It’s impossible to ignore – 2026 brought a new mood to Wall Street. Looking back at the previous stages of the artificial intelligence rally, the market lived by hope, visionaryism, and promises of the coming technological revolution. Investors willingly issued practically limitless blank checks, generously paying for future prospects. That’s starting to change. In trading halls and analytical reports ever more often, a dry, pragmatic question arises: where is the real return from the billions invested?
The investment community is beginning to become stricter. We are transitioning from a phase of unconditional faith in the magic of artificial intelligence to a phase of demands for real cash flows.
In such transitional moments, the informational field often polarizes itself, and it is important not to succumb to cognitive distortions.
If yesterday the market bought everything that is connected with neural networks, then today the loud voices of skeptics are already heard, stating that AI is the next inflated bubble which is about to burst. Jumping from one extreme to the other is the worst possible strategy. The truth, as always, lies somewhere in the middle. No catastrophe or dramatic collapse of AI is occurring. The technology works, implements itself, and bears fruit.
What we are observing today is not a crash of illusions, but an absolutely logical, healthy process of the maturing of the market. A change of the very methodology of valuation of technological companies is occurring. Emotional capital yields its place to mathematical calculation.
The change of the valuation paradigm means that the period of comfortable growth of the sector as a whole is complete. In 2023 and 2024, a simple strategy worked. It was enough to buy an index fund of the broad market or a basket of AI startups, and the rising tide lifted all boats without distinction. Now it seems this has come to an end. The market is entering into a stage of deep revaluation and harsh stress-tests for every separate business model.
The golden era for the strategy of stock-picking begins. Investors now are inclined to conduct a separation: who is capable of converting hype into stable margin, and who simply burns shareholder capital in the hope of a miracle.
The brightest example of a company which brilliantly passes this stress-test and confirms its fundamental value appears to be Alphabet (GOOGL). The numbers of Google speak for themselves.
The market values the company approximately at $3.5 trillion, which at first glance seems an astronomical sum. But if we look at the real financial flows, the picture becomes absolutely logical. For 2025, Google generated about $132 billion of net profit. At such a profitability the forward price-earnings ratio stands at around 24x. For a technological giant, which not simply invests into AI, but also successfully integrates it into its ecosystem, raising the return from advertising and cloud services, this is a more than conservative and adequate valuation. Google is an example of how an issued earlier vote of confidence fully pays off by a most powerful cash flow. But this doesn’t occur at all companies.
The revaluation will touch not only software giants, but also hardware infrastructure. The first years of the AI boom demanded raw, universal computing power for the training of giant base models. This was a period of total domination of the pioneers of the GPU market. But by the measure of how the industry transitions from the stage of training to the stage of application (inference), the economics of computing comes to the forefront — energy efficiency and lowering the cost per request.
In this new paradigm, a colossal potential is hidden in the manufacturers of custom chips (ASICs), and companies like Advanced Micro Devices (AMD). The market may be currently underestimating the fact that the optimization of expenses will become the main theme of the next few years. Architectures, tailored to specific, narrow tasks of AI with maximal energy efficiency, will become the main beneficiaries of this wave of rationalization.
Exactly here, in the segment of companies providing a cheaper and effective infrastructure, can hide the most interesting opportunities for an investor who is ready to look deeper than the obvious trends.
Besides the microeconomy of separate companies, for investors it is vitally important to remember fundamental macroeconomic laws. The growth of the economy and the formation of GDP are determined by the balance of demand and supply. And no matter how revolutionary and beautiful the technology of artificial intelligence is, aggregate corporate and consumer demand has its own harsh, objective limits.
The economy cannot punch above its weight. The budgets of companies on IT infrastructure and software are finite. In order for the corporate sector to massively pay for expensive AI subscriptions, update server capacities, and implement new solutions, it is necessary to take these funds from somewhere — either to radically cut other items of expenses, or to demonstrate an explosive growth of its own revenue.
In the period of peak euphoria, the market suffered from a peculiar mathematical blindness, practically fully ignoring the factor of competition. Investors willingly priced scenarios into the shares, under which both “company A” and “company B” and tens of prospective startups simultaneously will capture leadership. Capitalizations grew as if every player will receive 100% of the future market. However, this market is one for everyone.
Now arrives the stage of critical thinking. The market returns to the severe reality in which the size of the general pie is limited. In 2026, the place of every company will be valued exclusively proceeding from that, what real share of this pie it is able to bite off and hold.
We are entering into a phase of harsh competitive struggle inside the AI sector itself. In this struggle, inevitably, a stratification will occur. The winners will systematically occupy their niches and break away from the pack, generating profit, while the losers will rapidly lose their unjustifiably inflated capitalization.
The current year requires a completely different level of discipline. Successful investing now boils down to careful stock picking. The time has come to believe less the loud promises from the stage and pay much more attention to the dry lines of balance sheets.
We stand on the threshold of the earnings season for the first quarter, and this will become the first serious check of the new paradigm. The main task of the investor will be the close monitoring of reports and the merciless comparison of the real financial achievements of companies with those ambitious forecasts and bold expectations. It turns out, that he who is not able to confirm his multiples with real cash will collide with a severe revaluation.
It is important to emphasize once again: In this process there is nothing catastrophic. The market is not collapsing, it is simply changing in an absolutely healthy, rational direction. And for the thoughtful investor, this new, strict reality opens excellent opportunities for the purchase of really strong businesses at adequate prices.
A final, critical note for those who rely on technical analysis and algorithmic trading. It is vitally important to recognize when the fundamental “climate” of the market is shifting. Often, a technical model or trading algorithm can work flawlessly for six months or a year while market conditions remain relatively uniform. However, when the underlying fundamental nature of the market changes — when a new market regime takes over — these technical systems can suddenly break down and generate false signals simply because they were optimized for a different environment.
Therefore, traders who build their strategies around technical methodologies must pay close attention to these moments when the fundamental winds change direction. Recognizing a macro shift early is essential for recalibrating technical setups and avoiding the costly mistake of applying old rules to a fundamentally new market.
On the date of publication, Mikhail Fedorov did not have (either directly or indirectly) positions in any of the securities mentioned in this article. All information and data in this article is solely for informational purposes. This article was originally published on Barchart.com
