Cloud Wars Live with Bob Evans

In today's Cloud Wars Agent & Copilot Minute, I explore Microsoft's addition of Mistral Large 3 to its Foundry platform and why it's a major win for developers seeking open, enterprise-ready AI models.Highlights00:11 —Microsoft Foundry is a platform from Microsoft, designed for building, customizing, and deploying GenAI applications and agents. It allows users to access over 11,000 models. Recently, Microsoft added a new model to its offerings, Mistral Large 3. Microsoft claims it to be one of the strongest open-weight, Apache-licensed frontier models available on the Microsoft cloud.00:53 — An open-weight model is similar to an open-source model, but with some differences. In an open-weight AI model, the parameters used to train the model are publicly available, not just for use or viewing, but also for downloading and modifying. More developers are turning to these models because they offer flexibility.01:40 — Mistral Large 3 is one of the world's leading open models optimized for enterprise applications. It excels in instruction following, long-context comprehension and retention, multimodal reasoning, predictable performance, and applied reasoning. Unlike its closest competitors, Mistral Large 3 is fully open, with Apache 2.0 licensing.02:08 — Each new model is carefully selected. It's not just a free-for-all. With Foundry, Microsoft is demonstrating its expertise in the space by keenly understanding what developers want—in this case, an open model with real-world enterprise applications. The pace of AI development is such that even incremental changes and improvements dramatically impact how businesses operate. Visit Cloud Wars for more.

Sales, and Matt Hobbs, Cloud Engineering and Data Analytics Platform Leader and Partner at PwC US. Together, they explore how companies can stop overpaying for cloud and instead fund AI innovation by shifting spend from legacy and suboptimal cloud deployments into modern architectures, multi-cloud strategies, and enterprise-grade AI capabilities that actually move the needle on growth, margin, and new business models.Smarter Cloud, Bigger AIThe Big Themes:Built to Cost Less: Oracle entered the cloud market later and designed OCI from the “bare metal up” with off-box virtualization, a low-latency non-blocking network, and significantly lower egress pricing. That means Oracle’s own cost to deliver infrastructure is structurally lower, so they don’t need to “race to zero” with margin-crushing discounts. When customers compare OCI run-rates to first-generation hyperscalers, it’s common to see 40–70% savings at list-to-net, not just in special deals.Turning Technical Debt Into Innovation Budget: Hobbs notes that roughly 40% of internal tech budgets are often tied up in technical debt rather than innovation. PwC sees executives searching for ways to unlock capital for AI and growth initiatives, not just trim expenses. Its “Fit for Growth” program looks at where money is tied up in non-differentiating costs (cloud infrastructure being one of the biggest line items) and reallocates that spend into value-creating initiatives. When PwC runs side-by-side economics, they’ve seen OCI’s promised 40–70% savings show up in real deals.OCI + PwC: budget creation meets execution: The Oracle–PwC collaboration stands out, the guests argue, because both sides are relentlessly focused on the client outcome rather than maximizing any one platform. PwC validates OCI’s economics and brings the talent to design and execute migrations, process re-invention, and agentic AI programs; Oracle brings a cost-efficient, multi-cloud-friendly infrastructure designed for price-performance and portability.The Big Quote: “You can burn a lot of money chasing ghosts in this game if you really don't have a very specific use case." Visit Cloud Wars for more.

Bob Evans sits down with Will Grannis, Chief Technology Officer at Google Cloud, to unpack how AI is reshaping both technology stacks and corporate culture. They explore Google Cloud’s Gemini Enterprise platform, the newly upgraded Gemini 3 models, and the rise of agentic AI. Along the way, Will shares customer stories from industries like finance, healthcare, retail, and travel, and even talks about how his own team had to change its habits to benefit from AI.Inside Google Cloud’s Agentic AI The Big Themes:Models vs. Platforms in the AI Stack: Grannis draws a sharp distinction between AI models like Gemini and the broader platforms that operationalize them. Models determine how intelligent and capable AI workflows are “out of the box,” across tasks like reasoning, multimodal understanding, and conversation. Platforms, by contrast, are how a business injects its own data, processes, and rules to build differentiated IP, brand experiences, and competitive moats. In practice, that means thinking beyond a single chatbot to agentic workflows composed of models, data, tools, and multiple agents working together.Culture and Discipline: Grannis describes how even his own team initially struggled to build an internal ops agent to automate sprint reviews, status updates, and reminders. It was only after leadership pushed them to be an exemplar that the agent became reliable and valuable. Things as simple as putting status information in the same place on every slide suddenly mattered. The lesson: AI exposes hidden process chaos. To get leverage from agents, organizations must tighten their operating discipline and be willing to change how they work, not just bolt AI onto old habits.Rethinking ROI and Metrics: Traditional, siloed ROI metrics can kill transformational AI efforts before they start. Grannis cites research about AI projects dying at proof-of-concept stage and contrasts that with companies like Verizon, which used AI in the contact center to simultaneously lift revenue, reduce cost, and improve customer satisfaction by turning support calls into sales moments. Instead of chasing a single metric in isolation, he advocates for “bundles” of outcomes anchored in customer experience.The Big Quote: “We had to be more disciplined about how we conducted our own work. And once we did that, AI’s effectiveness went way up, and then we got the leverage.”More from Will Grannis and Google Cloud:Connect with Will Grannis on LinkedIn or learn about Gemini Enterprise. Visit Cloud Wars for more.

