A recently published industry outlook on the state of the software sector offers a useful, if somewhat sobering, snapshot of where the business of building enterprise software is actually heading in 2026: toward a period of intensifying financial pressure, an all-consuming pivot toward agentic AI capability, and a competitive landscape being reshaped by AI-first products that are forcing even the largest, most established players to rethink how they operate. Taken together, the findings paint a picture of an industry in the middle of a genuinely difficult transition, one where the winners are not simply the companies with the most impressive AI demos, but the ones that can translate agentic capability into durable revenue while managing margin pressure that has already begun eating into the profitability metrics that once made software one of the most reliably lucrative sectors in the broader economy.
The Full-Stack Agentic Platform Bet
The outlook’s central thesis is that established software players are increasingly focused on becoming full-stack, end-to-end agentic platforms, companies that help their customers build, run, orchestrate, and govern AI agents across every business function, rather than remaining primarily focused on their traditional, narrower software category. This represents a meaningful strategic broadening for many established vendors: a customer relationship management platform, for instance, is no longer simply competing on the strength of its CRM features, but increasingly on how comprehensively it can support the entire lifecycle of AI agents deployed across a customer’s sales, service, and marketing operations, from initial development through ongoing governance and monitoring.
This full-stack ambition is not merely aspirational marketing language. Salesforce, one of the most closely watched examples of this strategic pivot, has publicly raised its financial forecasts on the strength of its Agentforce product line, with sales from that product surpassing the half-billion-dollar mark, a figure that represents genuinely substantial commercial traction for an AI agent platform still in a relatively early stage of its product lifecycle. That kind of concrete revenue validation matters considerably to an industry that has, over the past several years, sometimes struggled to distinguish between AI capabilities that generate measurable customer value and AI features added primarily to satisfy competitive pressure and analyst expectations.
The Uncomfortable Math of Shrinking Margins
Alongside this strategic pivot toward full-stack agentic platforms, the outlook highlights a less celebratory trend: gross margins among the largest SaaS companies by market capitalization have been gradually compressing over roughly the past decade, a pattern that predates the current AI investment wave but has been accelerated by it. Running AI inference at scale, particularly for the more sophisticated multi-step reasoning and tool-use capabilities that agentic products depend on, carries meaningfully higher compute costs than the comparatively lightweight software infrastructure that traditional SaaS products have historically relied on. That cost structure shift represents a genuine, structural change to the software industry’s underlying economics, not merely a temporary investment phase that will naturally resolve itself as AI infrastructure matures and costs decline.
This margin pressure helps explain several of the broader industry patterns already playing out this year, including the trend among the largest software platforms toward building in-house AI models specifically to reduce dependence on, and the associated costs of, third-party frontier model providers for routine, high-volume workloads. It also helps explain why analyst forecasts increasingly emphasize enterprises’ expanding AI agent adoption in the same breath as noting that many of those same enterprises struggle to achieve what they would consider genuine AI maturity, since scaling this technology cost-effectively while maintaining acceptable margins has proven considerably harder in practice than the initial wave of agentic AI enthusiasm anticipated.
Adoption Is Racing Ahead of Governance, Again
A recurring theme across current research into enterprise AI adoption is a persistent gap between how quickly organizations are deploying agentic capability and how mature their governance and oversight structures actually are around that deployment. Analyst projections suggesting a large majority of enterprise applications will feature task-specific AI agents within the next several years describe an extraordinarily rapid adoption curve, one that is significantly outpacing the parallel development of the security, compliance, and operational discipline needed to manage that many autonomous or semi-autonomous systems safely and effectively across an organization.
This is not a new pattern in enterprise technology adoption cycles, but the current agentic AI wave compresses the timeline in which this gap typically resolves itself considerably. Previous waves of enterprise software adoption, from cloud migration to robotic process automation, generally unfolded over a long enough period that governance practices had time to mature roughly in step with adoption, even if imperfectly. The speed and breadth of this year’s agentic AI adoption curve leaves considerably less room for that gradual maturation process, raising the stakes for organizations that treat governance as an afterthought to be addressed once initial deployment has already scaled across the business.
