Artificial intelligence is no longer marketed simply as a means to boost efficiency. According to IBM’s diagnosis, it’s becoming a revenue engine, a driver of corporate leadership redesign, and a reconfiguration of the technological architecture supporting organizations. The question, however, is not whether companies will invest more, but whether they will be able to turn that investment into measurable results without getting stuck in perpetual pilots.
One of the main conclusions from the global study conducted by IBM Institute for Business Value (IBV), in collaboration with Oxford Economics, which gathers the insights of over 2,000 senior executives on how their organizations will evolve between 2025 and 2030, is clear: in this “Company 2030,” the dominant narrative is that AI will shift from being an added accessory to existing processes to becoming the fabric that connects decisions, operations, and products in real time.
From AI as Cost Savings to AI as a Driver of New Revenue
The most striking data point is the jump in revenue expectations. By 2030, 79% of surveyed executives expect AI to significantly contribute to their organization’s revenue. Today, that figure stands at 40%. It’s an expression of intent, but also a sign of uncertainty: only 24% state they have a clear view of their main sources of income in 2030.
Ultimately, the study describes a market where competitive advantage is less about “doing the same cheaper” and more about inventing what has yet to exist. In fact, 64% of executives believe that advantage will come from innovation rather than resource optimization. This logic also applies to investment: between 2025 and 2030, they anticipate AI investments will grow by approximately 150%.
The technological perspective is twofold. On one hand, AI will continue to pressure automation and productivity improvement. On the other, that productivity is conceived as fuel to finance innovation. The study estimates AI will boost productivity by 42% by 2030, and 67% of executives expect to have captured most of those gains by then. According to IBM, the result is a “flywheel”: automation that frees up resources, reinvestment, and ultimately, business model transformation.
The Major Bottleneck: Business Integration
The paradox is that enthusiasm coexists with a very concrete anxiety: 68% fear that their AI efforts will fail due to a lack of integration with core business activities. This is not a minor detail. IBM explicitly distinguishes between “adopting AI” (adding tools) and “creating integrated intelligence” (making it inseparable from strategy and operations).
Here, a key term emerges for any tech-oriented media: orchestration. The report emphasizes the need for a neutral layer capable of connecting business platforms, applications, and AI agents with real interoperability, ensuring data flows and decision-making traverse the organization without depending on silos or fragile integrations.
David Carrero, co-founder of Stackscale (Grupo Aire), summarizes from an execution standpoint: “AI doesn’t fail because of a lack of demos; it fails when you try to put it into production and discover that data, permissions, latency, costs, and traceability are not resolved. Integration isn’t a project; it’s an operational discipline.” In other words: without strong data infrastructure and governance, AI remains a promise.
Multi-model, Not “One Model to Rule Them All”
Another important idea for 2030 is the transition towards model portfolios. IBM argues that winning companies will not optimize a single model, but rather a dynamic set that adapts to regulations, purchase cycles, and market volatility. In this regard:
- 71% of executives see emerging AI capabilities as complementary tools within a portfolio.
- However, only 28% are confident about which models they will need in 2030.
- The technical debate intensifies: many organizations expect to operate in multi-model scenarios, and competition will shift toward custom models and proprietary assets.
The practical implication is clear: if the future is multi-model, value lies not only in the model itself, but in how the lifecycle is governed (MLOps), how data and applications are connected, and how the model is “packaged” into products and services.
Carrero approaches this from an architectural perspective: “Model size matters less than the ability to deploy where it makes sense: close to the data, with cost control, and with security assurances. For many companies, advantage will come from specialized models and orchestration, not from pursuing the largest model.”
Agents, Leadership, and an “Always-On” Organization
The report describes the “smarter enterprise” as an organization that is “always active”: continuously processing market signals, adjusting course in real time, and experimenting with new revenue streams in a more automated way. This vision is accompanied by a profound organizational transformation.
The study predicts AI will redefine leadership: 74% of executives believe leadership roles will change and new positions will emerge. It even anticipates that by 2030, 25% of corporate boards will include an AI advisor or a “co-decider.” Simultaneously, internal friction is expected: 68% view current organizational structures as an obstacle to capturing AI’s full value.
In workforce terms, the study warns of an accelerated obsolescence cycle: 57% expect most current skills to become obsolete by 2030, and by late 2026, executives project that 56% of staff will require reskilling due to AI-driven automation.
For a tech-focused media outlet, a critical point is that this organizational leap won’t happen without a platform: cross-cutting agents and automation require observability, access control, auditing, and hybrid capabilities to move data and compute where needed. “If you’re going to deploy agents in critical processes, treat them like production: monitoring, change control, operational boundaries, and decision logs. Otherwise, you risk automating hazards,” Carrero emphasizes.
Infrastructure as a Competitive Advantage: Hybrid, Scalable, and Governed
IBM underscores that the multi-model portfolio must be supported by a hybrid architecture—flexible, secure, and scalable—that provides “instant” access to models, data, and applications to the teams that need them. This aligns with market trends: real-time AI, data supply chains, and pressures related to sovereignty and compliance.
From a European perspective, Carrero adds: “In Europe, the conversation isn’t just about performance; it’s about jurisdiction, sovereignty, and control. A hybrid architecture allows choosing where data resides and where models run without disrupting the business.” In essence, it’s the operational translation of the conditions the report highlights for competitiveness: speed, integration, and proprietary assets.
The study’s conclusion is not complacent. 2030 isn’t envisioned as a straight line goal but a race of rapid iterations, bigger bets, and less room for improvisation. Companies that can turn productivity into innovation, and build AI portfolios integrated into their business, will be those that convert AI into revenue. The rest will continue accumulating tools.
Frequently Asked Questions
What does “AI-first company” mean compared to a company that only adopts AI?
An AI-first company redesigns tasks, processes, and products so that AI is a structural part of the work (with humans overseeing), rather than just adding AI tools over legacy workflows without changing decision-making or operational methods.
Why is business integration the main risk in enterprise AI projects?
Because real use cases depend on data, permissions, core systems, traceability, and accountability. Without an orchestration layer and solid data governance, AI cannot traverse departments reliably or reach production effectively.
What is a multi-model strategy, and why is it gaining importance in corporate environments?
A multi-model approach is based on portfolios: foundational models, small specialized models, proprietary models, and agents, combined according to task, cost, latency, compliance, and risk. It enables AI to be tailored to specific processes and avoids dependency on a single provider or architecture.
What are the implications of a board of directors including an AI advisor?
It affects governance, responsibility, and risk management: from how AI-assisted decisions are validated, to how models are audited, biases documented, operational limits set, and investments aligned with business outcomes.
Source: IBM Report on AI in 2030