CoreWeave’s next test: From GPU scarcity to a durable AI cloud

Backup and recovery evolved into data protection then security, cyber recovery and now business resilience. The scope continues to expand and now includes how attackers use AI; and how defenders can leverage AI to better protect their assets. In our new field research with our partner Qualitate, we explored a number of issues, including one of the less frequently discussed risks around what happens when the AI inside an enterprise makes a mistake. Broadly, our research indicates that confidence in recovery is running well ahead of capabilities. Our field work underscores the immaturity of cyber resilience and AI security and specifically, how agentic AI is putting more pressure on practitioners to not only secure their data but re-create an agent’s intent and confidently recover the state of a business. 

The AI buildout is making an enormous bet on what businesses will eventually bewilling to pay for intelligence. In our research this summer on Forecasting theAI Bubble, we argued that a key indicator to watch is productive utilization byenterprise customers. Specifically, we said if infrastructure supply catches upbefore productive monetization materializes, the risk of a bubble pop increasesto dangerous levels. We forecast that supply constraints and healthy capitalfunding appetites could provide the conditions for a “soft landing” and wepegged 2029 as the high risk year. That brings us to the next obvious question:what has to happen inside mainstream enterprises for their AI investments tobecome economically self-funding? Michael Burry’s analysis of Anthropichighlights the scale and interconnected nature of the commitments. Apollo’sresearch by Torsten Slok raises questions about how broadly the financialbenefits are showing up today. And the history of the PC and internet erasreminds us that productivity gains are not immediate - and that powerfultechnology can coexist with what initially seem to be poor investment returnsthat take time to appear.

In Breaking Analysis 322, we shared with you that our forecast of a $1T silicon business arrived 2 years earlier than we had forecast. The reason for our twenty-four month miss was the bottleneck had moved from GPUs to memory. Micron gave us a remarkable datapoint in its FY Q3: DRAM prices rose by more than sixty percent sequentially in the quarter, while the number of bits shipped barely grew. Today we address what that means. The key questions are: Is this temporary scarcity or a structural change in the economics of memory? We believe both forces are at work. This week we participated in a Webcast with Micron President and Chief Technology and Products Officer, Scott DeBoer. The roadmap he showed us underscores the importance of memory with respect to AI performance. New customer take or pay agreements change how risk is shared. At the same time, customers are being forced to adapt to higher costs. Qualitate’s proprietary research gives us a window into those decisions. Our question is whether these changes can make memory less cyclical, and what evidence would tell us that the familiar cycle is returning.

ServiceNow is making one of the more consequential and visible bets in enterprise AI. The company says AI Control Tower can discover and govern any agent, work across models and clouds, track cost and value, and shut down behavior that violates policy.That sales pitch plays directly to ServiceNow's greatest asset – its installed base. ITSM, CMDB, risk workflows, security operations and years of operating context give the company a distribution and data advantage most AI governance vendors do not have.Our research suggests that moat is real. But proprietary customer data shows a more nuanced picture. Control Tower is already strong as a governance workflow and system of action. It can inventory AI assets, connect risks to owners, route approvals and remediation, and preserve an audit trail.What the field has not broadly validated is the claim that Control Tower is a complete technical governance and runtime enforcement plane across the heterogeneous enterprise.

In this week’s episode, we explore the AI control gap and the starting point for addressing AI safety, namely alignment. We question prevailing narratives and propose a public policy regime that has more teeth than what has been put forth thus far. 

Ahead of CoreWeave’s Fully Connected conference, we have made a notable investment in proprietary customer research with Qualitate. Rather than simply repeat the earnings call, we went to the people evaluating, buying and running the infrastructure. This analysis draws on thirteen in-depth interviews and more than seven hours of interview time. We then compared that evidence with CoreWeave’s financial disclosures and the public statements of CEO Michael Intrator and CFO Nitin Agrawal.Our research indicates that GPU scarcity opens the door, but performance, cost and the operating experience give customers reasons to stay. Inference is growing alongside training. Notably, training is not declining at the expense of inference. Inference is growing on a very steep curve and training workloads continue to grow as well. At the same time, the hyperscalers remain deeply embedded in the application estate, and not every successful CoreWeave eval turns into a signed customer.The question is whether CoreWeave is converting an availability advantage into a durable AI cloud. We believe the customer evidence strengthens that case. It also shows exactly where the case still needs work.

