Kuehne + Nagel International AG (KNIN) Earnings Call Transcript & Summary

Ladies and gentlemen, welcome to the Kuehne + Nagel Management AG Q2 2026 Results Conference Call Live Webcast. I am Valentina, the Chorus Call operator. [Operator Instructions] The conference is being recorded. [Operator Instructions] The conference must not be recorded for publication or broadcast. At this time, it's my pleasure to hand over to Stefan Paul, CEO of Kuehne + Nagel. Please go ahead, sir.

Thank you very much, Valentina. Good afternoon, and welcome to the presentation of Kuehne + Nagel Second Quarter 2026 Financial Results. I'm CEO, Stefan Paul, and I'm joined today by our CFO, Markus Blanka-Graff, and once again, by our Chief AI and Innovation Officer, Alireza Nemati. Let's go into Page #2, half year 2026 results, volume growth and cost control drive recurring EBIT momentum. For the first half of the year, strong volume growth in Air Logistics and disciplined cost control across the group, drove a steady improvement in recurring EBIT. In Q2, recurring EBIT rose to CHF 381 million, an increase of 6% year-over-year and 24% sequentially, roughly double the typical seasonal gains. There are several factors behind this momentum. Our cost reduction program first announced last October and largely underway by year-end 2025 delivered savings of about CHF 50 million over the first half with a modest incremental uplift from Q1 to Q2. We expect this positive trend to accelerate in Q4 and to an annualized gross run rate of at least CHF 200 million in savings by year-end 2026. Discipline in cost control enhanced the profitability of the faster-growing business units, notably Air Logistics. It also mitigated the effects of annual wage inflation concentrated in Q2 and the slow development of Sea Logistics volume. The net effect was a group conversion rate of 16.9% in the second quarter versus 14.6% in the first and 16.2% in Q2 last year on a comparable basis, adjusting for the IMC reclassification. The combined sea and air conversion alone was 30.2% in Q2 versus 26.2% in Q1 and 28.1% last year in Q2, again, on a comparable basis. Recurring group EPS in Q2 improved by 6% year-over-year or 11% excluding currency headwinds. Headline free cash flow generation in Q2 was roughly comparable to last year's. Excluding disposal proceeds. There was a decline because of higher working capital outlays linked to growth, especially freight-based airfreight. Overall, we see continued evidence of market share expansion across multiple business units. Given the momentum in the first half and our expectations for the second, we are lifting our 2026 full year recurring EBIT guidance. As usual, Markus will cover this in more detail shortly. Alireza, we also present our AI upgrade, along with our expectations for material productivity improvements emerging in 2027. Let us now turn to our performance by business unit. Page #3, we start with sea freight. As always, cost control drives further recovery of unit profitability, left volume and GP per container unit and right EBIT per container. In sea logistics, unit profitability recovered significantly for a second consecutive quarter thanks to our cost reduction efforts and improved quarter-on-quarter volumes. Q2 volumes expanded by 8% Q-on-Q or in line with the average Q1 to Q2 uplift over the past 5 years. Year-over-year, volumes declined by 1%. The situation in the GCC is still a material drag on volumes. European and North American input rates from Asia were robust in Q2, but countered by weak demand in backhaul trades to Asia. Our share of SME volumes rose to 52% in Q2 from 50% in the first quarter. EBIT in Q2 improved sequentially by 24% to CHF 140 million, mostly due to cost management efforts. Volume growth and improved profitability at IMC also contributed to the increase. Year-over-year, EBIT declined by 6%, excluding currency effects, mainly due to lower yields. Average yields were stable Q-on-Q for the third consecutive quarter in line with the expectations we shared during our last earnings call. We anticipate a continuation of the solid trend in yield. The Sea Logistics conversion rate was 29% in Q2 versus 30% last year and 25% in Q1. Next is Air Logistics on Page #4, strong EBIT growth, driven by higher volumes and better mix. In Air Logistics, strong underlying volume growth, a beneficial shift in mix and ongoing cost control combined delivered a very robust EBIT improvement. Q2 volumes expanded by 13% Q-on-Q. This growth is well ahead of the average 9% sequential uplift over the past 5 years. Year-over-year, volume grew by 3% in Q2 despite a much tougher comparison of 9% last year. EBIT rose by 39% quarter-on-quarter to CHF 154 million in Q2, a 42% increase year-over-year, excluding currency headwinds. A very beneficial mix shift helped drive this EBIT improvement. We gained market share in higher-yielding hard cargo segments, including the technology sector, while the contribution from lower-yielding businesses like perishables and e-commerce declined. This positive mix shift is the main driver of better years, 8% better quarter-on-quarter and 10% year-over-year. The Air Logistics conversion rate was 31% in Q2 versus 26% last year and 27% in the first quarter this year. Next is Road Logistics on Page #5. Headline EBIT growth accelerates on market share gains. In Road Logistics, EBIT growth showed continued momentum in the first half, building on the signs of demand recovery, we first highlighted in Q4 last year. The Q2 EBIT totaled to CHF 36 million, a 29% improvement on last year on an organic basis, excluding currency headwinds of 7%. Net turnover grew by 16% year-over-year in the second quarter or 12% organically, both excluding currency headwinds. This marks a clear improvement on the respective 9% and 5% growth in Q1. These positive trends reinforce the view that Q4 marked an inflection point. We feel that the improved shipment volumes of recent quarters results more from a market share gain than a recovery of underlying demand. Demand for custom solutions remained firm, a consistent trend since Liberation Day in