On July 1, 2025, we hosted a live webinar titled “AI Valuations and Funding Trends in 2025”. The webinar was presented by Marcin Majewski, Managing Director and Filip Drazdou, M&A Director.
You can now watch the full webinar replay below. If you would like to download the presentation material used during the session, you can easily do so by clicking the download report button on the left (if you are using a computer) or by scrolling at the very end (if you’re using a phone).
Below is the full transcript of the discussion, edited for clarity, flow, and briefness.
Welcome and Introduction
Marcin Majewski:
Hello! Welcome to our AI Valuations and M&A webinar.
I’m Marcin Majewski, founder and managing partner of Aventis Advisors, and I’m here with Filip.
Filip, say hi and introduce yourself.
Filip Drazdou:
Filip Drazdou, I’m the Director at Aventis Advisors. Nice to see you all.
Marcin Majewski:
We’re very impressed that you decided to take some of your precious summer holiday time to listen to us. So I hope we’ll deliver some interesting insights into artificial intelligence.
Like most people, we were new to the AI space, but we started quite early, experimenting with AI in our own business even before ChatGPT was announced.
So we’ve been observing AI and its impact on M&A and investments for quite a while. In this webinar, we want to share what we’ve learned so far about the public markets, the private markets, and what we expect for the future of M&A.
There haven’t been many exits so far, but we expect plenty to come. We’re keen to see these exits come to fruition.
Anyway, without further ado, let’s jump into the presentation. We have 45 minutes today so let’s start with the agenda.
We’ll talk about:
- AI in the public markets – listed companies.
- Then, an interesting perspective on what actually qualifies as AI (Filip will cover that).
- Recent AI funding trends.
- What we’re seeing in AI M&A – exits, and how to position yourself as a founder or investor.
- Then we’ll summarize and wrap up.
We’ll also stay a couple of minutes at the end to answer your questions, so feel free to type them in and we’ll try to make this as interactive as possible.
The video will also be available on our website and on YouTube – so do subscribe to our channel if you can.
Filip, over to you.
AI Valuations in the Public Markets
Filip Drazdou:
Yes, I want to start with public market data, which usually gives the best and most comprehensive daily information on what’s going on.
We’ve had very successful indices before. For example, in SaaS or IT services – where you can track sentiment and industry developments.
So for some time, we were thinking: let’s do something like this in AI. But the question was – what is AI on the public markets? Obviously, Nvidia comes to mind, and maybe a few other smaller players. But which companies actually count as AI companies?
We came up with a pretty interesting methodology. We picked one specific date in the past (January 27), when DeepSeek was released and markets crashed. Nvidia dropped by 17%, and many other companies saw double-digit declines.
So we said: okay, the companies that experienced significant movements that day are likely somehow exposed to the AI ecosystem. We put together a list of those across different industries, and that became our AI Index.
We analyzed index values starting from November 2023 (around the launch of ChatGPT, or the birth of modern AI). Since then:
- The equal-weighted AI index is up 166%.
- The market-cap weighted version where Nvidia makes up 30% is up almost 150%.
The main takeaway: we are at an all-time high for AI companies. Sentiment couldn’t be better.
Before jumping into bull and bear cases, let’s talk about what’s inside this index.
We included the entire AI supply chain. On January 27, when markets dropped, it wasn’t just semiconductors but it was everything from electricity generation to oil and gas, semiconductor supply chains, data centers, and hardware providers. So our upstream AI index includes all these companies that contribute to AI infrastructure.
LLMs and AI apps will be part of this index eventually, but most of them aren’t publicly listed yet. So for now, we’re working with the infrastructure and supply side and it’s very interesting. AI today isn’t just OpenAI or Anthropic; it’s also the entire industrial base behind them.
Bull Case: The AI Investment Boom
Filip Drazdou:
The question now is what’s next? Where do we go from here?
It’s our tradition to make both a bull and bear case. Marcin, want to start with the bull case? How far can we go?
Marcin Majewski:
Yes. I don’t know exactly how far, but I’m pleased to be bullish because I genuinely think we’re going to go very far.
Is it a bubble? Hard to say. There are a lot of expectations and there will be even more.
But I think it’s justified. AI is going to change society and the way people live substantially.
Rebuilding the world to accommodate AI just takes a lot of money. And someone has to do it. And someone will profit from it.
