence per dollar compared to GPT-4o and five times higher than DeepSeek R1. And, in partnership with Dell and Nvidia, Kurian announced that Gemini will now run on Google Distributed Cloud for local deployments, including those that need to be "air gapped" for particularly sensitive applications. As part of the infrastructure push, Google announced that it is offering its global private network to customers. Pichai said the Cloud Wireless Access Network (Cloud WAN) contains over 2 million miles of fiber and underlies Google's services, delivering "over 40% faster performance while reducing total cost of ownership by up to 40%." I never take vendor performance numbers at face value, and obviously Google's competitors will have new offerings of their own. But it's interesting to see such a focus on not only performance but also cost. I know many CIOs who have been unpleasantly surprised by the cost of running AI models. This is a step in the right direction.

Google boards the AI agent hype train

Customers aren't sure, economy isn't great, tech looks cute, though Cloud Next This week Google joined a throng of tech vendors pushing the concept of "agentic AI" on an unsuspecting and perhaps unreceptive collection of enterprise users. Questions remain about how effective this tranche of tools will be at solving business problems and how much it might all cost. At its annual Google Cloud Next bash in Las Vegas this week, the Chocolate Factory opened the floodgates to a slew of product news. Among the announcements was a promise to introduce an Agent Development Kit (ADK), an open-source framework Google said would simplify the process of building business software that integrates AI agents - question-answering and task-performing software that makes decisions and forms its output using large language models. Agents can be seen as software that talks to other software. Google claims its ADK will allow customers to build an AI agent in under 100 lines of hopefully intuitive code. The ADK also comes with an "agent garden" of pre-built bots and tools, as well as more than a hundred pre-built connectors to common data sources. Meanwhile, the search and ads giant is set to introduce a new protocol designed to help agents from different vendors talk to one another. The Agent2Agent (A2A) protocol has attracted 50 partners signing up, such as Accenture, Box, Deloitte, Salesforce, SAP, ServiceNow, and TCS, which it said were actively contributing to it. This is giving enterprises false hope in accelerating human replacement, and we are not there yet Digging into the details, Google's agent garden is a pre-packaged collection of tools within the ADK allows users to access 100+ pre-built connectors, custom APIs, integration workflows, or data stored within Google cloud systems like BigQuery and AlloyDB. Google cited law firm Freshfields as an adopter of its agent vision. The smaller biz said it would use Google's enterprise data platform for AI, Vertex AI, to create bespoke AI agents for its legal and business processes, while also using search-bots to bring together company-held information. To power language processing and information gathering for users, Google announced the seventh generation of its Tensor Processing Unit, aka TPU, specialist AI-accelerating matrix-math hardware first used internally at Google. The chips code-named Ironwood - which you can read more about on The Reg here and The Next Platform here - have more than 10 times the performance of the earlier Trillium TPU. A fully decked out pod of Ironwood TPUs, rented from Google Cloud, can hit 42.5 exaFLOPS of FP8 compute, "meeting the exponentially growing demands" of generative models such as Google's recently announced Gemini 2.5. Google is not alone in seeing in AI agents an opportunity to extract more revenue from customers. Salesforce, for example, has been extremely bullish on the idea of business interactions being powered by its vision of gen-AI-powered automatons. CEO Marc Benioff told investors last year the company might overcome a fall in customer license numbers by charging users for each and every conversation they have with a bot, which he said was "a very high margin opportunity" for Salesforce. SaaS business application vendor Workday has also pushed AI agents on its new platform, even seeing them as a means to cut its own human headcount. Meanwhile, giant omni-vendor Microsoft is also in the agent game with its 365 Copilot range, powered by its $10-billion-plus alliance with ChatGPT maker OpenAI. But what about users, what are they getting out of it? In a recent research note, Gartner said folks should negotiate Microsoft Copilot Studio products with care as there is potential for extra costs to arise on top of direct licensing. "Microsoft 365 Copilot Chat could have a significant impact on the software and cloud spend of Microsoft clients. Governance of usage and cost is required," it said. Across all AI agent deplyments, users needed to be aware of the complexity of demands they could make on their cloud infrastructure. Yrieix Garnier, products veep at monitoring and observability vendor Datadog, said agents often operate across multiple platforms, such as Microsoft and Google. "Their behavior can vary significantly based on input, context, and chaining logic," he said. "A single user prompt can initiate a multi-step reasoning process, trigger API calls, or spawn other agents -- making usage and cost highly dynamic and difficult to predict. "To manage this complexity, organizations need to implement observability practices that go beyond traditional monitoring. This includes capturing and analyzing inputs, intermediate reasoning steps, and outputs for every agentic task. "Token usage, latency, model responses, and error rates must be tracked at a granular level. Just as with microservices, teams need traceability and context to understand where inefficiencies or errors are occurring -- and what they're costing. Guardrails are also critical: Without limits, agents can enter loops or generate excessive downstream tasks, leading to uncontrolled spend." Google's agent interface protocol is quite compelling ... Chances are that this is more marketing hype Chirag Dekate, VP analyst at Gartner, said Google was differentiating itself in the market with its investment in TPUs, its ability to interact with data sources, and the A2A protocol, which promises collaboration across agents from different vendors. "If you're trying to create an agented enterprise of the future, you need to have agents working with other agents to solve complex tasks and complex activities. Here, Google's agent interface protocol is quite compelling, because it enables you to have multiple agents work collectively or constructively with one another," he said. Nonetheless, limitations in the current generative-AI models means they are far from the level of intelligence needed to displace a human workforce. "Chances are that this is more marketing hype. This is giving enterprises false hope in accelerating human replacement, and we are not there yet," he said. At the same time, organizations may be put off investing in the vision of agent AI offered by Google, Salesforce, Microsoft, and others because of the economic uncertainty created by America's global tariff war, which has wiped trillions of dollars off global stock valuations and prompted warnings of worldwide recessions. "We are starting to see early signals of enterprises shifting towards cost optimization, as opposed to last year's focus around gen-AI. Businesses will continue to invest in gen-AI but cost optimization will likely take a much more serious tone this year round, because everything is changing. AI adoption rates will suffer, but AI innovation will not suffer because the cloud and hyperscalers are not slowing down. They are doubling down on investments," Dekate said. ®

Google Cloud makes its pitch to own multi-agent orchestration in the enterprise