the second quarter last year. This was also the case for demand in the UAE in response to supply chain disruption from the conflict in the Middle East. And lastly, we now see signs of more robust demand emerging from the tech sector. This was Road Logistics. Now let's move to Page #6, our contract logistics business, solid underlying profits with investments in growth. In Contract Logistics, recurring EBIT was solid on an underlying basis, recurring EBIT totaled to CHF 51 million in Q2, is 14% lower sequentially and 9% lower year-over-year, excluding currency effects. This result reflects some additional cost due to new contracts going live in Q2 as well as investments in people and systems to meet growing demand as well here in the tech sector. Net turnover grew by 4% year-over-year in Q2, excluding currency effects. This is roughly in line with the typical underlying growth of 5% over the previous four quarters. We saw continued market share gains once again across geographies and sectors. The conversion rate of 6% in the second quarter roughly matches the underlying prior year results. We are confident that the stable growth trajectory will continue as Contract Logistics currently has more than 30 new contracts in the implementation phase. This concludes my comments on the performance of the business units, and I will now hand over to Alireza for an update on the progress of our AI initiatives.

Thank you, Stefan. Since we spoke to you in March, we have continued to scale AI across the organization with the goal of creating sustainable operating leverage. We currently expect AI integration to yield productivity gains of at least 5% across the addressable [indiscernible] workforce and significantly higher output for IT engineer. The material traction emerged in 2027. Winning the future in global logistics means using AI to manage complexity and increase shareholder and customer value. We have filled four advantages for exactly that purpose. Starting from the bottom of the slide. We own and control our proprietary cloud-based IT platform and have maintained our own TMS systems for decades. We have made our data AI-ready. We are redesigning workflows with AI at the core, and we are equipping our people to scale adoption of AI, these mutually reinforcing advantages create a competitive moat that will help drive our future growth, margin expansion and differentiation. Integrated data enables smarter workflows, redesign workflows free people for higher value work, engage people build deeper customer relationships, and customer relationships not only generated data that makes our AI more smarter, but also generate the revenue that will drive our profitable growth. The benefits of our AI transformation will come from four areas of improvement: operational, developer, individual and customer. These benefits will establish a doable and compelling value proposition for logistic customers. We're increasingly seeking partners who have successfully integrated AI into their own organization. I will start with the operational benefits because that is where AI translates straight into better P&L performance and stronger customer relationships. Our efforts to centralize and standardize repetitive workflows are already having a positive effect across the group, and they lay the groundwork for AI-driven automation with material traction expected in 2027. In sales, AI has begun to generate customer briefings, ultimate meeting documentation and assist the contract review of incoming RFQs. Presently, we expect a productivity increase of around 10%. In coding, we are increasing our daily capacity and customer responsiveness. AI is now processing standard spot code and automatically escalates complex cases for human review. In customer integration, we can now complete bespoke EDI and API connection in less than a day. This means they have started to onboard customers faster at a lower cost. In our global services, we are using AI agents to progressively automate workflows end-to-end. This began the ticket handling and will extend to route verification, document management and proof of delivery validation, just to mention a few. In shipment visibility, AI agents have started to verify transport milestones autonomously. This improves response time and eliminates manual follow-up. An exception management for temperature-sensitive cargo AI agents work around the clock, verifying compliance and shipment conditions across carriers and languages. This will enable us to focus on fast, scalable intervention for high-value goods. Finally, in our warehouses, machine learning dynamically matches labor capacity to expected demand. Beyond operational benefits, AI is also helping our development teams build and maintain higher quality software faster. A growing share of our code is now AI generated. We already delivered products significantly faster and expect to increase output to engineer over the course of 2027, improving our speed to meet market demand. We are transitioning towards a state where AI orchestrates the entire software development process where our engineers retain ultimate responsibility for validation, decision-making and oversight. To increase individual productivity through AI, we are putting internal agents directly in the hands of our white-collar workforce from top management to office employees. As with our other efforts, we expect to see first traction from this efficiency over the course of 2027, although the monetary impact remains difficult to assess. We have begun rolling out enablement program that helps every level of the organization understand AI, trust it and use it to augment their own capabilities. As AI literacy rises, the same workforce can take more volume and more complex work, allowing us to grow without expanding our cost base at the same rate. The benefits for our customers will also be significant. By integrating our AI directly with our customer systems we become embedded in the operations as we expand our strategic partnerships. Speed to market is paramount by the window of opportunity remains open. We have already started