So I think we’re just at the beginning of the AI revolution. We’re going to see much more growth.
There’s a huge amount of infrastructure still to be built to accommodate AI. I believe we’ll have a 5 to 10-year investment boom, which will drive valuations to unseen heights.
What’s also unique this time is that it’s not just companies competing. Nation-states are also in the game. This is something we didn’t see in previous tech booms.
Much of the infrastructure will be redundant because it will be duplicated across geographies. Everyone wants to have sovereign AI.
So we’ll have sovereign AI for the U.S., for China, for EU countries, you name it.
And therefore, I think that investment in the grid, in data centers, it’s going to be immense.
As long as we haven’t built new fusion as a source of energy, I think we’ll still need a ton of such investment. Because it’s just expensive.
So I think there’s a long way to go for these companies.
Filip?
Bear Case: AI Market Corrections & Competitive Pressures
Filip Drazdou:
Yeah, it was a bit difficult to make a bear case here. A lot of these companies aren’t hype stocks or SPACs. They’re old-economy companies that have been around for 100+ years and are only now benefiting from all the investment and CapEx in AI.
If I were to make a bear case, I’d start with correlation to the S&P 500. If you think the market overall is overvalued, especially in the U.S., then in any correction, these companies are going to suffer.
We’ve already seen that with tariffs, which caused a pretty big decline for many companies. So if there’s another bear market in the U.S., these companies might drop faster than others.
My second point would be: right now, these companies enjoy very high profitability and margins. Nvidia is practically printing money in billions. But by the laws of economics, this will eventually be competed away. The total profit pool will decline as new players emerge, competition intensifies, and value is passed on to consumers or businesses.
So we might see an earnings decline and with that, stock prices could drop but not because demand disappears, but because profits are redistributed.
That said, the sample of companies is quite diverse. I personally don’t see a bubble here in AI.
What do you think, Marcin?
AI Valuations: Is There a Bubble in AI Valuations?
Marcin Majewski:
No, no. When we look at the broad segment and the supply chain, I think it’s reasonable. Even Nvidia looks reasonable to me.
Filip Drazdou:
Yeah, especially with the revenue growth and profitability it’s delivered so far.
Marcin Majewski:
Yes, exactly.
Filip Drazdou:
But it’s interesting to see that it’s not just Nvidia. We looked across different segments and analyzed which groups performed best since ChatGPT was released.
We built an equal-weighted index of the various AI-exposed companies and plotted them by total return. Nvidia – part of the chip providers group, sits around the overall index average because some other semiconductor companies haven’t done so well.
What really stood out to me was infrastructure and everything related to electricity, HVAC, data center construction. These “picks and shovels” companies that actually build the physical infrastructure for AI are the ones performing best.
That’s very telling. It’s not software for now because software like OpenAI isn’t listed yet, so we don’t know how it would perform in public markets.
So far, traditional companies have been doing very well. And that has big implications for investors. You don’t have to chase the next OpenAI or Anthropic.
There are plenty of solid opportunities in the public markets and probably even more in private markets in the suppliers to these AI infrastructure players. These are often fairly valued, but stand to benefit from the growing demand for services, construction, and equipment.
Advice for AI Founders
Marcin Majewski:
So Filip, if you were to advise a founder then would you tell them to chase unicorns, or go into HVAC and trades?
Filip Drazdou:
I think it depends on the risk appetite. But yes, I do believe there are opportunities in HVAC and electricity. Some of these companies are trading at mid single-digit EBITDA multiples.
Marcin Majewski:
No, but I mean if you were talking to a startup founder, would you tell them to go into HVAC, work in the field, or go into machine learning, building LLMs or AI wrappers?
Filip Drazdou:
I think it really depends on personality. We’re seeing a lot of people going into LLMs and AI. But it’s also difficult, because you’re not just doing research, you also have to run a business.
So, to each their own. I think HVAC isn’t as bad an opportunity as it might seem.
Marcin Majewski:
That’s a very balanced view. I’d say the risk-reward trade-off is probably better in more traditional trades than in AI.
Of course, the biggest outliers, the people who make the most money, will be in AI. But for a more reasonable return for the risk you’re taking, it might actually make more sense to go into something traditional and conservative.
There are many safe bets to make in those sectors, and they’ll also offer solid exit opportunities.