Welcome to the Cloud Wars Minute — your daily cloud news and commentary show. Each episode provides insights and perspectives around the “reimagination machine” that is the cloud.In today’s Cloud Wars Minute, I discuss why culture, mindset, and leadership matter just as much as technology in driving AI transformation, based off my conversation with Will Grannis, CTO, Google Cloud. Highlights00:30 — Will has been the Chief Technology Officer at Google Cloud, one of the world’s most advanced technology companies, for almost a decade. So Will’s perspectives on things are pretty powerful, especially in this notion of how corporations unlock the power of AI to drive great outcomes for those companies and their customers or their patients or their stakeholders.01:10 — One of the first things that Will talked about is the big AI unlock. He said you’ve got to start with thinking about putting the customer at the center of everything, and then build back, build out from there. So reverse-engineer what has to change inside the organization to ensure that the customer outcomes, the customer experience, the customer value, are at the center.AI Agent & Copilot Summit is an AI-first event to define opportunities, impact, and outcomes with Microsoft Copilot and agents. Building on its 2025 success, the 2026 event takes place March 17-19 in San Diego. Get more details. 02:27 — He talked a lot about the mindset. One customer example was recently BNY Mellon. BNY Mellon has added Gemini Enterprise for its Eliza AI platform, and that is being used now. The Chief Data and AI Officer at BNY Mellon said our AI strategy in the company is simple. He said it’s AI for everyone, AI everywhere, and AI for everything.03:19 — He said this is something that’s enabled them now to do more things for their customers. It allows their internal people to be much more productive, be more expansive in their analysis, so that they can provide greater value to their customers. Will said it’s been a huge change at the company.04:06 — So again, I hope you have a chance to check out the whole interview with Will Grannis, the Chief Technology Officer at Google Cloud. You can see it in the links here. Will’s a terrific guy. One of the things you’ll see here is he offers some pretty honest and candid assessments about challenges he himself has faced as the CTO at Google Cloud, and very candidly explains how he got around those. Visit Cloud Wars for more.

Key TakeawaysMullican expresses the value of the event, and given that AI and Copilot have matured, the sessions at the 2025 AI Agent & Copilot Summit NA will shift from a focus on experimentation to practical applications that deliver real impact.As a returning Programming Committee Board member, Mullican highlights the "unprecedented number of submissions this time around,” making it tough to choose sessions for 2025. The event will offer a mix of visionary talks, practical use cases, and deep dives to serve both tech and business audiences.Mullican shares how business leaders can use the event to "adapt to a new reality," as it aims to “cut through the hype” around AI and focus on practical realities. Even though AI offers transformative potential, most businesses will adopt it gradually, using tools like copilots to improve efficiency rather than undergoing radical overnight changes. "I expect that there's going to be a lot of discussion about. It's just these different layers, these different levels of AI hype and truth." Visit Cloud Wars for more.

In today’s Cloud Wars Minute, I look at how IBM and Cisco are teaming up to pioneer long-distance quantum networks.Highlights00:03 — Although not the only company invested in the development of quantum computing, IBM is certainly considered the most active. The company has the highest number of patents, a clear road map for fault tolerant quantum systems, and the most prestigious track record across quantum hardware, software and the commercialization of these tools.00:28 — Now, IBM and Cisco Systems have revealed plans to link a network of quantum computers over long distances — and the result, perhaps the introduction of the quantum internet. Before I get carried away on this, leaders from both IBM and Cisco have confirmed that the technology to power these networks doesn't yet exist, but they are working on it.00:59 — The bottleneck is getting qubits, the unit of information used by quantum computers, to travel along fiber optic cables between Cisco switches. IBM and Cisco hope to have the first proof-of-concept ready within five years, a network that connects individual, large scale, fault-tolerant quantum computers with the power to run computations over 10s to hundreds of 1000s of qubits.01:49 — So, why do we need the quantum internet? Well, beyond the massive enhancement in computational power, which is the primary driver for companies to enter this space, if quantum computing itself becomes widespread, we'll need quantum structures in the Internet to protect ourselves from our very own creation.02:28 — Technology is advancing at an unfathomable speed, and just like in the AI space, we need to ensure it's contained. In fact, researchers at IBM co-developed three of the four quantum resistant algorithms that the U.S. National Institute of Standards and Technology, or NIST, have earmarked for future standardization. Visit Cloud Wars for more.

