Apollo's Sambur says software's AI troubles will persist, noting the 'very large unknowns'

The technology sector, particularly the software segment, is going through a period of deep turbulence. Despite the recent rebound of the IGV Software ETF index, sector valuations remain in red territory — a decline of 20 percent since the beginning of the year is a signal that investor optimism has a short memory. Meanwhile, Sameer Sambur from Apollo Global Management expresses a view that many people in the technology industry prefer to overlook: problems related to artificial intelligence in software will not disappear quickly, and may even deepen. His analysis points to deep uncertainties that the market is still insufficiently pricing in.

Observing the capital market, it is clear that investors are juggling two contradictory narratives. On one hand, every announced progress in artificial intelligence — a new model, better performance, a new application — drives stock market rallies. On the other hand, when we look at the actual financial results of technology companies, it turns out that AI promises do not translate into revenues at the pace the market is pricing in. This discrepancy between expectations and reality is the crux of the problem that Sambur is trying to articulate.

History shows us that every significant technological revolution comes with a period of market disorientation. However, this time the scale of uncertainty seems greater than in previous cycles. It is not just that no one knows exactly when AI will bring concrete, measurable financial benefits — it is that there is fundamental uncertainty about the very direction of change in the software industry.

Is the market overestimating the transformative potential of AI?

When OpenAI presented ChatGPT, the technology world fell into euphoria. Suddenly all technology companies had to have their "AI strategy", investors were seeking exposure to artificial intelligence, and stock prices were rising at a pace that had no justification in actual earnings. Sambur suggests that this euphoria has not yet been sobered up. The market is still pricing in a scenario where AI will be a panacea for every business problem, while reality is much more complicated.

Practice shows something different. Many companies that invested significant resources in integrating AI into their products are reporting disappointing results. Not because the technology doesn't work — it does. The problem is that it does not always translate into business value that they can monetize. Software companies face a challenge: how to monetize AI in a way that does not alienate existing customers while also generating new revenue? The answer to this question is far from obvious.

Add to this the fact that the software market is extremely competitive. When one player introduces an AI feature, others immediately copy it. This leads to a situation where no single manufacturer can achieve significant competitive advantage simply by having AI. Meanwhile, the costs of developing and implementing these technologies are enormous. This is not a very encouraging combination for investors who expect high profit margins.

Great unknowns that the market ignores

Sambur speaks of "very large unknowns" — very large unknowns. This formulation is key to understanding his position. It is not that AI will not be important — it will be. It is that there is a huge number of variables that research and analysis cannot predict. What will be the long-term effects on employment in the technology sector? How will regulations affect the ability to monetize AI? What will be the actual performance of these systems in practical production?

These questions are fundamental, and the answers to them will shape the sector for a decade. Meanwhile, the market tries to ignore them, focusing on short-term gains in valuations. Investors who buy software company stocks often do so based on the assumption that AI will be manna from heaven for them. But is that assumption worth a 20-percent decline since the beginning of the year? Sambur suggests it is not — and that the market will struggle with this uncertainty for a long time.

It is worth noting that the great unknowns concern not only technology but also business. For example, what will the cloud services market look like when all AI providers are offering similar capabilities? Will it be a pure price war, or will new business models emerge? Will AI startup companies be able to compete with giants such as Microsoft, Google, and Amazon, which have enormous resources? The answers to these questions will determine which companies survive and which fail.

IGV index rebound as a false signal?

The recent rebound of the IGV Software ETF index was welcomed by part of the market as a sign that the worst is behind us. However, Sambur seems to suggest that this may be only a temporary recovery within a longer-term downtrend. The history of capital markets teaches us that rebounds are normal — even during bear markets. The problem arises when investors interpret these rebounds as a trend reversal, when in reality they are just "dead cat bounces."

The key question is: what drove this rebound? If it was fundamental changes in software company earnings, then one could speak of a trend change. If, however, it was merely a transient improvement in sentiment or stock purchases by index funds, then the signal is much weaker. Given that technology companies are still not reporting spectacular revenue growth from AI, it is hard to believe that the fundamentals have changed.

Moreover, the software market has problems that go deeper than the AI issue alone. Competition from open-source solutions, a shift in customer preferences toward SaaS rather than traditional software, and margin pressure — all of this operates independently of the AI revolution. The index rebound may be merely a short-lived respite before another decline, when these fundamental problems make themselves felt again.

How are Polish technology companies coping with uncertainty?

