AI Capex Skepticism Rises and OpenAI Fires Back at Google

Investor Skepticism Mounts Over an AI and Data Center CapEx…Bubble?

The AI narrative has drifted toward skepticism on multiple dimensions in the past week. Perscient's semantic signature tracking language expressing concerns that AI capital expenditure represents a risky bet with uncertain payoffs rose by 0.3 to a z-score of 1.2, while the related signature measuring skepticism about realized productivity gains from AI strengthened by 0.3 to reach 2.0. These movements reflect a growing chorus of voices questioning whether the staggering sums being poured into AI infrastructure will ever generate commensurate returns.

The latest Bank of America Global Fund Manager Survey, conducted November 7-13, revealed that investors have warned companies are "overinvesting" for the first time in two decades. This sentiment gained particular traction after researchers at the Massachusetts Institute of Technology found that 95% of organizations saw zero return on their AI investments. BCA Research has articulated what it calls the "Metaverse Moment" thesis, predicting that a shift in market sentiment will occur when a major AI company announces increased capital expenditure only for its stock price to subsequently decline, signaling a transition from enthusiasm (or expectation) to a more skeptical assessment.

In keeping with the theme, Perscient’s signature tracking the language of corporate skepticism toward big AI investments rose by 0.27 to a z-score of 1.90. SoftBank, which holds exposure to AI companies including OpenAI, Arm, and Perplexity, saw its stock price fall by 40 percent in November, with nearly $50 billion in market capitalization disappearing within days. According to the Financial Times, approximately $1 trillion in stock value has been wiped from several of Silicon Valley's heaviest hitters, including Oracle, Meta, Palantir, and Nvidia over the past week.

The comparison between current AI spending and historical technology bubbles has become increasingly prominent. Our semantic signature drawing parallels between AI capex and fiber construction during the dot-com boom remained flat this week, not that it matters, since it registered at a z-score of 4.63. As CNBC reported, these concerns have now spread to the bond market, where investors are growing uneasy about the rapid rise in public debt used to bankroll AI investments. Reuters noted that while leverage across most major companies remains low for now, the sheer scale of borrowing has begun to strain market confidence.

Yet the picture remains complicated. The signature measuring expectations for continued hyperscale infrastructure growth remains elevated at a z-score of 3.70, reflecting ongoing coverage of expansion plans even as doubts mount. This tension between skepticism and continued investment creates an uncertain environment for market participants trying to assess whether AI spending represents the foundation of a new economic era or an echo of past technology bubbles.

Anthropic Emerges as Competitive Force as AI Race Narrative Intensifies

While investors debate the sustainability of AI infrastructure spending, the competition among frontier AI laboratories has reached a new intensity. The semantic signature tracking language positioning Anthropic as leading the AI competition rose by 0.32 over the past week to a z-score of 2.53, now just 0.09 below its peak, and following closely on the heels of Google’s Gemini 3 moment just the week before. This surge coincides with Anthropic's Monday announcement of Claude Opus 4.5, the company's third major AI model launch in two months.

The release represents a milestone for the startup founded by former OpenAI employees in 2021, arriving on the heels of Microsoft and Nvidia investments that valued Anthropic at $350 billion. VentureBeat reported that Anthropic slashed prices by roughly two-thirds while claiming state-of-the-art performance on software engineering tasks. Perhaps most striking, the new model scored higher on Anthropic's most challenging internal engineering assessment than any human job candidate in company history.

The semantic signature tracking language asserting that OpenAI leads the AI competition remained flat at a z-score of 2.07, still stronger than average but lower than Anthropic's rising narrative presence. Google's position, as measured by the signature tracking language positioning Gemini as leading, likewise stayed flat at 1.51. Opus 4.5 will face stiff competition from other recently released frontier models, most notably OpenAI's GPT 5.1, released November 12, and Google's Gemini 3, released November 18.

This rapid-fire product development reflects the breakneck competition in AI. The September launch of Claude Sonnet 4.5 was followed by October's Claude Haiku 4.5, and now November's Opus 4.5. Such a release cadence would have been unthinkable even a year ago. The Verge noted that within weeks, the three titans of frontier AI introduced their most advanced models to date, representing a transition toward systems that can pause, reason carefully, and operate as long-running agents.

The Death of Consulting Narrative Just Won’t Die

We noted this a couple weeks ago, but the corporate competition driving new model releases continues to find its “efficiency” mechanism most acutely in the industry that is in the business of finding such, well, excesses in its clients’ operations: the management consulting industry. Perscient’s semantic signature tracking the density of language predicting that AI will eliminate consulting industry positions rose once again, this time by 0.2 over the past week to a z-score of 1.7.

Some of this is less narrative and more simple reporting of news. McKinsey & Co. cut about 200 global tech jobs in the past week as the consulting firm joins rivals in using artificial intelligence to automate some positions. The company is not ruling out additional reductions across different functions over the next two years. Two senior executives at Big Four firms estimated that, across the UK's largest consulting and accounting firms, graduate recruitment would be down by about half in the coming year, with some of that reduction reflecting anticipation of AI's impact.

The consulting sector is not alone. In just the last few months, PricewaterhouseCoopers let go 1,500 US workers, Accenture announced ongoing reductions, KPMG laid off 4% of its US audit workforce, and Microsoft has completely halted hiring in its consulting business. Big Think posed the question directly: if the whole economy is now being shaken by artificial intelligence, the consulting industry is facing a tsunami, with questions arising about whether consulting firms will need to employ as many analysts if AI can perform much of the research, synthesis, and analysis.

The upheaval is prompting fundamental reconsiderations of traditional staffing models. Some industry observers are betting on an "obelisk" structure with fewer layers and less reliance on junior staff, while others predict an "hourglass" pinched in the middle as AI automates mid-level routine tasks. The semantic signature tracking language arguing that AI competency is required for employment rose to a z-score of 1.12, reflecting media emphasis on AI proficiency as a requirement for professional employment.

Right now, this seems like a focused phenomenon with its sights trained squarely on management consulting and adjacent fields. For example, the broader signature measuring language claiming that AI will fundamentally change professional office jobs remained flat at a z-score of 0.61, while the signature tracking language asserting AI will generate novel employment categories stayed at 0.35. Elevated, but not in the same zip code, and remaining fairly stable. Employment growth in industries such as marketing consulting, graphic design, office administration, and telephone call centers has fallen below trend amid reports of reduced labor demand due to AI-related efficiency gains.

Still, Fortune reported this week that MIT researchers now estimate that AI can already replace nearly 12% of the US workforce, suggesting the window to treat AI as a distant future issue is closing. But given the growing skepticism about AI more broadly that is showing up in our narrative monitoring, it seems reasonable to question how much appetite companies, employees, and policymakers will have for this technology to displace jobs when they can’t be written off as junior consultants from McKinsey.