If last year's Google Cloud Next was about introducing generative AI to enterprises, this year's event in Las Vegas is firmly focused on the rise of AI agents and - more importantly - how Google intends to manage them. In what appears to be a significant strategic move, Google Cloud CEO Thomas Kurian has laid out a vision for multi-agent systems that positions the company as (what it hopes will be) the essential orchestrator of enterprise AI ecosystems. The pitch is clear: while other vendors may be pushing single model integrations and a garden-walled approach, Google wants to be the platform that manages multiple AI agents working together across different frameworks and vendors. If successful, it's a play that could secure Google a strategic position in enterprise AI architectures, particularly for organizations trying to reduce their dependence on Microsoft (as many customers have suggested to me today that they are trying to do). Kurian emphasized during his keynote: Google Cloud offers an open, multi-cloud platform that allows you to adopt AI agents, while connecting them with your existing IT landscape, including your databases, your document stores, enterprise applications and interoperating with models and agents from other providers. You get value faster from your AI investments. What's notable about Google's approach is its emphasis on interoperability. Rather than forcing enterprises into a single model ecosystem, the company is betting that businesses will want a mixture of AI agents working together - some built in-house, others from vendors, with different specializations and capabilities. To support this, Google has announced a number of interconnected, key technologies: Behind this ecosystem play is Google's (very reasonable) belief that enterprises will need multiple specialized agents, rather than trying to make a single general-purpose model work for everything. And that customers will want the freedom to pick and choose which agents they need for which tasks, whilst managing them in a way that allows them to work seamlessly together. Kurian said: With Google Cloud, starting today, you can build and manage multi-agent systems with Vertex AI and our new agent development kit. You can scale the adoption of agents across your enterprise without the newly released Google AgentSpace. And you can accelerate deployment with packaged AI agents that are ready for use today. This multi-agent approach seems directly aimed at Microsoft, which has largely hitched its wagon to OpenAI's models (albeit with some flexibility being allowed). By emphasizing choice and interoperability, Google is positioning itself as the more open alternative - something that's likely to resonate with enterprises concerned about over-dependence on a single vendor. What makes Google's approach particularly interesting is its effort to build a coalition around its agent interoperability standards. The company has secured support from over 50 technology partners for its Agent2Agent protocol, including major players like Atlassian, Box, Salesforce, SAP, ServiceNow, and Workday, as well as consulting heavyweights such as Accenture, BCG, Deloitte and KPMG. This isn't just about Google Cloud - it's about establishing this open approach as the de facto standard for AI agent interoperability in the enterprise. However, with it being a leading force, if it were successful, it would put Google in a central position regardless of which specific AI models gain favor in the future. The protocol itself enables agents to communicate with each other, securely exchange information, and coordinate actions across enterprise platforms. It follows five key design principles: embracing agentic capabilities, building on existing standards, security by default, support for long-running tasks, and being modality agnostic. In practical terms, this means agents can advertise their capabilities, manage complex tasks together, collaborate through messaging, and negotiate different user experience formats. It's essentially creating a common language for AI agents to work together - with Google as the translator. Powering this agent ecosystem is Google's Gemini model family, particularly the new Gemini 2.5 which the company is positioning as a "reasoning model" capable of thinking through problems step-by-step. The company has introduced both Gemini 2.5 Pro for complex reasoning tasks and Gemini 2.5 Flash for high-volume applications where speed matters more than depth. Both feature what Google calls "enhanced reasoning" - essentially the ability to work through multi-step problems methodically rather than just pattern matching. This reasoning capability is crucial for sophisticated agent behaviors, allowing them to interpret visual information, understand text, perform searches, and synthesize diverse inputs into coherent outputs. It's what enables agents to move beyond simple chatbots into genuinely useful autonomous systems. A key challenge for any AI deployment, however, is ensuring its outputs are accurate and factual - something Google is addressing through multiple grounding approaches. It arguably has an advantage here thanks to its access to a broad range of data outside of its enterprise business (thank you Google Search, Maps and its various other consumer products). Kurian said: For model factuality, we offer the most comprehensive approach to grounding on the market today, combining grounding with Google Search, grounded with your own enterprise data, Google Maps and third party sources. This multi-pronged grounding strategy lets enterprises connect their agents to existing data sources, business applications, and third-party systems. Google is offering over 100 pre-built connectors, integration with services like AlloyDB and BigQuery, and direct connections to enterprise applications like Oracle, SAP, ServiceNow and Workday. The company is even making Google Maps data available for agent grounding, providing geospatial context for location-aware AI applications. This variety of grounding options is of course aimed at addressing enterprise concerns about AI hallucinations and factual accuracy. Google is eager to demonstrate real-world adoption of its AI agent technologies. Kurian noted significant growth in usage: In just the last year, we've seen over 40 times growth in Gemini use and Vertex AI, now with billions of API calls each month." The company highlighted several customer implementations, including Revionics building a multi-agent system for retail pricing, Renault Group developing an agent for EV charger placement, and Nippon Television implementing a video analysis agent. Others like Gordon Food Service are rolling out AgentSpace to employees to enhance access to enterprise knowledge. On the ground at Google Next, it's clear that the company is putting customer stories front and center - we have been given a lot of access to customers already deploying the latest agent technologies. What's particularly notable about Google's approach is its emphasis on providing a comprehensive platform rather than focusing on the models themselves. While competitors may be touting the benefits of a single LLM and what it can do across its platform, Google is offering a broader ecosystem approach. Kurian said: Vertex AI gives you easy access to over 200 curated foundation models. We offer all of Google's models, Gemini, Imagine and our latest research models, as well as curated popular third party models and open source models - all now on Vertex AI. Again, this platform approach positions Google as a neutral broker in the AI model space, able to support whatever models enterprises prefer rather than forcing them into a specific choice. It's a strategy that could prove particularly effective if model preferences continue to evolve rapidly, as they have over the past 18 months (the only constant is change in this AI market). Highlighting why customers are eyeing Google's multi-agent, foundational platform for enterprise AI orchestration, Kurian said: Customers around the world are choosing to work with Google for three important reasons. First, Google Cloud, offers an AI optimized platform with leading price, performance, precision and quality. And, new today, everything you need to build and manage multi-agent systems. Our AI platform offers advanced infrastructure and databases, world class research leading models