Everybody wants to wrap around the data, because that’s where they’re seeing the most retention.
Marcella ArthurVP, Strategic Alliances, memoryBlueThe conversation about the AI race revolves mostly around models: who builds the best one, who deploys fastest, who secures the most compute. But that framing misses the more fundamental contest underneath it. The real competition is a land grab around data, and the hyperscalers racing to win it increasingly behave less like technology companies and more like the largest commercial real estate operators in the world. Their goal is to build the infrastructure, pull as many organizations inside it as possible, drive up consumption once they're there, and make the ecosystem so valuable that leaving becomes harder than staying. For any company building a go-to-market strategy on top of that infrastructure, understanding the real estate logic has become a competitive necessity.
Marcella Arthur is VP of Strategic Alliances at the sales development firm memoryBlue, where she works across the partner ecosystems and platforms that technology companies increasingly build on. She's spent considerable time thinking about what the scramble to own data means commercially and where it leaves the companies whose customers and workloads now live inside someone else’s environment.
"Everybody wants to wrap around the data, because that’s where they’re seeing the most retention," Arthur says. Once you see data as the property everyone's fighting over, the rest of the market starts to look a lot like real estate, too.
Compute is the new square footage
The clearest parallel, in Arthur’s view, is the credits. "AWS is offering credits, Azure is offering credits, Google is offering credits. Everybody is giving you as many credits as possible because they want to house the data," she observes. Housing it is the point, because once the data is inside, the provider can wrap policy around it and make leaving difficult. The economics rhyme with commercial property almost exactly. Where real estate measures value in location, occupancy, and expansion, cloud and AI infrastructure measure it in compute, workloads, and consumption. Marketplaces, security platforms, and AI models stand in for the amenities that keep a tenant from moving out.
The shift shows up most visibly in how things are priced. The market has moved away from per-user licensing toward consumption, Arthur notes, and she saw it again in a recent pricing conversation with a vendor. "One of the first things they said was, 'We don’t price per user. We don’t care how many employees you have. It’s all about consumption,'" she says. "I’m thinking, 'Oh my gosh, it’s all about the data again.'" Even the physical scramble follows the analogy, with bidding wars now breaking out over where to place the data centers that will house all of it.
What gets lost in the grab
The rush to pull data into the fold is being treated as a straightforward retention win, and that, Arthur believes, is exactly the problem. "What they’re not thinking about is the future of privacy and policy and compliance," she says, or what the arrangement means for data sharing and security down the line.
Her reframing is that organizations should stop treating data as an asset and start treating it as property, something with a location, an owner, a custodian, access rights, and security requirements, and something that can turn into a liability fast if the people managing it don't understand the responsibility that comes with it. The technology, she points out, is advancing much faster than most organizations’ governance around it.
The ICP you can't just assert
The same abundance of data has rewritten how a go-to-market motion gets built. It's no longer enough to declare a target and run at it. "You can’t just go in and say, 'This is my ICP.' That’s no longer the way forward," Arthur says. The honest version now starts with a hypothesis. "You say, 'I think this is my ICP. But within that, I bet there are decision-maker roles and personas making decisions differently,'" she says, some answering the phone at certain times, others responding to email, others heads-down on research. Quantifying those differences separates a sophisticated go-to-market motion from a crude one, and the data finally makes it possible, provided a team uses it to sharpen relevance rather than to justify more outreach.
For a services company, that changes the deliverable itself. Running a campaign and pushing it to market is no longer the product. "Now you have to show evidence of that campaign, beginning to end, in a very well-documented and orchestrated way," Arthur says, because the competition is every other tool looking at the same data through its own lens. The value is in what the interactions reveal. Every campaign should teach something, she asserts, every objection should refine the positioning, and even the negative responses should return market intelligence, so that a team is constantly learning while competitors are simply executing.
Partnerships become dependencies
The land-grab logic lands hardest on partnership strategy, because operating inside someone else’s ecosystem is now the default rather than the exception. Whether a company builds around Microsoft, AWS, Google, Snowflake, Salesforce, or another major platform, those relationships shape where its product lives, how customers discover it, and how data moves through its environment. That raises a question beyond the usual one of who can help you sell. "If you’re leaning on a hyperscaler and that hyperscaler is housing your data, part of it is, what are they housing," Arthur says.
A critical discipline is dividing the sensitive data from the non-sensitive, the data that keeps the business afloat from the rest. The deeper partnership questions, in her framing, are whose data practices align with yours, what dependencies you're creating, and whether a relationship makes you more valuable or simply more dependent.
Earning the right to use it
What decides this next phase cuts against the instinct driving the whole land grab: hoarding the most data settles nothing, because more data doesn't automatically produce more intelligence, and more intelligence doesn't automatically produce better decisions. The edge goes to the organizations that can separate useful signal from noise, apply context, and act on it in a way that creates value for the customer, which takes judgment and governance as much as technology. "The real advantage belongs to the organizations that earn the right to use the data," Arthur says, and that right is earned by treating trust, privacy, security, and responsible stewardship as things with real economic value. The infrastructure will keep changing. That principle, she believes, will not.





