The current AI market is split into two distinct financial camps: the companies building the physical infrastructure and the companies developing the actual AI models. Infrastructure giants like NVIDIA, Microsoft, and Google are reporting record-breaking revenues because everyone is buying their chips and cloud services. However, these companies are also spending staggering amounts of money—collectively hundreds of billions of dollars—to build the supercomputers required to stay ahead. The big question for investors right now is whether this massive “capital expenditure” will eventually pay off in long-term profits.
On the software side, model developers like OpenAI and Anthropic have reached massive valuations and multi-billion dollar revenues, but they are still burning through cash at an incredible rate. For example, while OpenAI’s valuation is nearing a trillion dollars, it is projected to lose over $14 billion this year alone because the cost of running and training these models is so high. Meanwhile, Anthropic is trying to reach profitability faster by focusing on business partnerships. This “high-growth, high-cost” cycle is forcing even established companies like Meta to restructure their teams to prioritize AI spending.
Finally, older tech leaders are facing a period of intense transition as they try to figure out how to sell AI to everyday consumers. Apple is currently at a major turning point, dealing with a slight decline in hardware sales while also managing the exit of long-time CEO Tim Cook. Meanwhile, Amazon is leaning heavily into its cloud division, AWS, to act as the primary platform where other companies can access various AI models. Essentially, the financial state of AI in 2026 is a high-stakes balancing act between rapid innovation and the massive operational costs of keeping the lights on.
__________
Pascal BORNET
Everyone wants the AI penthouse.
Almost nobody wants to pay for the basement.
What I keep seeing is the same pattern: companies want AI outcomes without investing in AI foundations.
The exciting layer gets funded first:
→ GenAI pilots
→ strategy decks
→ dashboards
→ executive demos
The foundational layer gets ignored:
→ definitions
→ data quality
→ metadata
→ lineage
→ ownership
And then people act surprised when things start to crack.
AI rarely fails because the vision was too ambitious.
It fails because the foundation was too weak.
That is the expensive mistake.
Foundations are not the boring part of AI.
They are the part that keeps everything else standing.
What do you think kills more AI projects: weak vision, or weak foundations nobody wanted to fund?
#AI #GenAI #DataQuality #DigitalTransformation #DataGovernance #BusinessStrategy #Innovation#FutureOfWork #Technology