and grounding model responses with Google quality search. If successful, this strategy could cement Google Cloud in a critical position within enterprise technology stacks. By owning the orchestration layer, Google could maintain relevance regardless of which specific models or agents gain favor in the future. It's a clever attempt to avoid making success dependent on winning the model race - something that's far from certain given the intense competition in foundation model development. Google's agent orchestration play comes at an interesting moment for enterprise AI. The initial excitement about generative AI is giving way to more practical considerations about integration, governance, and actual business value. How companies choose to incorporate AI into their technology stacks now will likely shape competitive dynamics for years to come. The company is clearly betting that enterprises will opt for an open, interoperable approach to AI agents rather than locking themselves into a single model ecosystem. This bet makes particular sense for Google, which lacks the enterprise software footprint of competitors like Microsoft or Salesforce, but has strengths in AI research, infrastructure, and platform services. Its multi-agent strategy represents a compelling alternative to the "single model embedded everywhere" approach favored by some competitors. By positioning itself as the orchestrator of diverse AI agents, Google is playing to its strengths (as it did with its multi-cloud strategy) while avoiding direct competition with entrenched enterprise software providers. The emphasis on interoperability and choice is likely to resonate with enterprises wary of over-dependence on a single AI provider, particularly given how rapidly the technology is evolving. The coalition Google has assembled around its agent protocols suggests this message is already gaining traction with significant players in the enterprise space. However, success will depend on whether enterprises embrace this more open, platform-based approach over simpler, more integrated options. Many organizations may prefer the convenience of having AI capabilities built directly into their existing systems rather than adopting a separate orchestration layer, regardless of the flexibility benefits. Not to mention that some of Google Cloud's competitors make it very difficult to allow customers to diversify their portfolio. What's particularly interesting is how Google's agent strategy could reshape enterprise software interactions more broadly. If successful, it could accelerate the trend toward natural language interfaces and abstraction layers that hide underlying system complexity - potentially disrupting traditional enterprise software user experiences. Although there has been little talk about that explicitly, given how much Google Cloud would have to deliver change in the market and drive adoption for that to become a reality. However, whilst there was plenty of 'cool' stuff to dig into today in terms of what AI agents and generative AI can 'do', what was most striking was that Google Cloud is speaking to the boring stuff that a lot of CIOs and tech leaders care about - interoperability, management, governance, security, etc. That's always a smart move. If it succeeds, it could secure a strategic foothold regardless of which specific AI models ultimately dominate.

Google Cloud Next 25 - CEO Thomas Kurian addresses Trump tariff uncertainty and reassures EU customers of data sovereignty

As geopolitical tensions rise and regulatory landscapes shift, Google Cloud CEO Thomas Kurian appears confident about navigating increasingly complex international waters. In comments made today at the vendor's annual Google Cloud Next conference in Las Vegas, Kurian offered customers and the market assurances regarding the ongoing turmoil of international tariffs under the Trump administration, as well as addressed speculation regarding the potential restrictions on US cloud services in the European Union. Kurian spoke candidly - yet diplomatically - about the volatile tariff landscape that has emerged since Trump's return to office in January. The administration's approach to tariffs has been, to put it kindly, notably inconsistent - initially threatening 60% tariffs on Chinese goods, then moderating the stance after market backlash, only to reassert aggressive positions weeks later. This pattern has extended to European trade relations, with proposed 10% blanket tariffs on EU goods followed by shifting exceptions for what Trump has termed "friendly countries" - a designation that appears subject to frequent revision. With a knowing smile on his face, Kurian said: I think you all see that the tariff discussion is an extremely dynamic one. But he did go on to compare the ongoing tariff situation to navigating global crisis' in recent decades: Obviously, we have an executive leadership team that has worked globally for many, many years. We run a global supply chain. We run a global distribution network. And we have been through many cycles like this, whether it's 2001, 2008, or when the COVID crisis happened in 2020. While Kurian seemed keen to lay out a 'stay calm and keep moving' stance in Google's ability to navigate the tariff landscape, he remained cautious about making definitive predictions: We are very confident that our executive team will find a path through it... I think we have to wait and see how this political environment and regulation evolves. Google Cloud's distributed global infrastructure might provide some insulation against tariff impacts, where its geographic diversification should help insulate it (and its customers) against some levels of uncertainty, However, it goes without saying that the fundamental challenge for Google Cloud - and indeed all US tech companies - is developing long-term strategy in an environment where trade policy lacks predictability or consistency. Kurian's diplomacy is wise considering that the safest bet regarding trade under Trump is that uncertainty is the only certainty (at present). Beyond tariffs, Kurian addressed another looming challenge: speculation that there is a potential collapse of the EU-US data transfer framework. Recent developments have raised serious questions about whether US cloud services could face restrictions in Europe. The Privacy and Civil Liberties Oversight Board (PCLOB), a critical component of the Transatlantic Data Privacy Framework (TADPF), has been weakened by Trump's removal of Democratic members. This has cast doubt on whether US cloud providers, if certain measures are removed, can continue operating in Europe without running afoul of GDPR and other EU privacy regulations. However, again, Kurian offered EU-based customers assurances on any situation that may unfold: In order to serve Europe on Europe's terms, we invested in sovereign solutions much earlier than competition. I think you will see the maturity of our sovereign solutions is yet another indication that we take European governments and their regulations and their needs [seriously]. When pressed on the possibility of an executive order that might prevent Google from providing services to the EU, Kurian outlined Google's technical safeguards: Technologically, we have solutions to protect against [an executive order that would prevent us from providing services to the EU]. These solutions include giving customers complete control over data location and encryption. He added: When you use any of our regions in Europe, you have sole control of the location of your data. You can encrypt it with your encryption key, as a customer. You can keep the encryption key away from Google in an off-site location you have. We don't have access to your environment, If we get a request to hand over a customer's data, you can deny us access to it, because you have the encryption key, and we do not have the ability to even reach the encryption key. For organizations with the highest security requirements, Google offers even more isolation: We also offer something for people who want to put the workload in the cloud but are worried about long term survivability. That's what we mean by Google Distributed