 |
| Original illustration by F. Stop Fitzgerald, PillartoPost.org |
Are AI fears overdone?
GUEST BLOG / By Angelo Kourkafas, Senior Global Strategist, Investment Strategy, CFA, EDWARD JONES COMPANY.
Key Takeaways:
--AI sentiment remains split, with strong earnings and rapid adoption offset by concerns about overbuilding, capital intensity, and business model disruption. ------
--Software valuations look unlikely to revisit prior peaks, but we don't think AI will render the industry obsolete. The performance gap between software and semiconductors has become extreme-potentially stretched enough to support a snap back.
--Labor market fears appear overstated, with historical patterns of creative destruction suggesting AI will likely reshape, not collapse, employment. AI driven disruption can bring short term pain but long term gains.
--Diversification may be an advantage as the AI cycle advances, fundamentals stay solid, and market leadership broadens across sectors and geographies.
Artificial intelligence (AI) continues to dominate market conversations, with sentiment whipsawing between dystopian fear and innovation-led optimism. On one side are the disruption concerns: rapid AI advancements, compressing valuations, challenging legacy business models, and raising concerns about knowledge worker displacement. On the other side stands the promise of higher productivity and an earnings backdrop that continues to beat expectations and, in many cases, accelerate.
Caught between these opposing narratives are U.S. large cap stocks, which have traded essentially sideways for four months. The range between this year's high and low in the S&P 500 is one of the narrowest on record, masking wide dispersion beneath the surface as weakness in tech has been offset by strength in other parts of the market*. We offer the following perspective on what's driving the current wave of AI anxiety, assess whether these fears are warranted, and outline ways to help "AI proof" portfolios.
The AI race is on, but the buildout comes at a price
NVIDIA, the world's most valuable company by market capitalization, reported quarterly results last week. With a valuation above $4.7 trillion and an 8% weight in the S&P 500, its results are a bellwether for AI and were among the most anticipated of the earnings season. As has been the case for several quarters, NVIDIA delivered record revenue, beat analyst expectations, and issued strong guidance, noting that computing demand continues to grow exponentially and that enterprise adoption of AI agents is accelerating.
Yet despite the strong fundamentals and confirmation that the AI infrastructure cycle remains in full swing, investors were not impressed. Shares fell 5%, reflecting growing concerns that AI demand may be overheating. Mega cap technology companies continue to pour capital into chips, data centers, and power capacity, with AI spending expected to reach roughly $700 billion this year, nearly doubling last year's levels. But skepticism is rising around the pace of this spending, the step up in capital intensity, and the ultimate returns these investments will generate.
We see two main takeaways from NVIDIA's results and the market's reaction:
--AI infrastructure demand remains strong, with no signs of slowing, and earnings for companies supporting the buildout remain well anchored.
--Investor skepticism around the cost, pace, and durability of spending is likely to keep a lid on tech valuations at this stage of the AI cycle. The silver lining, in our view: this caution, and the absence of euphoria, suggests we are not in bubble territory.
Disrupt, replace, or complement?
Another major pillar of the AI wall of worry isn't whether AI will create value-it's whether it will create too much of it, too quickly, disrupting existing business models and reshaping labor dynamics.
1) Software in the eye of the storm
Nowhere is this tension more visible than in the software subsector, which has declined 30% from peak and seen its weight in the S&P 500 decline from 12% down to 8%. The worry is that the new AI tools and agent releases are arriving at an accelerating pace, and this wave of automation may intensify competition, pressure pricing, or in the worst case, render parts of legacy software suites obsolete.
Yet we think it is a leap to conclude that AI will render the entire software industry obsolete. Over the past 100 days, the divergence between semiconductors and software has been extreme, in our view, potentially stretched enough for a snap back.
Recent announcements from Anthropic that position its tools as complements to existing systems rather than replacements (integrating with Slack, Gmail, Intuit, DocuSign, LegalZoom, FactSet) may help calm near term disruption fears.
