AI Demand Strong as Enterprises Shift to Value-Focused Spending

As demand for artificial intelligence (AI) technologies remains robust, a significant shift in corporate spending strategies is being observed. Companies are transitioning from 'tokenmaxxing,' an approach focused on maximizing AI tool usage regardless of cost, to 'valuemaxxing,' which prioritizes measurable business outcomes from these investments. This transformation is driven by rising AI costs and the increasing need for financial leaders to demonstrate the value derived from their expenditures.

Initially, many companies adopted a 'tokenmaxxing' strategy to accelerate AI adoption, encouraging employees to use AI tools extensively. Some executives even promoted high usage rates, with internal leaderboards at giants like Meta and Disney tracking token consumption. However, the costs associated with AI technologies quickly escalated. For instance, Uber reportedly exhausted its entire 2026 AI budget in just four months, Microsoft canceled most of its Claude Code subscriptions, and Walmart imposed limits on AI tool usage. This prompted Chief Financial Officers (CFOs) and Chief Information Officers (CIOs) to treat AI spending like any other operational expense, demanding proof of its tangible impact on delivery, quality, and modernization efforts.

The 'valuemaxxing' approach involves directing simpler tasks to more cost-effective models, reserving advanced models for complex work, and requiring transparent pricing and sustainability from vendors. Technology giants like IBM (IBM) are also advocating for this value-driven strategy. This shift is seen as part of companies' efforts to achieve a genuine return on their AI investments.

These developments have led to significant volatility in AI-related chip stocks. Wall Street began questioning the sustainability of record AI capital expenditures, triggering a sell-off in semiconductor stocks that wiped out over a trillion dollars in market value. The Philadelphia Semiconductor Index (SOX) and the VanEck Semiconductor Index experienced substantial declines. Despite the sell-off, some analysts view this as a 'mid-cycle reset' and maintain bullish price targets for chipmakers like Nvidia (NVDA) and Micron (MU), citing strong earnings growth and attractive valuations. Notably, demand for high-bandwidth memory (HBM) remains strong and outstrips supply, with customers securing multi-year supply commitments in advance. Micron's CEO highlighted 'unprecedented' memory demand and announced a $250 billion U.S. investment plan. However, Morgan Stanley cautions on chip stocks, citing concerns over limited pricing power as major cloud service providers increasingly design their own cost-effective chips.

AI infrastructure spending has become a major driver of the U.S. economy, accounting for approximately 60% of recent economic growth. This year, AI infrastructure spending is projected to exceed $700 billion, raising questions about whether these massive investments will ultimately justify lofty valuations. According to Goldman Sachs, companies like Amazon (AMZN), Microsoft (MSFT), Alphabet (GOOG, GOOGL), and Meta Platforms (META) are collectively expected to spend around $757 billion on AI infrastructure in 2026. As the cost of remaining competitive in AI rises, companies are increasingly shifting from free cash flow to debt issuance and equity for funding.

Analysts and market experts interpret the recent sell-off in the semiconductor sector as a 'mid-cycle reset'. Morgan Stanley maintains Nvidia as a top pick, confident that diversified demand from enterprise, sovereign AI, and industrial customers will continue to drive revenue growth. Micron's CEO anticipates a 'sustained, substantial multi-decade memory demand cycle' beyond current AI data center demand, potentially fueled by humanoid robots. Goldman Sachs projects that competitive pressures could push AI spending even higher, potentially exceeding $920 billion in 2027, surpassing current market expectations. Nevertheless, concerns persist regarding valuations reminiscent of the dot-com bubble era and overbought market conditions.