Cloud, air-gapped, where it runs fully detached. There is no connection to Google and there is no connection to the internet. We also have sovereign partnerships with European organizations, where a European entity does a supervisory role on top of our operations, so they can trust and verify our operations. They can also be the custodian of the encryption keys. Under Trump, it's clear that organizations in the EU can't rely on Biden-era security policies - the TADPF faces significant uncertainty. Biden's executive orders on the matter offered reassurances to the EU that European data would be handled appropriately, especially in a post-Snowden world, but should this framework collapse, EU businesses would face difficult choices: migrate to European alternatives, risk substantial GDPR penalties, or partner with US providers that have developed genuine sovereign capabilities. And it's important to note that executive orders aren't law and Trump could easily overwrite what came before - and could be used as a bargaining chip as relations between the EU and the US become increasingly hostile. However, Kurian doesn't seem worried. He said: As geopolitical environments and national regulations evolve, people evolve to meet those requirements. Beyond navigating regulatory challenges, as we outlined yesterday, Google Cloud is making a strategic bet on AI agent orchestration as a core differentiator in the enterprise market. Kurian detailed the company's vision for multi-agent systems that position Google Cloud as an essential orchestrator of diverse enterprise AI ecosystems. Central to this strategy is Google's Agentspace, a platform designed to address agent-to-agent integration challenges in the enterprise. Kurian said today: Agentspace came out of our observations that many organizations spend a lot of time looking for information and trying to get tasks done. I have SharePoint, I have Office, I have my Oracle applications, I have Salesforce, and I'm spending all my time trying to find information about my supply chain, my customers, and my internal benefits system. Agentspace offers three core capabilities, according to Kurian: Number one, you can use Agentspace to have Google search for your company, access these systems and make them searchable and discoverable. Number two, you can integrate that into a conversational AI system, so that the AI can help you do research on this information, summarize it and give you answers to complicated questions. And third, you can then ask an agent to do tasks on your behalf. To facilitate adoption, Google has developed extensive integration capabilities. He also said that whilst good progress has been made around building a coalition around a certain set of standards, there are more companies signing up in the pipeline: People then asked us, can you provide connectors to all these systems so it's easy to discover them? We already have 100 connectors live. There's another 300 under development. That makes it easy for people to adopt the technology without having to rip or replace anything you have. The agent-to-agent protocol is a key component of Google's strategy, enabling different AI systems to work together seamlessly: We designed it to give it to the community, and it's supported by lots of partners. And I think that is the approach, I think in the end, that will be successful. It's very similar to what we did with Kubernetes years ago. Kurian also addressed growing concerns about AI's environmental footprint, where he was keen to highlight Google's efforts to reduce energy consumption while improving model performance: First and foremost, we have done a lot of work these last two years to hugely reduce the amount of cost required for both the training and inferencing of models. If you look at just inferencing since January 2024, not even 2023, the cost of inferencing has dropped by more than 20 times. These comments challenge the assumption that AI's growth will necessarily lead to proportional increases in energy consumption. The public perception is that AI will lead to negative outcomes for the environment; but Kurian had a different take: A lot of times, people take the size of a model and the cost of the model and then, say, if the usage is going to grow 20 times, you will need 20 times more energy - without realizing that within even a 12 month period, we have reduced it by 20 times, so you don't need as much energy. We introduced water cooling many years ago for our processors. The new Trillium version that we announced yesterday is a water cooled system. But our earlier systems are also water cooled. Those provide another 35% to 40% more energy efficiency when you're actually serving models. And just to give you an example, we have more than seven times the water cooled AI systems as the rest of the world combined. Beyond hardware optimizations, Google Cloud is also investing in more sustainable energy sources. Kurian said: We are investing in new AI energy technologies. For example, recently we introduced, and we are working with a bunch of organizations for nuclear power. If you look at several of the locations in Europe, we have sustainable energy sources: wind, hydroelectric, solar. There are several regions, even with the increased AI demand, that are higher than 95% sustainable energy. For us, it's really important that people see AI as being a technology that can also drive both efficiency in the consumption of energy and also create new forms of energy because of the investments that we're making. Kurian's message at Google Cloud Next 25 could be summed up as "Keep calm and cloud on," but the subtext reveals how technology companies are carefully hedging their bets in an unpredictable geopolitical landscape. His diplomatic answers about Trump's tariff whiplash acknowledge the chaos while essentially saying, "We've seen worse, we'll figure it out." The comparison to previous crises was telling though, highlighting the scale of impact these trade wars could have on organizations. And despite Kurian's steady hand, it's fair to say that no one can predict which way the Trump trade winds will blow next week, let alone next quarter. For customers, this means planning for flexibility rather than certainty. And on Kurian's comments regarding the potential collapse of the EU-US data relationship, the company's early investments in sovereignty solutions now look less like regulatory compliance and more like strategic foresight. With all that being said, Google Cloud has been clear this week that it's leading with customers front of mind. Its approach to agentic AI addresses the interoperability and 'garden wall' challenges faced by organizations as they make decisions regarding vendor choices, whilst offering buyers choice and protection in an increasingly turbulent economic and political environment. That bodes well for the company, in my opinion.

At Next, Google Cloud makes a credible claim to lead the new era of enterprise AI - SiliconANGLE