The Cloud Infrastructure Layer Beneath It All
One dimension of this transition that receives less attention than the agentic platform strategy itself is how heavily the entire full-stack agentic vision depends on the underlying cloud infrastructure layer that most software vendors do not themselves control. Even the largest independent software companies pursuing ambitious agentic platform strategies are, almost without exception, building on top of compute infrastructure provided by one of a small number of major cloud hyperscalers, which means their own margin economics are directly exposed to pricing and capacity decisions made by infrastructure providers with their own competing strategic interests in the same agentic AI market. A software vendor absorbing rising compute costs to run increasingly sophisticated agentic workloads has limited leverage to negotiate those costs downward when the hyperscaler supplying that compute is simultaneously building competing first-party agentic products of its own.
This dynamic adds a further layer of complexity to the margin pressure story, since it means the software industry’s cost structure challenges are not purely a function of AI compute becoming more efficient over time, a trend that would eventually ease pricing pressure on its own, but are also shaped by the competitive posture of the infrastructure providers that most software companies depend on and, increasingly, compete against for the same enterprise agentic AI spending.
What This Means for Software Buyers
For enterprise software buyers evaluating vendors in this environment, the outlook’s findings suggest several practical considerations worth weighing more heavily than in previous procurement cycles. Vendors pursuing the full-stack agentic platform strategy are making a genuine long-term bet on their ability to support the entire agent lifecycle, not just an initial point solution, which means buyers should evaluate a vendor’s agent governance, monitoring, and orchestration capabilities with the same rigor traditionally reserved for core product functionality, rather than treating those capabilities as secondary features layered onto an otherwise familiar purchasing decision.
Margin pressure across the industry also has direct pricing implications worth anticipating. As vendors absorb rising AI compute costs while pursuing ambitious full-stack platform strategies, buyers should expect continued upward pressure on software pricing over the coming budget cycles, even as individual vendors compete aggressively on AI-specific feature pricing to win new agentic platform deployments. Navigating that combination, rising baseline costs alongside aggressive promotional pricing for newer AI capabilities, requires more careful total-cost-of-ownership analysis than software procurement has typically demanded in recent years.
What to Watch Next
Several developments over the coming months will help clarify how this transition plays out across the broader software industry:
- Continued margin trends among major SaaS companies in upcoming earnings cycles. Whether margin compression stabilizes, accelerates, or reverses will be a key indicator of whether the industry is successfully managing the cost structure shift toward AI-heavy products.
- Revenue traction for other vendors’ agentic platform products, beyond Salesforce’s Agentforce. Comparable revenue validation from other major software vendors would confirm this as an industry-wide successful strategic pivot rather than an isolated success story.
- Enterprise AI maturity survey results in the coming year. Continued gaps between AI investment intentions and reported AI maturity would reinforce concerns about adoption outpacing organizational readiness.
- Software pricing trends across major enterprise categories. Watch for whether rising AI compute costs translate into broader price increases across software subscriptions generally, beyond AI-specific feature add-ons.
- Consolidation activity among smaller software vendors unable to sustain the full-stack agentic platform investment required to remain competitive. Margin pressure combined with the scale of investment required to build comprehensive agentic capability may accelerate acquisition activity among smaller players unable to compete on that basis independently.
The Bigger Picture
The software industry’s pivot toward full-stack agentic platforms represents a genuine strategic response to where enterprise customer demand is clearly heading, and early revenue validation from products like Agentforce suggests this bet is paying off for at least some of the companies pursuing it aggressively. But that strategic success is arriving alongside real financial pressure, from compressing margins to the enormous compute costs required to deliver genuinely useful agentic capability at scale, and alongside a governance gap that is widening rather than narrowing as adoption accelerates faster than most organizations’ oversight capabilities can keep pace with. How individual companies navigate that combination of strategic opportunity and structural cost pressure is likely to determine which software vendors emerge from this transition as genuine long-term winners, and which discover that agentic ambition without a sustainable cost structure behind it was not, in the end, a winning strategy.
Leave a Reply