News Analysis
Sam Altman would very much like you to know that we are in a bubble. Or at least, a sort-of bubble. Or maybe something with the contours of a bubble but with the potential to become, over time, something more like the beginning of a long economic epoch.
“When bubbles happen,” the OpenAI CEO said recently, “smart people get overexcited about a kernel of truth.” It’s a neat way of saying what a lot of tech leaders are now reluctantly admitting: we might be racing toward something real, but we’re also definitely sprinting.
Others agree. Alibaba’s Joe Tsai and Bridgewater’s Ray Dalio both raised red flags about the scale and speed of AI investment, warning that the moment has shifted from sober innovation to late-stage exuberance. Last week, the market seemed to catch a chill. Nvidia, AMD and other AI-linked stocks took a hit. Many pointed to Altman’s bubble talk as the cause.
What’s left is the question nobody wants to answer directly: What if we already know this is unsustainable?
Is AI in a Bubble — or Just Early?
"Most companies are exploring AI because they feel they have to, not because they know what to do with it.”
- MIT's "The GenAI Divide: State of AI in Business 2025"
If the term “AI bubble” once sounded like a fringe worry, it’s now part of the mainstream discourse. Altman publicly acknowledged the possibility, framing the current moment as one where enthusiasm may have jumped ahead of reality. This is not the first time a technology cycle has moved faster than its infrastructure, but it is one of the most expensive. The AI sector has attracted billions in venture funding, driven by optimism around language models, copilots and synthetic media tools.
This month, the conversation turned from theory to consequence. A study from MIT, "The GenAI Divide," reported that only a small fraction of enterprise generative AI pilots have delivered tangible business impact. Many projects remain stuck in experimentation, either because of unclear use cases or a lack of operational fit. As the report noted, “Most companies are exploring AI because they feel they have to, not because they know what to do with it.”
Related Article: AI in the Real World: What’s Changing at Work, Online and at Home
Wall Street’s AI Reality Check
Investors noticed. After months of uninterrupted momentum, tech stocks tied to AI infrastructure and applications wavered. Analysts pointed to a growing gap between the size of the bets being placed and the maturity of the tools on offer. The sell-off was modest but symbolic. It suggested that markets are beginning to distinguish between theoretical potential and functional value.
Dr. Sarah Kreps, Director of the Tech Policy Institute at Cornell University, said AI does look like other bubble-ish moments. “Money is pouring in, virtually every startup is pivoting to AI and there’s a fear-of-missing-out dynamic. That herd behavior can create distortions — capital chasing AI because everyone else is, not necessarily because each new application has a sustainable business model. That was the dynamic in the dot-com era, when companies were valued sky-high just for having ‘.com’ in their name.”
At the same time, industry leaders continue to make large, directional investments. Some see short-term volatility as a natural correction, a pause in a longer trend rather than the end of one.
“Unlike some earlier bubbles, AI is already producing tangible advances,” Dr. Kreps added. “It’s driving measurable gains in productivity for some firms, and the core technology is here, it isn’t vaporware. The uncertainty is in the scale of its economic payoff and the distribution of that payoff across sectors. That’s where the bubble question bites: are investors betting on a future that’s much bigger than what AI can actually deliver?”
Speculation vs. Substance: What’s Really Driving AI Value
"LLMs won’t get us to AGI, and it’s hard to justify the enormous valuations in light of that fact."
- Dr. Gary Marcus
Emeritus Professor of Psychology and Neural Science, NYU
AI investment divides into two clear tracks. On one side are the speculative plays: startups building copilots, chatbots and synthetic media tools. These companies attract headlines and capital, but most have yet to establish sustainable revenue.
On the other side is the infrastructure that supports them. Semiconductor firms, data center operators and cloud providers already capture steady demand. Nvidia dominates the GPU market, and hyperscale providers report record requests for AI capacity. SoftBank is committing billions to large-scale data centers, notably through a $500 billion Stargate data center project with OpenAI and Oracle.