Filip Drazdou:
Yes, and if that manufacturing combines with some reshoring in the United States, there’s a lot of money to be made in producing things that were previously made abroad but are now brought back. For example, components needed for data centers.
Marcin Majewski:
Exactly. And it’s not just in the US. The same thing is happening in Europe. With deglobalization, supply chains are getting shorter, so local labor and manufacturing will likely grow across all major economic blocs.
That’s a super interesting shift.
AI Valuations in Private Markets & AI Funding Trends
Marcin Majewski:
So let’s move on to private funding.
Filip Drazdou:
Yes, funding.
Marcin Majewski:
We’re super excited about this because we like seeing things grow though there’s a lot of nuance here, and a lot to unpack.
Obviously, the headlines are familiar. The biggest funding rounds ever are happening in AI. That’s common knowledge. But if we dig deeper into the structure of funding, we start seeing patterns that are less obvious.
What we’ve observed is that more and more of the total AI funding is going to the winners. I think we can already call them that. Less and less capital is going to early-stage startups.
This shows that the industry is maturing. In some ways, it’s quite late to the game already.
2025 will probably be off the charts($110 billion as of June). I wouldn’t be surprised if we hit $200 billion in total AI funding by year-end.
And that doesn’t even include all the investments from the “Magnificent 7” into AI. So overall, AI investments might already be in the trillions of dollars.
It’s an exciting time, and there’s a lot of money to be made in this space.
Fewer Rounds, Larger Checks
In the next slide, we showed deal count. The trend is similar with fewer rounds, but more concentrated funding.
The number of funding rounds peaked in 2021, probably influenced by COVID-related shifts as well.
In 2025, we might see a recovery in the number of AI rounds. We may match the 2024 record, but I don’t think we’ll exceed the volume we saw in earlier years.
2021 likely marked the peak in deal volume, and I expect the concentration around the largest startups to continue.
AI Valuations in 2025: A Wide Spectrum
Valuations are super interesting. We presented a snapshot of 2025 valuations, and we refresh this periodically.
There’s a huge spread in EV/revenue multiples – from high single digits for some companies, to more than 100x for Perplexity.
The key driver of valuation is revenue growth. In the next slide, we compare OpenAI’s funding rounds and how its revenue multiple evolved.
In December 2023, they raised at 53x revenue. Then they grew more than 5x, and the multiple dropped to 42x. Then they nearly tripled again, and the multiple dropped further.
This is natural. As an industry matures, revenue multiples compress, and companies are expected to eventually turn a profit. We’re still some way off from that, but I expect AI companies to push harder on monetization.
We’ll likely see a broader contraction in revenue multiples, even among early-stage companies, especially as it becomes easier to make money and customers grow more willing to pay for AI.
AI used to be a novelty, often free. Now, many companies are monetizing aggressively.
What Drives Valuation Beyond Growth?
In the next slide, we highlight what differentiates valuations beyond just revenue growth.
It’s very important to evaluate the quality of the underlying business. Under the AI banner, many different types of companies operate but some business models are clearly more sustainable than others.
We divide them into three broad categories:
- Foundational Models: These are companies that focus on AI research and training. They power others’ applications. They tend to command the highest valuations, and in our view, that’s justified as they are more sustainable.
- AI Applications: These are robust businesses built on top of AI, solving specific problems in verticals or industries. They’re not just wrappers. These companies understand their clients deeply, their workflows, and apply AI in meaningful ways to deliver real value.
- Thin AI Wrappers: These typically offer minimal added value, just a slightly better interface than ChatGPT to do essentially the same thing. They’re heavily reliant on compute and are the riskiest model.
Currently, AI application companies are trading at 10–60x revenue. But they’re also quite sales-heavy and face disruption risks, especially if companies like OpenAI offer similar functionality for free.
In general, the closer a company’s product fits a real business case or problem, the better its valuation.
At the bottom, we have generic, me-too AI models. These tend to trade at 1–5x revenue multiples, which is what we’re seeing in the market today.
Marcin Majewski:
Filip, anything you’d add?
Capital Allocation in AI: Where the Money Goes
Filip Drazdou:
Sure – I’d like to add a bit on all slides. One interesting observation from the top fundraising slide is how it ties back to public markets. Almost all of the funding to OpenAI, Anthropic, and Cohere ultimately flows into infrastructure or model training.
There isn’t that much capital going into AI applications. Most of it goes to fundamental research, data centers, processors, heavy CapEx investment.