At Next, Google Cloud makes a credible claim to lead the new era of enterprise AI Google made a full-court press on artificial intelligence this week at its Cloud Next conference in Las Vegas, but the surprise is that it finally has a strategy and story that may give it a chance to lead the coming era of AI in the enterprise. Just as AI appears poised to change everything in cloud, the enterprise and beyond, the No. 3 cloud provider debuted new AI processors, a raft of services to create agentic AI features, an agent-to-agent protocol, more AI in search, enterprise access to its own massive cloud infrastructure, and a lot more. Who remembers that Google was seen as hopelessly behind in AI a year or so ago? Tariff insanity continues, with no sign of if or when it will end, at least before the 2026 midterms. Trump blinked on the broadest tariffs, sending stocks back up, but they went right back down Thursday, so who knows what's next? Those tariffs may have tanked tech stocks and our 401(k)s, but it's not seeming to slow down AI funding -- yet -- as a16z may be raising a $20 billion fund focused on AI and Signalfire raises $1 billion for early-stage AI startups. And Mira Murati's Thinking Machines reportedly is raising $2 billion all by itself -- in a seed round. But how long will it last? As the Journal says, tariffs are still coming for the AI boom. Elon Musk's DOGE aims to hold a hackathon to rewrite IRS software. What could go wrong? Here's all the news from this week, and you can hear more about this and other news on John Furrier's and Dave Vellante's weekly podcast theCUBE Pod, out later today on YouTube: Google Cloud Next was all about AI, of course -- but given it is after all a cloud event, you can read all the news and analysis in Around the Enterprise below. Breaking Analysis: Mapping Jensen's world: Forecasting AI in cloud, enterprise and robotics Stanford HAI's annual report highlights rapid adoption and growing accessibility of powerful AI systems How Apple fumbled Siri's AI makeover (per The Information): Internal chaos, poor leadership, John Giannandrea's AI/ML group dubbed "AIMLess" internally. Ouch. Mira Murati's Thinking Machines reportedly raising $2B in funding A16z in talks to raise $20B for AI investments -- more capital than VCs have raised so far this year SignalFire secures $1B to expand early-stage AI-focused investments OpenAI could reportedly acquire Jony Ive's AI device startup for $500M+ Rescale secures $115M Series D round to accelerate innovation with AI-driven digital engineering Arena AI raises $30M to accelerate innovation in hardware testing with Atlas Insurance technology startup Ominimo raises €10M in funding PyannoteAI raises $9M for its speech processing AI Deep Cogito releases open-source language models that outperform Llama TransUnion caps massive data migration project with new analytics and security services Riverbed rolls out new AI-powered observability features Stytch and Cloudflare partner to secure Remote MCP servers with OAuth Amazon Q Developer coding assistant now speaks multiple languages for coders worldwide Atlassian upgrades its AI assistant Rovo and makes it available to all customers Vectara launches open-source framework to evaluate enterprise RAG systems Camunda adds AI agents to its process orchestration platform Docker's latest release provides a foundation for local AI model development HubSpot unveils new AI agents and AI-powered workspace features for enterprise teams ThoughtSpot adds reasoning to its conversational agent TeKnowledge partners with Genesys to simplify the journey to AI-enabled CX Some big-picture thoughts after three days of roaming the conference: * This was nothing less than Google's coming-out party for the beckoning world of AI applications, especially in the enterprise. It's no secret that Google was caught flat-footed by the popularity of OpenAI's ChatGPT immediately after it was released in late 2022, despite Google possessing fearsome AI chops. And before that, it was caught flat-footed by Amazon Web Services' enormous success with cloud computing, despite Google having its own huge computing network. Now, AI has given Google another whack at capturing a large chunk of enterprise computing. "It feels like Google Cloud is the one-stop shop for AI now," Google Cloud CTO Will Grannis said. * CEO Thomas Kurian and Google in general seemed to have a crisper message about its position and ambitions than in previous years: It aims to become the go-to cloud for building and deploying AI-infused applications both in the cloud and in enterprise data centers. * In his keynote, Kurian called out several big advantages Google has in this new world of AI, and he wasn't just blowing smoke. Google Cloud offers a computing platform optimized for AI, it works in multiple clouds and soon in corporate data centers, and platforms for building the AI apps enterprises are now clamoring to create. "Enterprise AI has put Google on the main stage," new Google Cloud Chief Operating Officer Francis deSouza told me in an interview. TheCUBE Research's John Furrier noted that "we've always liked the nerd side of Google, but now they got the blocking and tackling done. They're not a dark horse anymore." Indeed, it's deSouza's job to nail down the processes needed to cement that blocking and tackling in the enterprise. * In other interviews, several Google execs emphasized their belief that the company has the strongest position in AI compared with its two larger cloud competitors, AWS and Microsoft, by virtue of developing its own AI models. (AWS and Microsoft do have their own, and likely more coming, but Microsoft mainly leans heavily on OpenAI, and Amazon's don't appear to be widely used yet.) One marquee project to try to bring that home: Google's work using AI to bring "The Wizard of Oz" to the immersive entertainment venue The Sphere in Las Vegas, where this week it showed some of its handiwork. * You could almost feel Google's renewed confidence, or perhaps an eagerness to exploit its renewed opportunity with the enterprise thanks to AI, which despite all criticism a year or two ago, it has deep expertise in -- deeper than AWS or Microsoft and most of the AI startups out there. "We've hit that inflection point over the last couple months," Raj Pai, VP of product for AI at Google Cloud -- and former AWS and Microsoft exec -- told me. * It didn't hurt that attendees seemed to feel it too -- 30,000 of them, 50% more than last year -- and the crowds at Mandalay Bay felt a little more like AWS' much larger re:Invent than previous Next shows. I even caught cellphone-waving crowds following Kurian on his tour of the expo hall Thursday -- not as frenzied as those who crowded Nvidia CEO Jensen Huang at his GTC event last month -- but the excitement was hard to miss. * Several new products stood out, including the new AI processor Ironwood, a seemingly big leap in performance and energy efficiency over the last tensor processing unit. It's for internal use, so Nvidia has little to fear, but as the seventh generation of TPUs, it's clear Google has some serious chip chops. * Another big entry was a preview of Cloud WAN, essentially offering Google's own global wide-area network for rent. It's significant that Alphabet CEO Sundar Pichai introduced it at the top of the keynote Wednesday. "It's one network for all enterprise needs," Google Cloud networking chief Muninder Sambi told me. * Likewise, Google's Gemini AI models will now run on Google's distributed cloud, enabling it to run inside enterprise data centers where some of the most important corporate data lives, including data that is too precious or regulated to move to the cloud yet. "Now, if you can't come to the cloud, Google will bring AI to you," said Amin Vahdat, Google Cloud's VP of machine learning, systems and Cloud AI. "We have found a way... to bring these very differentiated models to customers where they need them," added database chief Andi Gutmans in an interview. "AI is pulling customers into the cloud much faster." * Google also doubled down on AI agents, of course, in a number of dimensions, from an open-source Agent Development Kit to an Agent Engine for running them to an Agent Garden to find ready-to-use tools to an Agent Marketplace to an Agent2Agent protocol so they can all talk to each other -- even those not made by Google. (Though AWS, Microsoft, OpenAI and others are conspicuously absent. Google execs said they're working on that, but this will be a process as all prospective standards are.) * And though some of these new technologies are still in development and not fully available, it was no accident that Kurian spent much of his keynote highlighting and providing videos of customers such as Verizon, Toyota, Walmart and Reddit showing off that they're already getting business benefit from Google AI products. He also noted that services partners have already created thousands of AI agents. * So what about those tariffs? Kurian drew laughs in a panel session Thursday when he said in his customary deadpan way: "The tariff discussion is an extremely dynamic one." He added: "We have been through many cycles like this. We are very confident our executive team will find a path through it. We have to wait and see how this political environment and regulation evolves." * Of course, Microsoft and AWS