Wall Street Crash: Who Lost Trust and What Awaits the AI and IT Sector in Russia

Those who read my previous post on impressions from the Davos Forum 2026 may recall: software companies were beginning to face difficulties in communicating with investors regarding the future potential of artificial intelligence (AI). And then what we saw in late January to early February 2026 was a cascade of selloffs on Wall Street and in Europe of software company shares.

In Davos, fund managers with over a trillion dollars in assets under management spoke unequivocally about the prospects for IT companies. What happened on Wall Street came as no surprise to me. Some traders' leveraged trading alone could not have caused such strong selloffs, and the liquidation of such positions was a consequence, not a cause. The trigger was fundamental — a shift in sentiment among major investors. Retail investors were caught off guard.

I remember the turmoil in the markets at the start of the COVID-19 pandemic and how we had to quickly adapt our strategy at Otkritie Broker, but the volume of daily position liquidations in early February 2026 surpassed March 2020. In the end, hedge funds in the US dumped software stocks so quickly that their share in portfolios fell to just over 3% — an absolute low.

In Europe, for example, by the end of the first week of February, SAP shares had fallen 16.23% from the start of the year, and over the past 12 months their decline reached -27.57%. Looking at major software companies worldwide, the entire sector was down from its annual highs by an average of nearly 40%.

Many thought this was a death sentence for "software without AI" and cited the emergence of another AI model from Anthropic as the trigger for what happened on Wall Street in late January and early February. Yes, the model appeared, but let me explain how I see the situation.

What Is Really Happening...

The usual conclusion many draw is: "pure software companies have lost faith, but AI companies are doing great." But that's only part of the story. The current decline is also a disappointment among investors in the business model of the world's major neural network providers. The scale of capital expenditures they have announced — so that clients can continue to replace third-party IT solutions with software written using their neural networks — amounts to burning hundreds of billions of dollars. But to what end?

Here's what's happening. There was a software market, a market for IT solutions. Businesses bought it, there was competition between providers — more in some IT market segments, less in others. But there was competitive choice. Then neural networks arrived. Yes, with their solutions, vibe coding developed, where companies can now create IT products for themselves rather than buying them externally.

But! There's a problem on both sides. The user of public neural networks, vibe coding for their company, becomes increasingly dependent on these AIs, with their hallucinations. They don't know who has access to, where the information entered during vibe coding and testing the software created with vibe coding is stored, or how securely it is protected. These are huge risks.

And another problem: the early February crash on Wall Street and in Europe is not only about the bleak prospects of IT companies outside AI, but also about the equally uncertain prospects of the neural network providers themselves. Big Tech alone in the US plans to raise AI capital expenditures this year to $700 billion, increasing the pace more than 1.5 times compared to 2025. If we include the plans of other players worldwide, the AI sector wants to attract over $1 trillion — because that's exactly how much is needed to invest.

The development model built by the latest players is not just unprofitable — even though the cost of training AI is decreasing a thousandfold. It requires too much money from investors who cannot abandon all other investment areas. Otherwise, it begins to resemble the "tulip mania" that gripped the Netherlands in the 17th century, when literally everyone rushed to invest in just one thing — tulip bulbs. Similarly, global investors cannot invest everything only in chips, their production, data centers, and power plants. Incidentally, the crash of the tulip "bubble" also occurred symbolically in February — but of 1637.

The subtlety of the current situation is that most providers of well-known neural networks worldwide are private companies. Their securities are not listed on exchanges, so one cannot formally see investor disappointment in them. But it can still be discerned. One of the clearest indirect signs is the strong correction in Microsoft shares. By the end of the first week of February, they were down 17% from the start of the year. This is even despite Microsoft's reputation for innovation and interest in AI. The company has a partnership with OpenAI, the developer of ChatGPT. But the decline in Microsoft shares reflects, among other things, investor concerns about OpenAI's business model.

There are deeper signs as well. This is what is happening in the private credit sector for neural networks, which until recently kept them afloat financially. However, investors are losing faith. Interest rates for neural network providers on such loans have doubled over the year, and the volume of lending has shrunk by more than a third.

And What About the IT Market in Russia?

The most important development point is not software by itself, not software with public neural networks (vibe coding), and not building public AIs. The future lies in corporate neural networks, where vibe coding will operate within a company's closed perimeter. With all key processes tracked via blockchain. And better yet — in an ecosystem of partner corporate neural networks (more on this below).

In other words, the future of the IT market and AI solutions is to develop within companies themselves. Why is that? Much could be said on this topic, but briefly: "ordinary" companies outside the IT and AI sectors already have a business model that is understandable to investors. If a company is not sinking, it's the result of a well-crafted corporate strategy. Banks and brokers, by their very nature, are closer to investors, and corporate strategists are in the most effective communication with them. Incidentally, there is also an understanding there that replacing human communication with investor-clients and shareholders with AI bots is a major risk of losing trust, losing connection in financial relationships where, alongside rational arguments, there is a significant emotional component to client communications.

In fact, I see it this way: the digital era requires more personal meetings. The more "digital" an organization has, the more valuable human contact becomes for building trust. And of course, a person has the right to communicate with a person — and this demand has always been particularly felt when it comes to money and capital. Of course, in some instances AI assistants and AI bots may be appropriate, but not in those critical from a financial standpoint, and certainly not where communication is with Ultra-High-Net-Worth Individuals (UHNWI) — investors with large checks.

These "ordinary" companies, especially banks and brokers, have a professional approach to finance. Why did so many global neural network providers run for loans not to banks, but to companies in the private credit and private equity sectors? Because they cannot find common ground with banks. And why is that? Well, that's how it turns out.

This is generally a problem for many startups in fintech, blockchain, and cryptocurrencies: founders there often underestimate the crucial role that bringing in a corporate strategist with experience in the most classic financial sector — banks and brokers — plays for their ideas and ventures. But building long-term funding relationships through private credit and private equity also requires professionals. Underestimating this — and we've seen how even these funding sources have effectively become skeptical of even the most public and widely known neural networks.

From Corporate AI+IT Pairings to New Ecosystems

Sooner or later, the new trend will be recognized: now is the time to create IT solutions and launch neural networks within organizations, and over time, preferably within new ecosystems uniting several participants. In Russia, as an example, it would be optimal to build such partnerships: a bank + a brokerage company + a major exporter + a major importer + an energy company + a semiconductor manufacturing plant + data centers — all for the development of a unified IT+AI ecosystem. Within such an ecosystem, the first two participants are key from a financial standpoint. They can provide capital inflow. Importantly, everyone in this ecosystem is within Russia, ensuring technological sovereignty.