Is the Worst Over for Saas? & Top 5 Technology Stories | CRTV Episode 129

The SaaS apocalypse had a good run as a narrative. For a few months, the prevailing wisdom was that vibe coding, AI agents, and a generation of developers armed with Claude and Codex would roll their own software and walk away from the enterprise SaaS stack entirely. Episode 129 of ConstellationTV put that narrative in front of Constellation analysts Larry Dignan, Esteban Kolsky, and Martin Schneider, and the result was a nuanced, honest debate (in which they don't all agree).

The Debate: SaaS Apocalypse or SaaS Shakeout?

Esteban opened with the kind of directness that makes the CRTV debate worth watching. The SaaS apocalypse, he argued, was always a pricing concern dressed up as an existential crisis. Once vendors addressed the pricing anxiety, the narrative collapsed in about two days. The real SaaS companies, the ones with deep enterprise footprints, years of integrations, and the institutional knowledge that took a decade to build, are not going anywhere. You cannot replace three years and $25 million worth of legacy integration work with a vibe coding session.

Martin agreed on the broad strokes but pushed back on the edges. The established players with genuine systems of record are fine. The point products are not. The SMB-focused, single-function SaaS companies that never built deep integrations and never earned irreplaceable status in the enterprise stack are facing real margin compression, and AI is accelerating what was already on the horizon. The ones providing value will survive. The ones that were coasting on lock-in and switching costs without delivering real outcomes are already in trouble.

Larry added the observation that has been sitting under the whole conversation: zombie SaaS is real, most of it is owned by private equity, and whether AI is the cause or just the excuse for the shakeout is still an open question. It might not matter. The outcome is the same.

The Amazon angle was sharp. Larry raised AWS Connect's expansion into talent, supply chain, and healthcare as a sign of hyperscaler ambition in the SaaS space. Both Esteban and Martin were skeptical. The DNA of an infrastructure company is not the DNA of an enterprise software company, and that gap does not close easily. EMC tried it. CA tried it. The go-to-market motion, the customer relationship model, the way enterprise software gets sold and implemented, it is a different business. Google knows this about itself, which is part of why it has stayed in its lane more carefully than Amazon.

The bottom line from the debate: SaaS is not dead. But the shakeout is real, and it was coming with or without AI. The vendors that delivered genuine value and depth of integration will be fine. The rest are on borrowed time.

From Google Cloud Next: When Data Finds Its Meaning

In the next segment, Mike Ni caught up with Paul Pritchard, CEO of Overdose, a high-growth digital commerce agency, at Google Cloud Next, and the conversation cut to one of the most practical questions in enterprise AI right now. Most organizations do not have a data problem or an AI problem. They have a meaning problem. They have more data than they know what to do with, and none of it is telling a consistent story.

For Overdose, the turning point was Google Cloud's Looker semantic layer, built on BigQuery. Before, the agency was spending a month building dashboards that looked backward at the previous month's performance. Every channel claimed the sale. Every stakeholder had a different interpretation of the numbers. The CFO, the board, the marketing manager, all of them were looking at different truths and arguing about which one was right.

The semantic layer solved that. It established a single, grounded truth that every team, and every client team, could access in real time. From there, Overdose built a tool that takes a Looker insight, generates a creative brief, produces creative assets, and automatically publishes them to ad channels, with Looker tracking performance and triggering the cycle again when it is time for new assets.

The metric that matters: the business is now more profitable because the teams are focused on margin-oriented decisions rather than vanity metrics. Clicks and revenue are easy to chase. Profit is harder, and it requires knowing the truth first.

Paul's advice for any data or commerce leader starting this journey: begin at the data layer. Get to a single truth. Everything else builds on that foundation, and without it, everything you build on top is noise.

Larry's Five Tabs: What You Need to Know This Week

Larry's tabs covered five stories worth tracking across cloud, enterprise software, hardware, and an emerging pattern that deserves more attention than it is getting.

1. AWS Connect has expanded well beyond the contact center into talent, healthcare, supply chain, and workflow automation through Amazon Q. The strategic intent is vertical integration, and it is a direct challenge to the SaaS vendors in those categories. Whether it works is a separate question, but the ambition is clear.
2. Google Cloud is in the pole position on AI infrastructure. First quarter revenue was up 63%, putting the platform at an $80 billion annual run rate, roughly half of AWS but growing faster. The AI stack, models, TPUs, and tooling is the most fully built out of any hyperscaler right now, and Google Cloud Next made that case comprehensively.
3. ServiceNow's Knowledge conference introduced an AI control tower update worth watching closely. The new value calculator tracks AI spend, agent performance, and business outcomes in one place. For enterprises drowning in AI pilots and unclear on returns, a tool that actually tries to answer the ROI question is not a minor feature. It could be the killer app for the control tower.
4. Apple's earnings flagged something that is going to ripple across every hardware and device category: memory supply is getting tight, and costs are going up. Macs are particularly constrained because demand for Mac Minis and Mac Studios from developers building with local AI models is outpacing supply. Every device maker will face versions of this problem, and the question of what gets passed to consumers remains open.
5. The most interesting tab was the last one. GM and Rivian both reported earnings where software and services revenue is becoming a meaningfully larger share of the business. Rivian's software and services revenue rose 49% year over year, driven in part by its partnership with Volkswagen. Caterpillar is heading in the same direction. The pattern is clear: every company is becoming a software company, and AI is accelerating that transition faster than most investors are pricing in.

Why You Should Go to Canvas 2026

Constellation Research VP and Principal Analyst Chirag Mehta closed the episode with a question that cuts across every conversation about enterprise AI: Is AI making everyone faster, or is it making teams better?

Individual productivity gains from AI are real and measurable. A product manager can pressure test an idea in minutes. A designer can generate concepts almost instantly. An engineer can explore multiple implementation paths in the time it used to take to explore one. But faster individuals do not automatically create better-aligned teams. In fact, AI can amplify the small misalignments that already exist in an organization. More ideas, more plans, more prototypes, and the same fundamental questions left unanswered about what the team is actually building and why.

That is the collaboration challenge Miro is addressing with Canvas 2026, and it is why Miro is on the Constellation ShortList. The category is moving beyond digital whiteboards into platforms that help cross-functional teams structure work, apply AI, and move from ideas to coordinated outcomes. Shared context is becoming infrastructure. It is how strategy connects to execution, and how organizations ensure AI produces better decisions, not just more output.

Canvas 2026 is built around that conversation. If you are a product, engineering, design, or technology leader working through the same challenge, it is worth your time. Register here before May 19th.

Watch the full episode on YouTube and subscribe to Constellation Research for weekly enterprise technology debates, customer stories, and the five things you need to know.