Generative AI is in the “Trough of disillusionment”
Right now, generative AI is going through a critical transition phase. It’s what Gartner calls the “Trough of Disillusionment.” Basically, this is when a technology that once generated massive interest starts to lose momentum. The hype is settling down, and people are beginning to see its limitations. That’s a good thing. Failure is part of progress, and this phase is when real, durable value gets built.
We’ve seen big-name failures recently. Humane launched an AI pin with promises of revolutionizing how we use devices, it didn’t land, and the company shut down. Microsoft put Copilot into their software ecosystem, early feedback flagged it as awkward, poorly integrated, and occasionally creepy from a privacy standpoint. These missteps have made a lot of people skeptical about AI’s readiness for prime time.
John Lovelock from Gartner points out that many companies have pushed early experiments with generative AI as far as they can, for now. That’s expected. With any new tech, companies will sprint at first, trip a few times, and then get smarter about where the real use cases are. According to Gartner, we’ll move out of this disillusionment phase around 2026. That’s when generative AI will stop being a flashy add-on and start delivering serious productivity gains.
If you’re running a business, don’t abandon AI now, it’s not the end of the story. It’s the part where serious builders separate real opportunity from noise. Technologies that matter follow this curve. Generative AI is no exception.
Global spending on generative AI is set to surge despite current setbacks
Even with the critiques and failed experiments, money is still moving toward generative AI. Businesses, investors, and tech leaders see long-term value in these systems, even if short-term execution hasn’t lived up to the early predictions. What we’re seeing is a recalibration, not a retreat.
Gartner predicts spending on generative AI will climb to $644 billion in 2025. That’s a 76.4% increase from 2024. From a business standpoint, this signals more than hype. It’s a shift in capital toward long-term structural capability. Smart leaders are doubling down, not on novelty, but on infrastructure and functionality that scales.
This investment direction makes sense. The initial wave taught us what doesn’t work. Now the focus is on integrating AI where it counts, in operations, service delivery, internal workflows, not just repackaging old products with an “AI” label slapped on them. We’re going to see smarter product design, more selective applications, and a shift from grand demos to measurable outcomes.
If you’re a decision-maker, this is not a time to reduce exposure to the space. It’s time to refine your approach, get better at spotting practical use cases, and make AI enhancements part of your larger digital architecture. The capital markets are seeing what’s on the horizon. Growth won’t be a straight line, but movement over the next five years will be substantial.
The services sector is anticipated to lead generative AI adoption
Generative AI isn’t scaling evenly across all industries. The services sector is moving fastest, and with good reason. Services operate on process efficiency, customized delivery, and scale. Generative AI applications and managed services align well with those needs. Rather than building complex models in-house, companies in this space are finding value by embedding pre-trained AI into business logic, workflows, and client tools.
Gartner forecasts a compound annual growth rate of 132.5% for generative AI applications and 131.8% for generative AI managed services. That’s growth driven by practical deployment, not headline experiments. We’re talking intelligent content creation, document summarization, customer interaction tools, and internal support improvements, basically wherever time and labor can be compressed without sacrificing quality.
Service providers need speed, adaptability, and cost efficiency. AI helps reduce lag between input and output. That could mean helping a legal firm prep documents faster or enabling a consultancy to deliver client-ready research in hours instead of days.
As an executive, don’t wait for some universal AI solution. Focus on key parts of your service delivery chain that are highly repetitive, content-heavy, or customer-facing. Then look at targeted generative AI solutions already available through large providers or consulting partners. You don’t always need to build. You need to select, refine, and deploy fast where there’s clear ROI. The services sector is proving this market is maturing where the business case is strong. Follow value, not noise.
Consumer-Facing generative AI in hardware does not necessarily mirror genuine market demand
Generative AI is rapidly showing up in consumer hardware, smartphones, laptops, and PCs are now shipping with built-in AI functionality. But here’s the catch: that doesn’t mean consumers are asking for it. The push is coming from manufacturers, not from clear usage signals in the market.