For investors, the distinction is critical. Consumer-facing applications may deliver bursts of growth but also carry the highest risk of collapse when sentiment shifts.
Dr. Gary Marcus, Emeritus Professor of Psychology and Neural Science at NYU, offered this assessment: “LLMs won’t get us to AGI, and it’s hard to justify the enormous valuations in light of that fact. LLMs will continue to exist, but in my opinion the companies that sell and develop them are overvalued.”
Infrastructure, by contrast, aligns with enterprise adoption and the physical requirements of scale. Current volatility is testing both sides of this divide.
Related Article: How Real Is AI Washing? 4 Companies — and 1 Rock Band — Caught Faking It
The Next Phase of AI: Post-Bubble Strategy
A market correction would change the shape of AI’s growth. Companies with untested products could struggle, while firms with proven models and infrastructure would gain ground. Demand for chips, cloud capacity and large-scale data centers continues to expand, and those sectors remain closely tied to real adoption.
Enterprises face a different set of pressures. Most pilots still fall short of measurable ROI, yet leadership teams feel compelled to build strategies around AI. This is creating a shift toward projects with clear business cases and operational fit, rather than broad experimentation.
For investors and operators alike, a few indicators will matter most:
- Earnings reports from chipmakers and cloud providers, which track demand for infrastructure
- ROI data from enterprise deployments, which show progress beyond pilot phases
- The scale of new data center projects, which reflect long-term confidence in AI workloads
Dr. Kreps described this as a paradox that companies cannot escape. “There’s a real risk of oversaturation. Not every AI startup will survive, and many investors will lose money. On the other hand, ignoring AI entirely is not a safe strategy either. If you’re a company that sits out, you risk being left behind by competitors who do successfully integrate AI. That’s what drives this paradox: companies feel compelled to play the game, even knowing that many bets won’t pay off.”
The next phase will depend on how quickly companies move from experimenting with AI to embedding it in core operations.
Webinar
Oct
16
Agentic AI Playbook: Real-World Customer Service Use Cases You Can Deploy Now
Boost self-service by 30% and slash call volume by 63% with agentic AI.
Register
Webinar
Oct
22
Beyond Storage: Smarter Content, Bigger Impact with DAM + AI
Discover how the DAM + AI duo makes content smarter, stronger and more accessible.
Register
Webinar
Nov
6
Fix the Content Bottleneck: Build a Better WebOps Strategy
Content stalled? Dev overloaded? You’re not the only one. Learn how streamlined WebOps bridges the publishing gap.
Register
Webinar
On demand
CMS Briefing: A Live Look at What’s Next in AI-Driven Platforms
Learn how leading organizations are using AI‑driven tools to publish faster, personalize smarter and stay secure.
Watch Now
Webinar
On demand
Ready or Not: How Data-First Organizations Are Unlocking Agentforce Potential
Learn how to cut through the noise, activate Agentforce and build a Salesforce AI strategy that actually delivers.
Watch Now
Webinar
On demand
AI in Customer Service: Faster Resolutions, Happier Customers
Don’t let rising demand burn out your team. See how to build a smarter, more resilient support org.
Watch Now
From Gold Rush to Gut Check
Everyone wants to know if this is a bubble. The truth is less dramatic and more inconvenient: most of the money going into AI is chasing things that do not work yet. The MIT study showed it clearly: 95% of enterprise pilots failed. That means executives bought the promise, spent the money and came away with little to show.
Yet the spending continues. Data centers are being built at a historic pace. Nvidia keeps selling every GPU it can make. SoftBank is planning a half-trillion-dollar project on the assumption that demand will keep growing. These are signs of an industry straining to keep up with its own ambitions.
The bubble conversation matters because bubbles change behavior. They attract people who want to get rich quickly, and they punish those who arrive too late. If AI is in that kind of moment, then the real winners will be the ones who survive the eventual correction with something useful still in hand.
