ICR's Anton Nicholas: AI Is Now a Show-Me Business, and 2026 Is the Reset

Anton Nicholas at Nasdaq: investor focus has moved from AI curiosity to deployment

"56% of CEOs report no measurable revenue increase or cost reduction from AI investments" 56% of CEOs report no measurable revenue increase or cost reduction from AI investments. That is the reset. In 2026, investors are no longer paying for AI curiosity. They want evidence that AI is producing something useful for revenue, costs, or product value.

From exploration to deployment

Earlier this month at Nasdaq MarketSite, ICR's Anton Nicholas said investor focus has shifted from whether companies are exploring AI to how effectively they're deploying it. That changes the scoring system. A compelling AI narrative used to earn patience. Now management teams are expected to show measurable AI strategies.

Why the gap between story and proof matters more now

Recent 2026 ROI data points to a broad execution problem, not just a technology problem: only 12% of CEOs report achieving both revenue growth and cost reduction from AI. That helps explain why credible deployers may be more protected than companies still asking the market to wait while pilots multiply.

BCG and Deloitte data explain why AI spend is rising even as ROI lags

The strange sight on today's balance sheets is easy to miss if you only look at headlines: companies are spending more on AI even though early results are mixed. That is not random. It looks more like fear of falling behind than confidence that the business case is already settled.

Why the wallet is opening anyway

BCG found that corporations expect to double AI spending from 0.8% to about 1.7% of revenue in 2026. At the same time, 56% of CEOs report no measurable revenue increase or cost reduction from AI. The gap looks irrational until you look at who is making the call. Nearly three quarters of CEOs say they are their organization's main decision maker on AI, and half say their job is on the line if AI does not pay off.

When leadership treats AI as a survival issue, budget often keeps moving forward even while the cash-in-register evidence is still thin. In plain English, this looks less like "we have the proof" and more like "we cannot let a peer get ahead while we sit this out."

Where the spend is getting stuck

AI is not just software. It is a change program that runs into old workflows, data plumbing, and operating habits. 72% of AI projects in infrastructure and operations fail to fully meet ROI expectations, and 41% of organizations cite data access and integration as the top barrier to progress. That helps explain why spending can rise while profit impact remains hard to see.

For investors, this is the practical watchpoint. Rising AI spend is not automatically bullish. If the money is going into scattered tools and endless pilots, it may behave more like accumulated overhead than a clean profit lever. If the spend is concentrated where process pain is highest and leadership is forceful, today's weak results may be the messy middle rather than the final verdict.

The bull case is longer payback, not no payback

Bulls do not need every project to pay back next quarter. They need investors to understand that AI may take longer to show up than older software categories did. A recent Deloitte survey found most respondents still expected satisfactory ROI within two to four years, even though that is far slower than the seven-to-12-month payback investors often expect from technology investments. Only 6% reported payback in under a year.

So the real debate is not whether AI works at all. It is about timing, discipline, and whether companies are using the extra time to fix processes and data before scaling.

How to split public AI names while the market resets its expectations

Split the universe by deployment, not messaging

The practical move is to separate public AI names into two buckets before the next earnings cycle hardens the market's view: sellers still marketing capability without deployed results, and buyers that can show AI changing the P&L or the product. The tension is sharp because boards and investors want patience for real transformation, but not endless patience. Investors are already asking for positive returns within six months or less, while focus has shifted from whether companies are exploring AI to how effectively they're deploying it.

Sellers: story logic is weakening

I would be more inclined to sell or avoid exposure where AI still looks like a portfolio of interesting pilots, vendor demos, or strategic ambition. The warning sign is not spending by itself. It is spending without a clear path from test to scaled use case to measurable business impact. That fits the broader lesson that AI success takes nuance and planning, not a catch-all "we're doing AI" message.

Buyers: operating programs deserve more patience

The better bucket contains companies treating AI as an operating program rather than a press-release theme. Watch for three things:

• clear workflow fit, not just model access
• controlled pilots that have already moved into production
• management commentary that ties AI to auditable process or financial outcomes

That is also why management credibility matters. The fact that the new CEO Anton Nicholas is now representing ICR underscores a simple point: investor dialogue is moving toward execution, clarity, and follow-through, not polished ambition alone.

What would change this view

This framework gets harder to defend if AI spending starts producing durable revenue growth, stronger competitive positioning, or sustained cash generation. Some companies may still be in a longer build phase, with satisfactory ROI within two to four years. But until that shows up more broadly, the cleaner approach is to favor deployers over dreamers: firms with good workflow fit, sensible payback assumptions, and a credible path from implementation to impact.