Salesforce’s next growth opportunity may come from customers spending less time in its interface—and having agents do more of the work.Coming out of Dreamforce 2026, we believe that is the shift worth exploring. The progression from command lines to graphical interfaces, browsers and mobile is entering another phase: an agent can generate an interface around the task, inside Claude, Slack or another client. The customer no longer has to start with the application’s screen. The starting point becomes a customer outcome.But generating an interface is not the same as understanding the business. That is where our System of Intelligence (SoI) framework comes in. The system of engagement (SoE) connects people and agents. The system of intelligence supplies the business context. And the system of agency (SoA) turns that understanding into action. Salesforce’s larger ambition is to connect these elements through what it calls an enterprise AI harness: the data, business knowledge, workflows and controls that let models do useful work across systems. This goes well beyond answering questions.The customer evidence from our good friends at Qualitate, suggests an exciting opportunity…but the data also shows a company in transition. Qualitate’s latest channel checks find that most organizations interviewed remain in Agentforce pilots or proofs of concept. Yet among customers who have modeled or experienced the financial impact of headless access, 75% expect Salesforce spending to increase—not decline. Working outside Salesforce’s interface could actually mean doing more work on its platform.That creates both an opportunity and a challenge. More consumption can generate more revenue, but it does not automatically generate more customer value. Pricing must make sense as pilots scale. And if customers increasingly work through their preferred AI environment, Salesforce must demonstrate why its business context and execution capabilities remain an essential underpinning of the customer experience. The point is, the interface can move elsewhere, meaning the next purchase may not automatically stay within Salesforce.Welcome to Breaking Analysis #327. Salesforce After Dreamforce…How $CRM (the company) can grow beyond it’s own interface. In this episode, George Gilbert joins us to map Salesforce’s post-Dreamforce direction into our System of Intelligence framework. We combine his firsthand observations with new Qualitate customer research to examine headless access, Agentforce maturity, competition from AI-native agents, and the changing relationship between pricing and business outcomes.

The cloud shared responsibility model was initially not well understood by many customers. In fact, early adopters often believed that simply having data in the cloud meant that Amazon, or a SaaS vendor were responsible for safeguarding it. Amazon had to educate its customers and partners that security and compliance duties were split between the vendor and the client organization. In short, the vendor was responsible for securing the cloud resources but you, the buyer, were responsible for securing what you put inside the cloud; based on your policies, priorities and budget. We believe a similar but much more consequential dynamic is unfolding with respect to agentic AI. Specifically, Cloud computing divided responsibility by infrastructure layer. Agentic AI distributes authority across a chain of models, platforms, clouds, partners and customers. Our premise is the industry now needs a shared accountability model for the decisions, actions and outcomes that chain produces.In short - The cloud shared-responsibility model told customers who secures what. The agentic shared-accountability model must define who can do what, who can stop it, who can prove what happened and who pays when it goes wrong.Welcome to episode 326 of Breaking Analysis. Beyond Shared Responsibility - When AI Acts, Who owns the Blast Radius. In this Breaking Analysis, Principal CUBE Research Analyst Krista Case and I explain why the agentic era demands a new accountability model. We’ll draw on learnings from last week’s CrowdStrike Fal.Con event, where the post Mythos moment and the OpenAI/Hugging Face “accident” were front and center. We’ll also draw on other new datapoints, including conversations with CISOs at Fal.Con and Palo Alto Networks’ earnings print from last week, to unpack what we’ve defined as a new AI accountability model. We’ll also test this new model against our Sovereignty framework, developed by Amit Govrin and assess sovereignty in the context of business recovery. We’ll explore the sequence of events that leads up to the ultimate question of who pays when something goes wrong? 

Breaking Analysis with Dave Vellante

Breaking Analysis with Dave Vellante and Amit Eyal Govrin