AI Investment Enters a Reality Check as Risk Appetite Cools

The enterprise AI boom isn’t slowing, but the way companies are investing in it is changing. After an initial wave of aggressive spending driven by hype and fear of missing out, new data suggests that investors and enterprise leaders are becoming far more cautious, measured, and disciplined in how they approach AI.

According to recent data from Solvd CIO & CTO AI Research 2026, which surveyed 500 US CIOs and CTOs at large enterprises, organizations are still investing in AI, but with tighter scrutiny, clearer expectations for returns, and a growing willingness to walk away from underperforming projects.

So much of today’s conversations around AI are focused on the extremes, but the situation on the ground is far more nuanced.

“So much of today’s conversations around AI are focused on the extremes, but the situation on the ground is far more nuanced,” says Mike Hulbert, CEO of Solvd.

“The reality is, most companies are still in active experimentation mode; only 20% have found high-value use cases for AI at this stage. As the market matures, we have to rethink traditional IT approaches. Solvd is helping to fill these gaps with deep expertise, experience, and strategic execution.”

From Hype to Discipline

The numbers reflect a market that is pulling back from unchecked optimism. Nearly half (49%) of technology leaders say expectations around AI are becoming more data-driven and less hype-based. More tellingly, 72% of companies say they are likely to shut down AI projects that fail to meet KPIs within the next year.

Read more: How AI is Changing Business: Hybrid AI is Coming

This marks a fundamental shift in risk appetite. Instead of treating AI as a long-term experimental bet, enterprises are increasingly managing it like any other capital investment subject to performance metrics, accountability, and return thresholds.

Boardroom pressure is also rising. According to the report, 82% of CIOs and CTOs say boards are questioning the scale of AI spending. This scrutiny is forcing companies to prioritize efficiency over expansion, and outcomes over ambition.

Failure Is Forcing Caution

The cooling of risk appetite is also being driven by experience. AI hasn’t delivered uniformly and enterprises are feeling that reality firsthand. Many enterprises are struggling to realize significant benefits from AI, with only a small group seeing real value.

The Solvd study finds that 80% of companies have experienced at least one AI project failure due to lack of visibility and oversight. Even among those seeing returns, 70% describe outcomes as only “small to moderate” ROI.

This aligns with broader industry trends. Companies like IBM and Accenture are doubling down on governance and measurable ROI. IBM has declared “the era of AI experimentation is over,” while Accenture is positioning itself around secure, scalable deployments, indicating that enterprises are no longer willing to fund open-ended AI bets.

Similarly, several high-profile generative AI pilots across industries, from retail chatbots to internal copilots, have been scaled back or restructured after failing to deliver meaningful productivity gains. Even large tech adopters are quietly shifting budgets from exploratory AI initiatives to those tied directly to cost savings or revenue impact.

Experimentation Continues but with Guardrails

Despite the caution, companies are not abandoning AI. In fact, 90% of CIOs and CTOs still expect increased investment in innovative AI initiatives. The difference lies in how that investment is being deployed.

Organizations are now balancing experimentation with discipline. As the Solvd data shows, 100% of companies have begun establishing AI governance frameworks, while 66% are taking a proactive approach to governance, and 50% say their governance models are still evolving. This suggests that enterprises are building the internal structures needed to reduce risk before scaling further.

At the same time, adoption remains slower than expected. Three-quarters of respondents believe that 50% or less of their workforce will use AI daily by the end of 2026, highlighting the gap between investment and real-world integration.

External Dependence Reflects Market Uncertainty

Another sign of reduced risk appetite is the reliance on external expertise. More than half (59%) of the companies are leveraging cloud provider AI services, while many depend on consultancies and specialized AI firms to implement initiatives.

This reflects a fragmented ecosystem where enterprises prefer to partner rather than build from scratch, reducing execution risk in an uncertain landscape.

Cloud leaders like Microsoft, Amazon Web Services, and Google Cloud are benefiting from this shift, as companies opt for managed AI services instead of investing heavily in proprietary infrastructure.

A More Mature AI Economy

Still, AI adoption is not slowing down but maturing. Enterprises are no longer chasing AI for its own sake, they are demanding proof.

The reality is, most companies are still in active experimentation mode; only 20% have found high-value use cases for AI at this stage.

As Hulbert notes, the market is far from binary. Companies are neither all-in nor pulling back entirely. Instead, they are recalibrating.

“The reality is, most companies are still in active experimentation mode; only 20% have found high-value use cases for AI at this stage,” he says.

Read more: Why legacy systems are the real AI bottleneck in the mid-market

That reality is reshaping investor behavior. Even though risk-taking hasn’t disappeared, it has become conditional. AI investments now need to justify themselves quickly, operate within governance frameworks, and deliver measurable impact.

In effect, the AI gold rush is giving way to a more sustainable phase, where capital is still flowing, but far more carefully.