John Lovelock, VP Analyst at Gartner, made the point clearly, consumers will end up buying AI-enabled devices simply because they come standard in the next upgrade cycle. When new smartphones or PCs hit the shelves with generative AI included, most buyers aren’t looking for those specific features. They’re just upgrading their devices, and AI happens to come preloaded.
That disconnect matters. When hardware adoption metrics are used as a proxy for consumer enthusiasm, it creates distorted feedback loops. Companies interpret higher unit sales as validation that the AI bet is working, even if users barely touch the features. This misalignment can lead to inefficient R&D spending and product decisions driven by internal innovation goals rather than actual user behavior.
Gartner estimates generative AI on smartphones alone will see a CAGR of 171.6%, with $438.4 billion in spending forecast by 2026. These are massive numbers. But executives need to look deeper. What are people actually doing with those capabilities? What features stick, and which ones are bypassed after the first week? Those are the right questions.
If you’re a product or operations leader, don’t chase trends just because the supply chain is delivering them. Focus on what drives real engagement. If embedded AI adds to product stickiness, fine. If not, reconsider the allocation. Adoption at scale doesn’t always mean value at scale. Know the difference, it saves time, capital, and brand credibility.
Gen AI technology consulting is emerging as a growth area
As generative AI becomes more complex and central to business strategy, one thing is clear: most companies aren’t equipped to handle integration on their own. That gap is creating rapid demand for Gen AI-focused consulting services. Not theoretical advice, hands-on, execution-focused support around architecture, compliance, product alignment, and internal workflows.
This is about bringing in human consultants who understand the changing AI landscape and know how to map it to real business needs. From aligning product features with current AI capabilities to designing responsible data governance policies, organizations are relying on outside experts to avoid expensive missteps.
Gartner projects a compound annual growth rate of 111.8% in Gen AI technology consulting. These aren’t just inflated future bets, there’s real money changing hands right now. Companies are looking for clarity: when to use generative AI, what tools are worth the investment, how to structure implementation, and how to mitigate legal and reputational risks.
For C-suite leaders, this is a practical signal. If your internal teams lack the bandwidth or specialized knowledge to adapt at the pace generative AI is evolving, you shouldn’t hesitate to bring in top-tier consulting partners. That decision can save months of delay, rework, or exposure. But choose partners who deliver implementation, not just frameworks. This market will flood quickly with surface-level advice. Go deeper. Prioritize firms with real-world AI deployment experience and a track record in regulated, complex environments.
Right now, having a clear generative AI integration strategy is an operational advantage. In 18 to 24 months, it will be a basic expectation. Get ahead while the window is still open.
Hardware innovation in generative AI does not automatically guarantee consumer satisfaction
Several recent generative AI hardware launches have failed to meet expectations, despite high investment and media attention. The lesson is straightforward: adding generative AI to a product doesn’t mean users will find value in it. If features are poorly integrated, confusing, or invasive, user trust drops, and adoption declines just as fast.
Microsoft’s Copilot rollout is a prime example. Early users responded with criticism over security concerns and inconsistent behavior. Instead of enhancing productivity, the tool introduced friction. Meanwhile, Humane’s AI pin was positioned as a breakthrough wearable interface. It didn’t deliver, so the company shut it down. These aren’t isolated failures. They show what happens when product teams push AI-enabled features without grounding them in user need.
What matters now is precision. Delivering generative AI through hardware only works when the experience is well-designed, intuitive, and clearly beneficial. If it’s rushed or treated as a marketing layer, users will ignore it—or actively avoid it.
For executives overseeing product portfolios, this is a risk management issue. Focus on signal, not noise. Engage in direct user testing before rollout. If AI features don’t support a clear use case, skip them. And be prepared to scale back or iterate fast post-launch. Frustrated users are vocal, and negative feedback spreads quickly—especially when it involves technologies that impact privacy or workflow.
More broadly, be aware that consumer sentiment around AI is still forming. That means any product experience that disappoints today affects confidence in tomorrow’s launches. Protect the long-term value of your AI efforts by making the short-term experiences solid, respectful, and optional. You don’t earn trust through capability. You earn it through impact.
