As AI without a product Lift the billions In Monster valuations, Openai‘S Altman itself myself He called this moment “AI bubble”. But the firms that managed AI wave are themselves very vulnerable to the investors with any profits.
With current valuation, there is a real risk of OPENAI is not going to refund money to later investors. His combustion of money until 2029 will arrive apparently price $ 115 billion (Despite AR AR in the amount of $ 20 billion for 2025). Earlier supporters have already closed themselves, but the current expectations of future profit that motivate investors can be misleading.
We are in the race to the bottom
The basic AI models – LLM used by most AI tools – are now in the creation phase. When GPT first fell into our lives, it was a real technological step forward, depicting transformational, profitable applications to firms.
But since then, every recent progress was getting smaller. The implementation of GPT-5 Openai was deeply disappointing for users who demanded a return to the previous version.
With the fundamental fundamental AI firms – GoogleIN XAIIN AnthropicIN FinishIN Openai – Working on relatively marginal improvements in your basic models, the price is now real sales. Switching costs between the fundamental suppliers are very low for many AI users, and Openai desires to get and stop customers Aggressively lower prices with recent models.
And yet their operating costs will remain a bit in the foreseeable future, because their huge demand for infrastructure is consistently increasing.
Real winners
In fact, the most clear winners of the current AI AI are GoogleS, Microsoftsi NvidiaS provides selections and shovels. Opeli and others were financial advantages for these industry giants, and OpenNAI recently Commitment to purchase $ 300 billion IN Oracle Calculation power over the next five years.
While the costs of infrastructure or calculation requirements are one way or the other falling – and I do not see that this happens in the near future, a quantum breakthrough or otherwise – billions of realized profits will go to the pocket of GPU suppliers, energy and other resources, and not foundation models.
Another real competition for profits takes place among firms that make up AI tools, which shall be practically useful for the life and jobs of individuals, trained on the basis of knowledge specific to firms that the company they support.
Consumers and business users logically go from generalized AI platforms to more specialized tools that are aimed at specific needs. But in this area OpenAI competes with an infinite variety of startups and technological giants. Their early advantage is the first much less necessary than having a unique team working on UX or learning to bolster, and as a substitute of getting a huge pool of historical user data.
And although GPT and other LLM for profits supply many applications of other firms, one other existential threat is free, open source models, equivalent to Deepseek offering a real alternative.
Opeli would have to interrupt into this layer of use by force to have a significant a part of profits that have more specialized AI products. But so far they have not proved that they have a clear advantage in comparison with competitors on the fundamental arenas (Codex vs. OpenAI’s Codex vs. Girub“S Copilot, Dall-E vs. Midjourney).
The way forward for the AI investment
So why do investors clog lots of of thousands and thousands in Opeli and others?
It’s partly ego. This and they hope that their “horse” ultimately wins the whole cake. I do not think one player will go to the top. The separation of assorted fundamental providers of AI becomes too irrelevant to supply investors a clear winning scenario.
Even in terms of probability that the fundamental market of artificial intelligence consolidates about a few players, they may still not return to multiples that can generate roi to later investors.
Users will spend money on AI applications that significantly affect their lives. I see many profits from AI firms that are terrified in a specific industry area of interest and transform its results – equivalent to the recent discovery of medicines, Pionowy Enterprise applications, healthcare and even consumer applications.