Entrepreneur Andrew Wilkinson, co-founder of Tiny and owner of brands including Dribbble and Letterboxd, recently shared his perspective on X, addressing AI companies’ infrastructure spending and the economics of AI labor.
Wilkinson’s comments come as AI infrastructure investment continues to accelerate across the industry, with Anthropic securing major backing from Microsoft and Nvidia to expand its compute capacity on Azure.
Wilkinson said that concerns over AI companies’ data center spending overlook the potential of AI software, writing, “If you think AI companies are overspending on data centers, you don’t understand math or business.”
He highlighted the declining cost of AI labor, noting that digital employees are becoming far cheaper than human talent. “Employees (at least digital ones) are about to be 1/10th the cost of the worlds most talented humans,” he said, adding that as businesses adopt these technologies, “there will be near endless demand for this technology in the coming years.”
Wilkinson also referenced milestones in AI development, stating, “As of Sonnet and Opus 4.5, we have crossed the chasm,” and emphasized the accessibility of advanced AI talent: “You can now purchase a genius level programmer’s time, 24/7, for hundreds of dollars a day. That will soon drop to a few dollars per day.”
Commenting on the economic impact, Wilkinson said, “The amount of economic value and innovation that is about to be unlocked in software alone, is unprecedented. Anyone who has their hands on the tools knows this. Anyone who denies it, either doesn’t understand how good it is, or is in denial.” He concluded with a quote from Upton Sinclair, writing, “Never expect a man to understand something that his paycheck depends on him not understanding.”
Industry analysts have highlighted that global AI infrastructure spending could reach into the trillions over the coming decade, as cloud providers and startups expand data centers and specialized compute capacity, according to analysis cited by Trade Brains. While these investments support model training and innovation, economists caution that they also raise questions around long-term sustainability and return on investment.
Additional research from Gartner suggests that global AI infrastructure investment continues to expand rapidly, with Gartner saying worldwide AI spending will total $1.5 trillion in 2025, driven by growth in data centers, AI‑optimized hardware, and software.
Financial observers have drawn parallels to past technology booms, noting that some AI company valuations may not match near-term revenue expectations. However, proponents argue that AI’s transformative potential justifies continued investment, especially as research from PIMCO suggests that investors are pricing in future economic value creation even amid current cost and financing challenges.
Wilkinson’s comments underscore that declining AI labor costs could make advanced AI more accessible to smaller firms and startups, even as infrastructure spending remains high. Experts from McKinsey have found that generative AI tools have boosted productivity in certain development tasks by up to 40% in observed cases, such as product management and coding activities, highlighting the potential for similar efficiency gains in software work.