Highlights00:24 — I think we’re seeing some very interesting competitive battling between Palantir and ServiceNow, head to head, emerging over two categories. One, this AI platform position that they’re both talking about relentlessly as sort of the core of their strategic value. And secondly, they’re both on this kick now about who is the defining software company of the 21st century, of the generation.01:57 — Both are talking up the power of the AI platform as the foundation for what companies need to succeed in this new AI-centric future. Which one has the right approach? How are they trying to position their companies, their capabilities, their ecosystems, to be able to take advantage of that? These are some of the things that I get into in detail in a later article.02:19 — We’ve also got both claiming that their results are just the greatest thing ever relative to the Rule of 40, which is for high-growth companies. You want to have a combination of 20% growth or more for revenue and 20% growth or more for margins, to have those equal 40 or more. Both exceed that number.03:26 — Who’s got the momentum? Who’s positioning themselves right in this sweet spot for businesses moving into this AI future, right? Who’s going to be able to put together the tools that control the agents, that let all the data that’s needed for IE come through? How are they able to build applications easily on top of this?04:05 — But right now, I think ServiceNow and Palantir are the ones that have the most focus on this AI platform vision. And it was fun to sort of look at this. In the article, we look at some of the comments from the Q3 earnings call from ServiceNow CEO Bill McDermott, and also from Palantir’s Chief Revenue Officer, Ryan Taylor. Visit Cloud Wars for more.

Key TakeawaysAgents take center stage: Bensch says last year’s event was focused on Copilot basics and early-stage AI adoption, but 2025 is the year of agents. Organizations are asking how to make agents work securely, how to tap into data through offerings like Work IQ and MCP servers, and how to build governance that actually holds up. Attendees can expect deep dives into security, data extraction, Dataverse, Finance & Operations, and all the new features unveiled at Microsoft Ignite. “Everyone’s hearing agents, agents, agents,” he notes. “So, how do I get them to work?”Sessions designed for real adoption—not marketing fluff: This year’s call for speakers exploded from ~160 last year to over 500 submissions, giving planners a far stronger pool of practical, hands-on sessions. The programming committee prioritizes real-world implementations, lessons learned, and “the good, the bad, and the ugly” of AI projects. He stresses that this event is not a product-pitch environment. Instead, sessions will help attendees understand how to build, deploy, and scale agents across modern work, business apps, and development workflows. “We’re looking for empowering people,” Bensch says.Where strategy meets execution: Bensch explains that most attendees will fall somewhere between the starting line of AI adoption and mid-stage Copilot integration—but everyone is looking to connect the dots between strategy and execution. From governance to Dataverse to legacy-system integration via computer-use capabilities, sessions will show how companies can extend agent intelligence far beyond Q&A. The setting, including intimate sessions, world-class speakers, and networking events like golf and pickleball at Torrey Pines, creates space for candid, high-impact conversations attendees won’t find at massive trade shows. Visit Cloud Wars for more.

In today’s Cloud Wars Minute, I review the remarks and insights from CEO Marc Benioff and Salesforce from the recent Q3 earnings call.Highlights00:20 — After an approximately three-year hiatus, CEO Marc Benioff and Salesforce demonstrated their growth mojo through the Q3 numbers. While some of the numbers aren't quite as robust as they were in Salesforce's earlier days, in the earnings call, Benioff emphasized growth, innovation, and new things coming that are reflective of where the company was in the past.01:07 — In its first 22 years, Salesforce had unprecedented growth — 20 years of 20% or higher growth. No other publicly traded company has done that. In the last few years, with some pressure from institutional investors, Salesforce had to shift its focus from growth and innovation to margins and profits. During this time, the character of the company has evolved, especially with the AI Revolution and the introduction of Agentforce.01:55 — Something that struck me was the exuberance of CEO Marc Benioff on the call and his excitement about lots of numbers that indicated things are headed in the right direction. He shared details and commentary about their Q3 numbers as well as the vitality and energy around new products.02:58 — Benioff was proud of the stats around Agentforce customers moving into production. The number of those was up 70% sequentially quarter to quarter. This demonstrates how quickly Agentforce customers are able to deploy the technology, get it into use, and start getting the essential business outcomes.03:38 — This is important because the biggest winners are always the customers in the Cloud Wars because they get to benefit from the incredible competition. It further triggers waves of relentless innovation unlike anything the world has ever seen. Visit Cloud Wars for more.

In today's Cloud Wars Minute, I discuss how Palantir’s latest customer wins reveal a shift from point solutions to end-to-end AI strategies.Highlights00:19 — Palantir's got a customer that, very shortly after signing a deal for a fairly limited AI engagement, said it wanted to go enterprise-wide with an investment that's up 8x over what the original one was. This signals where CEOs are driving these AI-centered business transformations to ensure that they go end-to-end here and have enhanced business outcomes as the goal.01:21 — This medical device manufacturer CEO went to the Palantir executive team and said, “What would we need to do to make this an end-to-end business transformation powered by AI, designed to drive greater outcomes?” And very quickly, they rewrote the whole deal. So five months after this initial thing, they had a new deal going on — eight times bigger.02:31 — So again, there we see this notion of Palantir doing some different things and engaging with customers in different ways. So, I think broadly what we're seeing here, overall, with these AI-centered business transformations is: time is the enemy — not your traditional competitors, not necessarily the new ones. It is this notion of time.03:40 — If you're on the customer side here, I think you've got to look at this and take a hard view of what's going on inside your company. Is it a lot of these little disjointed trials that could have a nice little upward bump in efficiency or cost savings? Or are you shooting for the moon here? Visit Cloud Wars for more.