will have their turn this year to make their cases, Microsoft at its Ignite conference in November (not to mention its Build conference next month) and AWS at re:Invent in December. The interesting thing, as theCUBE Research Dave Vellante said this week, is that unlike some technology markets where the No. 1 and No. 2 players utterly dominate, we now have a large and still-evolving market in cloud and AI where the No. 3 is not so far behind anymore. That's about all the time I have to sum up at this point, but next week, I'll look to add some more details as I empty out my reporter's notebook -- in particular about how agents may redefine and decompose software applications. Suffice to say, Google has big enough ambitions that it merits a couple of weeks of attention. But meantime, check out all our coverage of Next: AI, hybrid and multicloud: From Google Cloud Next, the driving forces behind cloud infrastructure's next big leap Google's cloud-based AI Hypercomputer gets new workhorse with Ironwood TPU Agent2Agent: Google announces open protocol so AI agents can talk to each other Google rolls out updates for building multi-agent AI ecosystems With agentic AI, Google Cloud is transforming almost every aspect of app development Google Cloud opens its global network for enterprise wide-area network use Google boosts AI and agentic features across BigQuery and AlloyDB Google uses AI wizardry to recreate 'The Wizard of Oz' in fully immersive 3D Google brings the power of AI to healthcare with new partnerships AI Mode in Google Search gets new visual search capabilities New Workspace AI tools from Google focus on workflow automation and content creation Google launches Unified Security platform and unveils Gemini agents for threat detection Google upgrades Android Studio with enterprise-grade Gemini AI tools Samsung integrates Google's Gemini into its Ballie home robot IBM revamps the venerable mainframe to run generative AI workloads and agents Dell lays out broad set of enhancements aimed at data center modernization Snowflake announces full support for Apache Iceberg Snowflake to begin military service after obtaining Department of Defense IL5 authorization Amazon's Andy Jassy reiterates the need to spend billions on building out AI infrastructure Networking startup Tailscale raises $160M at $1.5B valuation Incident.io raises $62M to build AI agents for incident response Silicon photonics startup nEye raises $58M to light up AI data centers British GPU-as-a-service startup NexGen Cloud raises $45M Observability startup Groundcover bags $35M in new funding to take on Datadog Fern gets $9M in funding to automate the creation and maintenance of software development kits Tariffs on China to increase to 104% as White House talks about building iPhones in the US Tech stocks surge after Trump temporarily lowers tariffs to 10% for all countries except China But not for long: Tech stocks give up some of Wednesday's gains following tariff pause Court rules UK Home Office's legal battle with Apple over data must be conducted in public Trump orders DOJ to investigate pair who disputed his allegation of election fraud Absolutely outrageous. Xanthorox AI emerges as a new malicious threat in cybercrime communities Darktrace uncovers spam bombing campaign used to mask targeted cyberattacks First quarter of 2025 sets record for ransomware attacks and threat groups SpyNote Android malware resurfaces in campaign using spoofed app install pages SentinelLabs exposes AkiraBot spam tool powered by OpenAI-generated messages Okta enhances Auth0 and core platform with expanded identity security features Fortanix introduces Armet AI for compliant, secure enterprise generative AI deployment CodeSecure and FOSSA partner to enhance visibility into open source and binary code Tessell reels in $60M for its database-as-a-service platform Hawk secures $56 million to expand financial crime detection platform Aurascape launches with $50M in funding to bring security and observability to AI apps Portnox raises $37.5M to scale cloud-native zero-trust access control Corsha raises $18 million to expand machine identity security platform Unosecur raises $5M in seed funding for AI-driven identity security Elon Musk's DOGE aims to hack the IRS and create a single API for easy access to U.S. taxpayer data

Google's cloud play: integrated AI from infrastructure to apps - SiliconANGLE

Google's cloud play: integrated AI from infrastructure to apps We predict Google's overall cloud business will reach $54 billion in revenue this year. Our models indicate that Google Cloud Platform will contribute to more than half of that revenue for the first time ever. Despite this performance, Google faces criticism even with its advanced technology and double-digit growth. The reason? It's still 2.5 to 3 times smaller than Amazon Web Services Inc.'s and Microsoft's cloud businesses. Regardless, Google has developed an artificial intelligence-optimized stack spanning from infrastructure to application-level agents. This stack includes custom hardware, integrated databases and a maturing development environment in Vertex AI that supports both proprietary and third-party models. We believe this makes Google a leader in providing tools for building enterprise agentic systems. Additionally, Google's transparent acknowledgment at its Next 2025 conference in Las Vegas this past week that enterprises won't move all their data to the cloud demonstrates an understanding of the reality that enterprise information technology requires hybrid and multicloud capabilities. By allowing integration with existing data, applications and workflows, Google aims to speed up the time-to-value for AI initiatives by bringing AI to the data regardless of its location. In this Breaking Analysis, we dig into Google's cloud strategy and how the company is integrating AI across its infrastructure and application portfolio to accelerate growth. We'll share our latest data on how Google Cloud is faring against leaders AWS and Microsoft Azure, including new survey data that suggests Google is closing the gap in AI adoption thanks to momentum in machine learning and AI. Google Cloud Platform's growth is outpacing its rivals, and we believe its differentiated AI capabilities -- demonstrated in its planned collaboration to bring and AI-driven version of "The Wizard of Oz" to the big screen at Las Vegas' The Sphere (pictured) -- are a key tailwind. We'll also examine the broader cloud market outlook, Google Cloud's improving financial metrics, and key areas to watch as the cloud wars evolve. To set the stage, the combined revenue of the "Big Three" U.S. cloud providers - AWS, Microsoft Azure and Google Cloud - is projected to reach $358 billion in 2025. The graphic above compares the annual revenues of the three providers over the last two years with our 2025 projections. Microsoft and AWS remain substantially larger in cloud than Google -- by roughly two-and-a-half to just over three times in terms of total revenue. Based on our CY 2025 estimates, Microsoft leads with approximately $178 billion, AWS follows at $126 billion, and Google ranks third at $54 billion. However, we believe it's significant that Google's cloud business already exceeds $50 billion and is still growing at a rate of around 30%. Imagine a company such as Cisco Systems Inc. simultaneously gaining share across multiple sectors and achieving 25% to 30% growth -- that's effectively what Google Cloud is doing today. In our opinion, it's also important to note that Google Cloud can leverage not only the substantial cash flows generated by Google's advertising, search and YouTube businesses, but also the massive infrastructure underpinning those properties. We believe this integration of resources is one of Google's core advantages in the cloud market. The chart below breaks down the percentage mix of infrastructure as a service, platform as a service, and software as a service for Microsoft, AWS and Google. Microsoft and Google currently derive most of their cloud business from SaaS, each with less than half of their revenue coming from IaaS and PaaS last year. Microsoft stands out as a dominant