We know a lot about SaaS, and many are familiar with SaaS valuations. In comparison, seeing 20 to 30 times revenue for frontier AI companies might seem high, but it’s not outrageous. At the peak of the SaaS bubble, many public companies traded at similar multiples and with much slower growth and no profitability.
We don’t know the exact profitability of these AI companies, but at a high level, their multiples don’t seem overstretched. I’d say valuations are fairly reasonable. Investors seem to better understand how to analyze and price these businesses now. You’re not just handed a premium because you’re “an AI company.”
M&A Activity in AI
Marcin Majewski:
Let’s move on to AI M&A. We’re still early. The volume is small. I’d say minuscule but the deals that do happen are very high profile.
We analyzed the last two and a half years of activity. It’s still slow. In 2023, we recorded around 100 exits we classified as AI M&A, primarily majority buyouts. To be clear, we’re not counting funding rounds here.
In 2024, we had 138. In 2025, we’re already at 80, and it looks like this year will set a new record. The activity is picking up, but we’re still in the early innings. The real peak is probably a few years out.
What’s especially notable is how short the cycle is – from founding to exit. Some AI startups are being sold before they even go to market. We’re seeing more of that now. It’s becoming more competitive and more difficult to grow, so the exit window is getting shorter. Currently, it’s around five years, but we’ve already seen companies exit after six months.
That’s something we’ve never really seen before. It’s amazing, but also incredibly challenging for founders. You really have to time your entry and exit with precision. This game is only for the best operators.
Marcin Majewski:
Filip, anything to add?
Filip Drazdou:
No, I agree. AI M&A will continue to grow, but it’s still hard to know which companies are genuinely AI and which are just rebranded SaaS with some AI features added.
But as you said, more and more of these companies will be acquired quickly.
How to Value an AI Company
Marcin Majewski:
We’re often asked by our clients how to value their AI company. So we wanted to share some real-life experiences and practical guidance, especially in an M&A context.
First, the hype and press coverage around AI distort the reality. Yes, there are a few very high-profile acquisitions, where large, well-funded buyers pay huge prices for companies with key synergies or talent.
And those prices are usually justified. For example, companies like Meta face massive downside risk if they don’t win in AI. To mitigate that, they’re willing to pay top dollar.
But when you look at the broader group of AI startups, many of them are struggling. They’re still searching for product-market fit, they have high churn, and their cost structures are tough because many resell compute from others, which eats into their margins.
Unlike SaaS businesses that typically had 80–90% gross margins, we’re seeing AI companies operate at 40–50%. It’s tough to build something sustainable with those economics.
What makes it even harder is that the next cohort of AI startups will have even more efficient tools, often built by today’s generation. That will accelerate competition, drive margins down further, and raise customer acquisition costs.
Where Value Accrues
Most of the value today goes to the infrastructure layer, whether it’s the energy providers or the compute layer (e.g., model developers and LLMs). If I had to choose, I’d bet on the “picks and shovels” rather than the overpopulated application layer.
At Aventis, we try to bring a conservative, rational approach to AI valuation. In our view, AI companies are not fundamentally different from other businesses. The same principles apply.
You need:
- Solid fundamental analysis
- Clear unit economics
- KPIs like revenue growth, customer retention, CAC to LTV ratio, etc.
Being “AI” doesn’t get you a premium by default. If anything, in the absence of a clear buyer-driven synergy, exits can even be a red flag. A proactive exit search could signal that something isn’t working.
However, if a buyer approaches you now, and the fit is right, it’s probably a great time to respond.
The Valuation Playbook
On the next slide, we laid out our usual playbook for valuing a tech company.

- Income-based approach: Discounted cash flow. You forecast future business performance and value the projected cash flows. Very difficult in AI, because most companies aren’t profitable yet but it’s still doable, albeit speculative.
- Market-based approach: You compare valuation multiples from funding rounds or public comps. Not many M&A benchmarks yet, but in general, if you raised at 10x revenue, you should aim to double revenue to exit at 5x and make investors whole.
- Asset-based approach: Often based on replacement value. Founders sometimes like to steer conversations toward this, but we don’t recommend it in AI. The technology depreciates extremely quickly. The tools and techniques evolve fast, making what you build obsolete in a matter of months.
So, in summary, we aim to stay grounded. We want to make sense of what’s going on without giving in to hype or panic.
