Alex Davis, CEO of Austin-based investment firm Disruptive Tech, has issued a stark warning about the U.S. data center market, predicting a significant financing crisis by 2027-2028. In an end-of-year letter to investors on Monday, Davis expressed deep concerns about speculative data center development that could threaten the AI infrastructure ecosystem.
Davis, whose firm has invested nearly $350 million in AI hardware startup Groq, criticized the “build it and they will come” strategy employed by speculative landlords constructing data centers based on anticipated future demand. He emphasized that hyperscalers—the largest cloud computing companies—will ultimately own their own data centers, leaving speculative developers vulnerable.
Disruptive Tech’s portfolio includes major technology companies like Airbnb, Spotify, and Slack, as well as AI-focused firms such as Shield AI, Palantir, and Databricks. The firm’s substantial investment in Groq came ahead of the AI startup’s $750 million capital raise announced in September, followed by a $20 billion licensing deal with Nvidia in November.
In his investor communications, Davis warned that AI has spawned “too many business models with no realistic margin expansion,” adding ominously that “this will not end well.” His concerns focus specifically on the financial stress that speculative data center landlords will place on the system, stating his preference to “back the owner/users, not the speculative landlords.”
The warning comes amid unprecedented data center construction across the United States. A Business Insider investigation revealed that by the end of 2024, companies had filed permits for 1,240 existing or planned data centers—nearly four times the number filed in 2010. This explosive growth is driven by the massive computational demands of AI systems and large language models.
These facilities, which store, process, and distribute enormous amounts of data required for AI training and inference, consume massive quantities of land, water, and electricity. Cities have accelerated construction rates by offering tax incentives to attract data center development, creating what Davis views as an unsustainable speculative bubble in the infrastructure supporting the AI industry.
Key Quotes
I am also deeply concerned about the ‘speculative’ data center market. The ‘build it and they will come’ strategy is a trap.
Alex Davis, CEO of Disruptive Tech, warned investors about the unsustainable approach of building data centers based on speculation rather than confirmed demand, highlighting risks in the AI infrastructure market.
If you are a hyperscaler, you will own your own data centers.
Davis explained why speculative data center landlords face existential risk—the largest technology companies that drive demand will ultimately build their own facilities rather than lease from third-party developers.
We want to back the owner/users, not the speculative landlords, and we are quite concerned for their stress on the system.
The Groq investor outlined his firm’s investment strategy, emphasizing preference for companies that own and operate their own infrastructure rather than speculative real estate developers.
AI has spurred too many business models with no realistic margin expansion. This will not end well.
Davis expressed broader concerns about AI market sustainability beyond just data centers, suggesting many AI-related business ventures lack viable paths to profitability.
Our Take
Davis’s warning represents a rare moment of public skepticism from within the AI investment community, which has largely maintained bullish sentiment despite mounting infrastructure costs. His position is particularly noteworthy given Disruptive Tech’s massive $350 million stake in Groq, demonstrating that even committed AI investors recognize systemic risks.
The predicted 2027-2028 timeline is significant—it suggests the crisis won’t be immediate but will emerge as financing terms mature and speculative projects fail to achieve projected occupancy rates. This could create a bifurcated market where well-capitalized hyperscalers like Microsoft, Google, and Amazon thrive with proprietary infrastructure while smaller players struggle to access affordable computing resources.
The comparison to previous technology infrastructure bubbles is unavoidable. The telecom overcapacity crisis of the early 2000s followed similar patterns of speculative overbuilding. If history repeats, we may see consolidation, bankruptcies, and a reset in data center valuations that ultimately reshapes the AI industry’s competitive landscape.
Why This Matters
This warning from a major AI investor signals potential turbulence ahead for the infrastructure underpinning the artificial intelligence boom. Data centers are the backbone of AI development, housing the powerful computing systems needed to train and run large language models and other AI applications. If Davis’s prediction of a 2027-2028 financing crisis materializes, it could significantly impact AI companies’ ability to scale operations and access necessary computing resources.
The concern about speculative overbuilding reflects broader questions about AI market sustainability. While demand for AI computing has surged, Davis suggests the market may be building ahead of realistic demand curves, potentially creating a mismatch between supply and actual usage. This could lead to stranded assets, financial losses for investors, and potential disruptions in AI infrastructure availability.
For the AI industry, the implications are significant: companies may face higher costs for data center capacity, reduced availability of infrastructure, or consolidation that favors large hyperscalers with their own facilities. This could create barriers for smaller AI startups and reshape competitive dynamics in the sector, potentially slowing innovation and limiting which companies can effectively compete in compute-intensive AI development.