After a meteoric rise in artificial intelligence (AI) enthusiasm across the technology and investment landscape, a new mood is overtaking enterprise IT: skepticism and “AI fatigue.” Jefferies senior analyst Brent Thill, speaking on Yahoo Finance’s Market Domination Overtime, painted a nuanced picture of an industry in the midst of a hype cycle, where excitement has so far outpaced actual results.

Enterprise AI: More Hype Than Harvest

Recent studies, including a widely cited report from MIT, have underscored the slow path toward return on investment (ROI) for corporate AI spending. According to Thill, 95% of organizations surveyed reported little or no measurable ROI from AI since adoption began. “Most deployment is still at the experimentation phase,” said Thill, noting much of the AI industry’s current revenue—less than 5%—actually stems from successful AI-driven software products.

This data aligns with findings from top data and AI companies such as Databricks and Snowflake. While consumer-facing AI applications like ChatGPT and Perplexity have captivated the public imagination, enterprise adoption requires overcoming complex hurdles: integrating legacy systems, restructuring workflows, improving data security, and retraining employees.

The Long Road to AI Payoff

Thill drew parallels to earlier technological revolutions—the cloud, broadband internet, and e-commerce—where expectations ran high, but mainstream adoption took several years to deliver. “You didn’t buy everything on Amazon on day one, nor did the cloud replace on-premise infrastructure overnight,” he remarked. For AI, executive commitment remains strong, but companies are shifting from experimentation to a phase of reflection and selective investment.

“We’re in the digestion phase,” said Thill. “Everyone is trying every tool—it’s an AI bake-off. Now, CIOs are taking a breather, assessing what actually delivers business value, and redirecting resources accordingly.” He cautioned that many of Wall Street’s hopes for immediate upside are overblown, predicting a slower burn before a real inflection in enterprise ROI from AI materializes.

Infrastructure: A Safer Bet than App Vendors

Despite underwhelming short-term returns for application vendors, the demand for cloud-based infrastructure has surged. Thill highlighted technology giants like Microsoft and Oracle as well-positioned beneficiaries. “All these apps—current and future—require robust infrastructure. Microsoft Azure and Oracle Cloud are critical enablers for AI and data workloads; thus, investors may find these safer, more reliable bets in the near term than pure AI software startups,” he explained.

This theory aligns with recent market trends. In their most recent earnings reports, Microsoft’s Intelligent Cloud revenue hit $34.3 billion, up 23% year-over-year, with Azure leading the way. Oracle’s cloud infrastructure revenue also surged, powering double-digit growth. These gains suggest that, even as standalone AI apps wait for their breakout moment, the foundational technology layer is already being built and monetized.

‘AI Fatigue’ Sets In, But Commitment Stays Strong

The notion of “AI fatigue”—where businesses and employees grow weary of the endless cycle of announcements, pilots, and demos—has begun to take root. Industry insiders liken the current state to “testing all the cupcakes at a school bake sale,” but few solutions have distinguished themselves in production at scale. Companies must also manage change fatigue among users and IT teams, who are asked to adapt rapidly while results remain ambiguous.

Yet, commitment from IT leaders is undimmed. Global surveys from Gartner and IDC show that 83% of CIOs plan to increase AI investment over the next 24 months, viewing the technology as essential for productivity, customer engagement, and long-term competitiveness. The challenge is shifting from proof of concept to scalable, secure, and trusted implementations.

Barriers: Integration, Security, and Skills Gaps

Three main hurdles are slowing the path to ROI:

• System Integration: Legacy IT environments are notoriously difficult to connect to modern AI systems, requiring both technical overhaul and business process restructuring.
• Security: As AI systems consume vast amounts of data, concerns about privacy and compliance have escalated, necessitating new investments in cybersecurity and governance.
• Workforce Training: Scaling AI tools requires extensive retraining, pulling workers away from their core responsibilities and sometimes triggering organizational resistance.

Successful deployments, experts say, are those with clear business cases, executive buy-in, and iterative, phased adoption strategies. Companies with robust data cultures and agile IT setups—such as financial services, major retailers, and some healthcare providers—are already seeing early wins, but these remain exceptions, not the rule.

Looking Ahead: Patience, Persistence, and Evolution

Thill remains “very bullish on AI in the long term,” forecasting that, as underlying technology and best practices mature, measurable ROI will accelerate. Historical cycles show that as use cases crystalize and infrastructure investments bear fruit, the market typically shifts from skepticism to sustained growth.

For now, the watchwords are patience and pragmatism. Investors are advised to track infrastructure leaders and monitor sector use cases that demonstrate proven business impact. Meanwhile, enterprises must keep a sharp focus on targeted adoption, security, and change management to break through the fatigue and unlock the true value of AI in the years ahead.