Closing Remarks and Q&A
Marcin Majewski:
If Filip has nothing more to add, I think we can wrap up.
Filip Drazdou:
Yes, I think we can.
Marcin Majewski:
So that’s it. And yes, we did generate a meme with AI! That’s probably the only thing in the deck we used AI for.
There’s a lot of hype and FOMO in this space. We hope this brought some common sense and clarity. Please feel free to reach out to us with any questions or if you want advice on your investments or exit strategy.
We’re happy to offer balanced, rational guidance.
Eventually, exits will happen, you just need to approach this space deliberately, with a clear head.
Audience Q&A
Marcin Majewski:
We have a bunch of questions. Shall we address them?
Filip Drazdou:
Yes, the first one I see could take all our time, but it’s probably the one on everyone’s mind:
“Is AI another dot-com bubble?”
Marcin Majewski:
There are a lot of similarities. But I think it’s different in one key respect: AI companies are generating real revenue or at least trying to. That wasn’t the case during the dot-com bubble, where you could raise money just based on eyeballs and web traffic.
Today, you at least have to sell a bunch of tokens to justify a valuation.
Marcin Majewski:
Next question: Is the hype slightly dying out?…
Filip Drazdou:
Yes, I think the hype is slightly dying out. Investors are now much better informed about AI. As you said earlier, they’re valuing companies based on fundamentals and financials not just the technology or the label of being “AI.”
Today, companies have to answer tougher questions:
- How many users do you have?
- How will you make money?
- What’s your gross margin?
It’s harder to get funded now. And many early investors got burned by startups that had an initial wave of traction but then ChatGPT added the same functionality, or it became part of Copilot, or integrated into Google Gemini. Google is rolling out new products at every conference.
You really need to make sure the AI application you’re funding at seed or Series A stage can survive. There was a big wave of AI note-takers, for example and now that feature exists natively in Google Meet. Excel got AI-powered formula support. Google Sheets now has “=AI” prompting Gemini.
So the implication is clear: there will be big winners but likely among large companies with deep pockets. It’s going to be more difficult for smaller players.
Marcin Majewski:
Exactly. Investors are more discerning now. You don’t get their attention just by saying you’re doing AI. You need a solid team and a clear go-to-market strategy to attract capital.
Q: Will foundational models be commoditized?
Marcin Majewski:
Great question. I don’t think we’re qualified to go deep on this from a technical perspective, but from an investment lens there’s more to foundational models than just the models themselves.
They require massive energy, hardware, and infrastructure to create and operate. I think only a handful of companies will be able to afford maintaining them. So I expect we’ll end up with an oligopoly – just a few big players making the majority of the money.
That makes it unlikely that foundational models will become a commodity. They’re too rare and too expensive.
Q: Are fundraising multiples as useful as public company comps?
Marcin Majewski:
Fundraising comps often show higher multiples because the companies are earlier in their lifecycle and have higher growth rates. Public company comps are lower, reflecting maturity.
Eventually, these two will converge. As private companies like OpenAI reach late stages, their multiples will likely align more closely with mature peers like Google or Microsoft.
Filip Drazdou:
Yes, we use fundraising comps a lot to set ceilings and manage expectations. For example, if OpenAI raises at 20x revenue and grows at a certain rate, a smaller niche company won’t get that same multiple, it’ll be much lower.
So those benchmarks are useful to anchor valuations, even if they aren’t apples-to-apples.
Q: At what point do the losses at big AI companies become a red flag?
Filip Drazdou:
Excellent question. We’ve seen SaaS and marketplace businesses survive with losses for a very long time. I used to be skeptical, but Amazon, Uber, and Airbnb changed that narrative. Different models, but the same lesson.
If a company gets big enough, losses are less of a concern, especially in frontier industries like AI.
The key signal is growth. As long as you’re growing, investors look past the losses. But once growth stalls, then the profitability question becomes unavoidable.
Marcin Majewski:
Exactly. Companies like Uber eventually managed to “turn on” profitability. So it’s possible though definitely more challenging for smaller players.
Economic Environment and AI Resilience
Marcin Majewski:
My view is that as long as interest rates and the cost of capital remain relatively low, no one minds that these companies are loss-making, provided there’s reasonable growth.
However, if we experience another bout of inflation and interest rates go up as a result, essentially triggering a serious economic crisis. I believe that could lead to a significant reduction in AI-related spending.