SaaS player. However, it's important to note that the vertical axis in the chart measures percentages rather than absolute dollars -- so we should not interpret Google's IaaS and PaaS revenue as larger than Microsoft's in raw terms. Our models suggest that Google's IaaS and PaaS business (GCP) will surpass 50% of its total cloud revenues this year, reflecting a shift toward infrastructure and platform offerings, driven in our view by Google's AI prowess. Notably, AWS is almost exclusively an IaaS/PaaS company, with only a small slice of SaaS (such as Amazon Connect for contact center). In our opinion, that composition can be viewed in two ways for AWS: either a missed opportunity in the SaaS arena or a point of differentiation, or an upside opportunity, especially since AWS remains highly profitable without significant SaaS contributions. We believe Microsoft's higher margins stem from its longstanding software DNA, while Google, still early in its cloud profitability curve, is steadily moving toward a more profitable line of business. Despite their differing portfolio balances, all three providers continue to grow, and AWS' ability to stay highly profitable while heavily focused on IaaS/PaaS is particularly noteworthy. This next slide from Enterprise Technology Research illustrates the core methodology behind its quarterly surveys, highlighting what we refer to as Net Score granularity. The vertical axis represents Net Score, which is calculated by taking the percentage of respondents that are adding or significantly increasing spending (the greens) and subtracting those that are decreasing or completely discontinuing spending (the reds). Specifically, the lime green portion (7% in the April survey for Google) is new customer adds; the forest green (38%) is the percentage of customers increasing their spend by more than 6% year over year; the gray indicates flat spending; the pink (7%) is spending down (6% or worse); and the red (5%) represents churn or isolation. When you net out the reds from the greens, you arrive at the Net Score, shown in blue line -- solidly in the thirties, which we believe is a respectable figure (noting anything north of 40% is exceptionally high). The yellow line in the chart, referred to as Pervasion, measures the level of penetration in the data set -- that is, the percentage of survey respondents that identify as Google customers. According to ETR's data, Pervasion for Google jumped from 26% in January 2023 -- just after ChatGPT's unveiling and the AI boom -- up to 33% today, a substantial seven-point gain in only a couple of years. We believe this trend sets up and allows us to test the premise of today's Breaking Analysis - that AI has provided a meaningful tailwind for Google Cloud, driving increased adoption and expansion within the ETR survey base and the broader market. We believe Google is rapidly closing the gap with AWS in the critical domains of machine learning and AI, which we see as a prime driver of Google's overall cloud growth -- particularly for GCP. We witnessed this firsthand at Google Next 2025, where the company showcased its AI-led innovations. The data below reflects ETR's Shared Net Score in the April survey, which had a strong sample of 1,859 IT decision maker respondents. Focusing specifically on the ML/AI sector, we isolate AWS and Google. On the vertical axis is Net Score, or spending momentum, and on the horizontal axis is Pervasion or penetration in the data set (that is, how many of these 1,859 organizations use each platform). In January 2023, AWS' N value was 175 respondents/accounts versus Google's 127, making Google about 73% the size of AWS within that ML/AI cohort. Fast-forward roughly two years, and the Ns for both firms have skyrocketed: AWS is cited by 449 respondents, Google by 425 -- now 95% of AWS' footprint. In the upper right corner of the chart, we see net scores for AWS and Google at 60% and 56%, respectively. Remember, anything north of 40% is already very strong; this data is extremely positive for both players. Because AWS is so dominant in the marketplace, Google's jump often goes unnoticed, but the numbers tell a clear story: Google is gaining meaningful ground in ML/AI adoption. Our research suggests that GCP's momentum in AI-centric workloads is a key contributor to this growth. The chart below tracks year-over-year growth rates for the three major hyperscale providers -- AWS, Azure and Google Cloud Platform -- going back to 2019. Importantly, this view isolates IaaS and PaaS revenue, excluding the SaaS component that boosts Microsoft's and Google's overall cloud numbers. We believe the data makes clear that, although growth rates have naturally moderated from pre-pandemic levels, Google is still substantially outpacing its competitors, after falling behind Azure's growth in 2023. Specifically, for 2025, we have GCP growing at about 34%, Azure at 27%, and AWS at 17%. In our opinion, these figures may understate AWS' potential upside -- Amazon CEO Andy Jassy's recent comments in his annual letter suggest some factors could push AWS' growth higher, and we plan to revisit that possibility in a future Breaking Analysis. Nonetheless, the contrast remains stark when we compare these hyperscale growth rates to overall IT spending (the yellow line in the chart). Traditional IT budgets are only growing in the low-to-mid single digits, while IaaS/PaaS is consistently above the 20% range. Our research indicates this divergence underscores a powerful, long-term shift to cloud infrastructure, with Google in particular benefiting from faster acceleration off a smaller base. This next slide focuses in on IaaS and PaaS revenue and provides a look at our 2025 estimates. The chart is labeled "Revised," reflecting the adjustments we made after reviewing leaked Microsoft/Activision court documents tied to statements in Microsoft CEO Satya Nadella's internal memos. Last year, Microsoft restated certain Azure definitions, swapping out slower-growing security and mobile solutions in favor of rapidly expanding AI and generative AI categories -- effectively boosting Azure's growth profile. In our opinion, those changes were largely immaterial to our existing models because we had already factored this into Microsoft's Azure revenue. Nonetheless, the above view shows AWS, Azure, GCP and Alibaba -- the last we consider a quasi-hyperscaler, though opinions on that front differ. We estimate that the combined IaaS and PaaS market for these four will exceed $250 billion this year (roughly $255 billion to $256 billion). Within that landscape, Google Cloud Platform continues to outpace its peers in growth, powered by the ML/AI tailwinds we highlighted earlier. Notably, we see GCP's IaaS/PaaS growth rate at roughly twice that of AWS -- albeit at a smaller base. Still, this translates into a roughly $28 billion IaaS/PaaS business within Google Cloud, a figure that Alphabet does not formally report and we must estimate. Because Google does not break out GCP revenue specifically, we rely on multiple data signals, including external surveys and market modeling. Another trend we're watching is Google Cloud's profitability, which recently turned positive, confirming Google's progress despite still ranking third among the hyperscalers. In our view, the overall cloud market remains vast, extending well beyond these top three -- into on-premises environments, edge deployments, global regions and new AI-driven or "agentic" services. The main takeaway is that though the Big Three command enormous revenues, there's significant headroom for further expansion, and Google's emergence into profitability marks a notable milestone in this multi-horse cloud race. This slide tracks the operating margin of Google Cloud as a percentage of revenue from Q4 2022 through the most recent quarters. At the end of 2022 -- just as ChatGPT was making headlines -- Google Cloud was running at an operating loss of around 2.5% or its cloud revenue. Since then, with the exception of a dip in Q4 2023, we've seen a steady improvement, with Google Cloud now achieving a positive operating margin of 17.5%. Though that number doesn't rival AWS or Microsoft