Businesses need to carefully define the scope and boundaries of generative AI use cases
As generative AI becomes more capable and more embedded in tools we use every day, companies are under pressure to decide where, how, and why to deploy it. You can’t treat generative AI as a universal solution. It needs precise roles and clear limits.
The questions facing executives are practical. Should generative AI be used to handle tedious support tasks to improve speed and consistency? Can it be trusted with core business functions that affect reliability or customer-facing decisions? Is it useful for internal communication tasks like summarizing meetings or composing quick replies? Or is the risk of diluting quality, creativity, or legal safety too high in some environments, like media, design, or regulated industries? These questions don’t have default answers. They need conscious, use-case specific evaluation.
Gartner’s outlook points to 2025 and 2026 as decisive years. This is when companies will be forced to standardize AI policies across departments and make confident calls on what’s in and out of scope. Those who delay may find themselves exposed, either through compliance issues, brand erosion, or inefficiency.
Executives should already be working with technical and legal teams to define boundaries. Decide which teams can deploy generative AI tools and under what constraints. Create guidelines around human oversight, security, and intellectual property use. And most importantly, revisit these frameworks regularly. The technology is moving fast, and rules that made sense a year ago may not serve you six months from now.
This is about shaping AI’s role to serve your operations, not disrupt them. The companies that set clear parameters now will have more control, more agility, and fewer surprises later.
Diverse growth rates across AI segments highlight a varied landscape of opportunity and risk
Generative AI is not growing in a uniform way. Some areas are moving fast, others more slowly. That split matters for planning, investment, and execution. You can’t treat the AI market as one single trend, it’s a layered ecosystem with very different growth trajectories across applications, platforms, infrastructure, and hardware.
Gartner’s data lays this out clearly. The fastest-growing segments include AI on smartphones, forecasted to grow at a 171.6% compound annual growth rate (CAGR), and AI-optimized Infrastructure-as-a-Service (IaaS), set to expand at a 139.9% CAGR through 2028. In comparison, AI-optimized servers are expected to grow at a much slower 34% CAGR, while generative AI applications in software are projected at 64.7%, and generative AI infrastructure at 69.2%.
What this tells us is simple: not all generative AI investments will deliver at the same speed. High-growth segments are mostly consumer-facing and cloud-native, where pace and scale are driven by fast user interaction, broad distribution, and lower deployment friction. On the other hand, hardware and infrastructure grow more slowly due to complex procurement cycles, integration demands, and cost thresholds.
For C-suite leaders, this is about allocation. Resources, whether capital or people, need to be placed where the returns will align with business cycles and product timelines. You don’t need to be everywhere in AI. You need to be in the right places with the right timing.
Use this growth divergence as an internal filter. If you’re in cloud services, telecom, or consumer platforms, you should probably move faster. If your focus is on enterprise hardware or on-premise deployments, plan for slower ramp-up and longer cycles. Align teams accordingly.
Tracking the pace of different AI segments isn’t about chasing momentum, it’s about matching your investments to where real traction is already happening. That’s how you minimize risk and position your company for solid, scalable impact.
Recap
Generative AI is recalibrating. The hype is cooling off, but the investment is accelerating. That’s a signal, not a contradiction. For business leaders, the takeaway is clear: it’s time to move from curiosity to competence.
The next 24 months won’t be about flashy demos or overpromising tools. They’ll be about building real value, executing quieter, smarter bets that align AI to process improvement, customer engagement, and operational lift. You don’t need to use AI everywhere. You need to use it precisely, where the return is clear and the risk is contained.
If you’re making budget decisions, define the scope. Align AI initiatives with business outcomes. Bring in the right partners. And resist the pressure to participate in trends that don’t serve your core goals. This phase rewards discipline over speed.
The companies that focus on structured integration, targeted use cases, and reliable oversight will walk out of this trough with actual leverage—not just AI features, but AI performance. The rest? They’ll have a lot of shelfware.