So, while that risk exists, as long as the financial environment remains stable, I don’t see widespread pressure on AI businesses.
AI Tools Used by M&A Advisors
Marcin Majewski:
Filip, what AI tools do we use as M&A advisors?
We use quite a lot, actually. ChatGPT is deeply embedded into our workflows. We’re also using Gemini more and more – it’s starting to play a larger role.
Sometimes, we have implementations with Claude Perplexity for some use cases as well.
Filip Drazdou:
So yes, we used Perplexity for a while, but after ChatGPT added search functionality, it felt a bit redundant to use a separate product just for that.
We’ve tested many tools. A lot of them aren’t quite there yet, but people are very eager to test AI applications. So if a company reaches out with an AI tool tailored for M&A, we’re very likely to try it.
But in most cases, the tool doesn’t deliver at the level we need.
Marcin Majewski:
Exactly. It’s easy to get initial traction in AI but retaining customers is much harder.
We’ve come to the conclusion that, like many professional services firms, we need to transform into a technology firm. We plan to invest heavily in building our own AI agents and systems.
The nature of junior analyst work is changing rapidly. While what we do talking to people, advising, isn’t changing much, the way we gather and analyze data is transforming quickly.
Is the Rule of 40 Still Useful in AI?
Marcin Majewski:
We got a great question: is the Rule of 40 still useful in AI?
I think it’s still relevant. The Rule of 40, profit margin plus revenue growth is a good framework. While many AI companies aren’t profitable yet, investors still care about both sides of that equation. If you’re both profitable and growing fast, that’s a strong recipe for a successful exit.
Filip Drazdou:
Yes, but it depends on the business model. The Rule of 40 originated in SaaS, where companies had high customer acquisition costs and recurring subscription revenue. In that context, it made sense to trade off profitability for growth.
If an AI company has similar dynamics, it works. But if it’s more like e-commerce or model reselling where margins are low and customer retention is weak, you need profitability upfront.
AI Valuation Insights and SaaS Comparisons
Filip Drazdou:
One more we discussed before the webinar was the “DeepSeek effect” from January 27. We looked at which companies’ stocks rose or fell that day.
Surprisingly, SaaS companies didn’t move much. So even though they might mention “AI” dozens of times in earnings calls, public market investors aren’t buying it. There’s a clear distinction: AI companies and SaaS companies are being treated as separate classes.
Disruption by Google: A Real Threat
Marcin Majewski:
A really interesting question came in on voice and video apps being disrupted by Google.
That’s definitely happening. Google’s AI video generation, transcription features in Meet, and integrations into Sheets and other tools – these are serious threats to companies with narrow use cases.
In the past, Google wouldn’t bother addressing smaller, niche SaaS products. But now, with AI, it’s economically feasible for them to deliver those capabilities directly.
This time, Google is moving fast and I think they’ve learned from past failures in social media. They were among the first to invest heavily in AI, and now they’re executing extremely well.
Filip Drazdou:
We had a line somewhere in the deck: “The bigger the TAM, the riskier it is.” That applies well here.
ARR in the AI World: Still Valid?
Marcin Majewski:
Another great question: is ARR outdated in the AI world?
Filip Drazdou:
It depends on the business model. If it’s a subscription-based business, ARR is still very useful. You can apply all the familiar SaaS metrics (ARPU, LTV/CAC, retention).
But if you’re reselling tokens or doing one-off transactions, then ARR becomes less meaningful.
Marcin Majewski:
Ultimately, whatever metric you use, it needs to add up in a DCF. ARR is a shortcut to understanding value but EBITDA, cash flow, and sustainable economics matter more in the long term.
It’s one tool in the toolkit, but never the only one.
Closing Remarks
Marcin Majewski:
That brings us to the end of today’s webinar.
If you’d like to speak with us, whether to build your AI strategy, exit plan, or investment roadmap – feel free to reach out. We’re happy to take your call and advise you.
We hope you found this session insightful. Please share it with colleagues, friends, or family who might benefit.
We’ll distribute the recording and transcript shortly.
Thanks again for joining us during your summer holiday, especially those of you in Europe, where July is typically a slower month.
Filip Drazdou:
Thanks everyone. Have a great rest of your day.
Marcin Majewski:
Thank you.
Filip Drazdou:
Thanks, bye.
Marcin Majewski:
Bye.