Also, many leading software companies are investing aggressively in their own AI strategies, positioning themselves to adapt rather than be displaced. We think it is likely unrealistic to expect sentiment around disruption to fully reverse, but a tactical rebound is likely in the weeks ahead in parts of the market that have been penalized the most.
2) AI may reshape, not destroy, the labor market Beyond company level disruption, the debate has shifted to the labor market, especially after a viral fictional essay argued that rapid AI adoption could displace white collar jobs in a self reinforcing loop. We think those fears are overstated.
That said, we think it would be pollyannaish to suggest there will be no impact.
Some studies do find links between AI exposure and unemployment, with occupations that embraced AI most intensively, such as software development, showing the largest unemployment gains. Announced job cuts have picked up in recent months. Still, according to the Challenger report, AI was cited as a reason for 7% of layoff plans announced in January and 4.5% for 2025 overall, suggesting AI is meaningful but not the primary driver.
Some of the recalibration likely reflects post pandemic over hiring as well. We think history offers some useful perspective. Past technological revolutions, from electrification to personal computing, followed a familiar pattern of creative destruction. Initial disruption gave way to major productivity gains, faster economic growth, and higher living standards.
While AI may displace certain tasks more directly than earlier technologies, its impact on productivity and profitability tends to lower prices, boost investment, and increase consumption.
These forces tend to create demand for new types of work, just as they have in previous cycles. Notably, about 60% of today's jobs did not exist in 1940, underscoring how innovation reshapes labor markets rather than eliminating them.
In addition, AI tools and lower service costs may help accelerate entrepreneurial activity. The recent increase in new business applications is encouraging.
Bottom line: We think AI driven disruption can bring short term pain but long term gains. Productivity improvements from AI should ultimately help support growth and create new opportunities. For now, jobless claims show no clear signs of labor market weakness. We expect low hiring-low firing conditions to persist, with any downside surprise in employment or inflation likely to open a path to more aggressive Fed rate cuts. Potentially that's the signal the bond market is sending with the 10-year Treasury yield falling below 4%.
Considerations for helping AI-proof portfolios As highlighted in Edward Jones Company annual outlook, we think AI and innovation matter-but so does diversification, which has shown its value this year as market leadership has rotated significantly. Because AI represents both a risk and an opportunity, and because it is difficult to predict which companies will emerge as long term winners, we recommend the following: Own the full AI value supply chain.
Rather than concentrating exposure in a handful of mega-cap names, investors may consider diversification across the broader ecosystem: Compute infrastructure (semiconductor companies), Cloud providers (platforms that host and distribute AI models), and Enterprise software (AI tools embedded into core business workflows).
Add sectors that could benefit from AI productivity. As AI moves from development to deployment, its benefits should expand beyond large-cap tech. Potential beneficiaries include Industrials (automation, robotics, predictive maintenance); Health care (drug discovery, diagnostics); and Financials (fraud detection, risk management, customer service). Balancing growth with value. Value stocks have begun to regain leadership this year and may continue to perform well if the global manufacturing cycle accelerates. Real asset and "old economy" sectors, such as energy and materials, carry lower perceived AI disruption risk and can add resilience to diversified portfolios.
Diversifying internationally.
AI is not only a U.S. story. Global beneficiaries include Taiwan and Korea (semiconductors), Japan (robotics and automation), and Europe (industrial automation), offering additional opportunity and valuation diversification, in our view. Not losing sight of the big picture While the pendulum around AI may continue to swing between fear and excitement, we believe investors should not lose sight of what we think remains a positive backdrop for stocks.
Corporate earnings continue to strengthen within and beyond AI, the U.S. economy appears to be on track for another year of solid growth, employment remains stable, the Fed maintains a bias toward easing, market seasonality may improve in the months ahead, and market leadership is broadening. In this phase of the AI cycle where growth is real but skepticism tempers valuations, diversification may provide an edge.