yet, we believe it reflects meaningful progress. Google's improving profitability is largely the result of scale efficiencies -- as the business grows, fixed costs are spread across a larger revenue base -- and operational discipline under Google Cloud CEO Thomas Kurian's leadership. Kurian has driven an enterprise mindset, focusing on efficiency across engineering, go-to-market and support operations. Of course, there are headwinds. Like its peers, Google is pouring capital into its infrastructure, and those capex investments must be depreciated over time. Depreciation flows through the income statement and can be a drag on operating profit, especially for businesses still building out their base. This creates a lag between the time Google spends the money and when it shows up as a cost, hitting operating margins even as growth accelerates. Still, Google's 17.5% margin remains well below AWS' approximately 35% and what we estimate to be Microsoft Cloud's high 30s to low 40s -- with the caveat that Microsoft's "Intelligent Cloud" includes more profitable elements such as on-premises software and SQL Server, which we have attempted to exclude from our cloud forecasts. That makes direct comparisons somewhat murky. Nonetheless, AWS has achieved industry-leading efficiency in IaaS/PaaS, driven by a mature business, global scale and high levels of automation. In our view, Google's margin expansion story is just beginning. But as long as capex continues to surge -- and depreciation keeps pressure on the income statement -- Google Cloud's path to margin parity with AWS and Microsoft will take time. Still, the direction is clear, and the underlying fundamentals are strengthening. The slide below revisits a chart we've used before to visualize estimated capital expenditures for 2025 across the major tech platforms -- Amazon, Microsoft, Alphabet (Google) and Meta Platforms. Our estimate for total capex among these four companies exceeds $300 billion for the year. Though some reporting in the financial press recently cited a lower number, we believe that reflects confusion between 2024 and 2025 projections, or perhaps a misunderstanding of fiscal versus calendar year reporting, which we mix here as well. What's particularly notable here is the significant uptick in Alphabet's capex relative to its peers. Though Amazon continues to invest heavily, a substantial portion of its spending goes into its retail infrastructure, such as fulfillment centers and logistics. By contrast, Alphabet's spending is focused squarely on data centers, custom silicon such as its tensor processing units, AI infrastructure and networking, all of which feed directly into its cloud and AI ambitions. Microsoft, we believe, has slightly moderated its pace of capex, with internal estimates showing roughly $85 billion this year -- still a massive number, but lower than the trajectory it was previously on. Its calendar year 2025 spend will probably approximate that of Alphabet. The key takeaway is that Google is aggressively scaling up its infrastructure investments, and it is doing so in parallel with a strategic push into AI and agentic computing. Though this level of spending creates a near-term drag on operating margin due to depreciation (as discussed in the prior slide), we see it as a necessary long-term play. In our view, these capex investments are not only sustaining Google's competitive position, but actively powering innovation across the cloud and AI ecosystem. Far from being a drag on the industry, this wave of investment is helping define the next decade of computing. As we wrap this analysis, there are several strategic dimensions we'll be closely watching that could shape Google Cloud's trajectory over the next 12 to 24 months. At the center of the story is a key question: Can Google convert its leadership in foundation models and its deeply integrated AI stack into a lasting competitive advantage over AWS and Microsoft? The data suggests Google is gaining share -- it's growing faster than both competitors, particularly in IaaS and PaaS. But sustaining that growth will hinge on execution and differentiation. One of Google's assertions is that it is the only hyperscaler that owns a leading foundation model. Though Amazon might contest that with its proprietary large language model Nova, it remains unproven at this stage with respect to keeping pace with leading foundation models. Google's view is that its longstanding investment in AI -- through DeepMind and internally developed models such as Gemini -- gives it a unique edge. What sets Google apart, in its own words, is the tight integration of its foundation models across its entire stack. Much like Oracle Corp.'s longstanding advantage from integrating software with hardware, Google sees this vertical integration of AI as a moat that could define its value proposition. Another area we're watching is when and how customers choose Google Cloud, particularly in relation to data-centric and AI-priority use cases. Google clearly sees AI as its wedge -- and partnerships with Nvidia Corp. and even mentions of Dell Technologies Inc. at the keynote reinforce this push to bring AI "to the data," including in sovereign and hybrid contexts. That right strategic fit -- where AI and data gravity converge -- is critical to Google Cloud's expansion. From a developer perspective, Google is making aggressive moves to infuse generative AI into development workflows. TheCUBE Research's George Gilbert posits that Google appears to be two years ahead of Microsoft in bringing gen AI to the developer toolchain. As theCUBE Research's Paul Nashawaty has pointed out, GKE, short for Google Kubernetes Engine, is being positioned as the backbone for high-performance AI inference and training -- effectively evolving into a next-gen AI supercomputing platform. This is an area where Google has a strong technical foundation, and if it can win developer mindshare, it could set itself apart in AI-native application development. Then there's the data layer. Google made a bold claim on stage that BigQuery has five times the customer count of the two leading competitors -- a not-so-subtle reference to Snowflake Inc. and Databricks Inc. That figure raised eyebrows (and warranted some skepticism), but ETR data does suggest Google has greater account penetration than either Snowflake or Databricks, even if the 5X claim might reflect a long tail of small customers not captured in enterprise datasets. Regardless, it affirms that Google has a world-class data stack, built on native cloud principles, separation of compute and storage, and integrated AI. Google is also making notable progress in what we'd call semantic alignment -- harmonizing structured and unstructured data for agentic AI applications. We believe this will be critical in the next wave of enterprise intelligence, where AI agents require consistent, governed access to data. In that light, Google's stack -- anchored by Vertex AI and supported by Looker, BigQuery and open data formats -- is positioning itself to power this next-gen intelligent enterprise. Security is another frontier. Google acquired Mandiant to bolster its security capabilities, and there's a lot of curiosity around its acquisition of Wiz, a cloud security posture management company. Google executives didn't discuss Wiz on stage, but we view this as a highly strategic move. If Google can combine Mandiant's incident response with Wiz's cloud posture analytics and integrate that into its cloud platform, it could deliver a differentiated security model that rivals AWS' native capabilities. Bottom line: This year's Google Cloud Next may have been the company's most important yet. With a cloud business now exceeding $50 billion and growing at nearly 30% in 2025, Google has momentum, a differentiated stack and a strategy that appears well-aligned with where the market is heading -- AI, data and developer-led growth. The next chapter will be about scaling, executing in go-to-market, and proving that its vision of integrated AI isn't just technically elegant, but commercially winning. We'll be tracking all of it closely and as always appreciate your input.