The Polish technology industry is observing these turbulences with mixed feelings. On one hand, Polish software and IT companies have access to the same investment capital and can use the same AI technologies as their Western counterparts. On the other hand, Polish companies have fewer resources to invest in long-term research and development without guarantees of return.

Many Polish technology startups are trying to find their niche in the AI market. Instead of competing directly with giants, they are looking for specialized applications — for example, AI for healthcare, logistics, or manufacturing. This is a wise strategy, but also a risky one, because each niche can be attacked by large corporations with unlimited budgets. The uncertainty that Sambur speaks of is particularly acute for Polish companies that do not have a financial cushion to survive a longer period without profitability.

At the same time, Polish technical talent is sought after by global companies. Many Polish engineers work on AI projects for foreign companies, which is both an opportunity and a threat. An opportunity, because Polish specialists can learn from the best. A threat, because the Polish industry is losing talent that could build innovative solutions here.

Profit margins under pressure — is this the end of the era of high profitability?

Historically, the software industry has enjoyed impressive profit margins. Once a product was ready, duplicating and distributing it cost virtually nothing. This allowed companies like Microsoft or Salesforce to achieve margins at the level of 30-40 percent, or even more. However, this era is slowly coming to an end. Why? Because AI is changing the dynamics of competition.

When everyone has access to the same AI models — whether through OpenAI API or through open-source solutions — product differentiation becomes difficult. If a competitor can quickly copy your AI feature, the only way to maintain an advantage is to continuously invest in R&D. This raises operating costs, which reduces margins. Additionally, pressure from customers to pay for what they actually use (pay-as-you-go model) also reduces average revenue per customer.

Sambur suggests that this pressure on margins will persist for a long time. Software companies that still expect margins at the level of 40 percent will be disappointed. The new reality may look more like margins at the level of 20-30 percent — still solid, but far less impressive for investors who have become accustomed to previous levels. This is a change that will slowly but systematically lower sector valuations.

Pessimistic scenario versus reality

It is worth considering whether Sambur is being overly pessimistic. Perhaps AI will really change the software sector for the better, and we are only at the beginning of this transformation? History shows that such transformations do happen — the Internet, mobility, cloud computing — all went through periods of hype and disappointment before ultimately changing the world.

However, there is a fundamental difference between previous technological revolutions and AI. The Internet, mobility, and cloud computing were primarily about infrastructure — new distribution channels, new ways to access data, new ways to store information. AI is something different — it is about automating thought processes. This is much more fundamental, but also much more unpredictable. No one really knows what the job market will look like when AI is able to do things that currently require people.

Sambur's pessimism seems justified not because AI will not be important, but because its impact will be far more chaotic and unpredictable than current market valuations assume. The market likes predictability — revenue growth of X percent per year, margins at Y percent. AI introduces unpredictability on a massive scale. This is precisely what Sambur calls "very large unknowns."

What should be an investor's strategy in this environment?

If Sambur is right — and his analysis seems logical — what should be the strategy for investors? First, one should be cautious about companies that promise quick AI revenue without a clear business model. Second, it is worth looking for companies that have solid business fundamentals independent of AI — that is, companies that made money before AI and will make money after AI.

Third, one should be prepared for the market to be volatile. Rebounds like the recent IGV index gains may repeat, but they do not signal the end of the bear market. Fourth, it is worth tracking what the big technology companies are doing — Microsoft, Google, Amazon. If they are investing in something, it is probably worth looking at it more carefully, because they have access to the best market information.

For Polish investors who want exposure to the technology sector, this may be an opportunity to buy stocks at reduced prices — but only if they have a long-term investment horizon and can withstand additional volatility. Selling in panic during declines and buying during rallies is a path to losses. It is better to have a clear strategy and stick to it, regardless of short-term market movements.

Is the software sector at a turning point?

Summarizing Sambur's analysis and market context, it seems that the software sector is indeed at a turning point. This is not a turning point in the direction that everyone expects — that is, rapid growth and huge profits thanks to AI. Rather, it is a structural turning point in which old business models stop working and new ones have not yet been formed. This uncertainty will persist for the coming months, and possibly years.

The decline of the IGV index by 20 percent since the beginning of the year is a signal that the market is beginning to understand this. The recent rebound is likely just a temporary improvement in sentiment. The real change in the fundamentals of the software sector will take much longer, and its direction remains uncertain. This is precisely what Sambur is trying to say — and he is right that it is worth listening to him.