to evaluate the readiness of industrial and tech companies for this kind of partnerships. As we codevelop AI capabilities with key customers, the integration deepens with every transaction and creates switching costs that competitors cannot easily replicate. For example, better predictive ETAs and proactive exception management enhances customer trust and the durability of our top line growth. So let me turn to our cloud-based AI platform. which powers the transformation. Our proprietary platform gives us control over quality, cost, resilience and speed of deployment. This approach to AI integration is rare in a market where most of our peers depend on external vendors for their core TMS. Instead of every team building its own AI stack, our platform provides the shared infrastructure once, so teams solve business problems rather than rebuilding plumbing. And every new capability deploys enterprise-wide at marginal cost. Our core AI platform has three layers, again, from the bottom of the slide. The foundation layer connects all our internal and external systems. The middle layer lets us ship new AI capabilities across the whole organization and speed by ourselves. No vendor negotiations, no integration queues. Importantly, this means we are not tied to any single model provider. We deploy whichever model is best for each task and swap as the technology evolves. At the top, it's a single pane of glass, one access point that puts our AI agents into the tools of people already use every day. So let me close where I began. Our AI strategy is focused on creating sustainable operating leverage. This includes at least 5% productivity gains across our addressable white-collar workforce and significantly higher output per engineer with material traction emerging in 2027. We are embedding AI as a core operating logic, not as a peripheral tool. It is a key enabler for future growth, margin expansion and competitive differentiation. With that, I'm happy to take your questions during the Q&A, and I will now hand over to Markus.

Thank you, Alireza. Good afternoon all, and thank you for your continued interest in Kuehne + Nagel. I would like to start where Alireza left off with some additional details related to our AI initiatives. The initial traction we expect to deliver in 2027 relates to our core white-collar workforce and sea logistics, air logistics and functional areas such as sales, finance, IT and HR. We're also working on use cases, specific to road and contract logistics. But the focus areas we highlight here are expected to deliver the first visible traction in 2027. On the left side of the slide, you can see four pie charts specifying the addressable workforce and the associated cost base. This is the scope for our short and midterm opportunities. On a long-term basis, this picture may change to a larger population, and we will continue providing you updates on a half year basis. The size of this white-collar workforce is just over 25,000 FTEs and represents a cost base of approximately CHF 1.7 billion. That's roughly 35% of our total staff cost. Based upon the initiatives that are underway, we see current scope of productivity improvement of at least 5%. Alternatively, a similar increase in business volume without cost increase. We project an annualized EBIT impact of CHF 100 million to CHF 150 million to be realized by the end of 2027. This estimate is based upon run rate gross profit and staff costs in first half 2026 alongside the number of process E&E and logistics orders per FTE and gross profit per order. Let me emphasize that these values are based on current cost levels for AI services. If these costs were to change in the future, we will include this in our regular update. When it comes to technology, I'd like to emphasize the strategic advantage of TMS ownership and in-house development as it should afford us greater speed and depth of AI deployment. Our initiatives to develop a broad-based AI literacy and effective use of copilots at the individual level is a critical foundation for all of our AI efforts. All our initiatives pay into better customer experience and service quality. Lastly, we are assessing additional use cases, including potential top line initiatives. We will update on these efforts when appropriate to do so. We will continuously assess the scope and provide our next detailed update no later than March 2027, alongside the presentation of our full year results. I would now like to turn to the regular review of our financial performance for first half year and second quarter of 2026. Looking at the income statement for the second quarter, I would like to describe the key drivers behind the year-over-year earnings demand. First, we saw an inflection in year-over-year gross profit trends in Q2 with a return to growth, led by Air Logistics volumes and yields. Second, a higher proportion of these gross profit growth converted to recurring EBIT growth, supported by our cost reduction program, which provided about CHF 50 million of savings over the first half year. And lastly, underlying EBIT growth was masked once again by a 6% foreign exchange translation headwind in Q2 as last year's result, consolidated at a significantly stronger balance. Turning to working capital development. We saw the base increased to more than CHF 1.6 billion over the past quarter, an increase of 6%, reflective of strong sequential volume development as well as higher rates. However, the relative development improved with net working capital intensity declining from 6% at the close of the first quarter to 5.5% midyear, back within our guidance corridor of 4.5% to 5.5%. This corresponds with a re-expansion of the spread between DSO and DPO to nearly 3 days. The CHF 97 million increase of core working capital in the second quarter, contrasts with the CHF 86 million inflow during the same period last year. This primarily reflects the growth in air logistics supported by a large proportion of dedicated freighter services. Now have a look at how this fits into overall free cash flow generation in Q2. In Q2, we produced CHF 116 million of free cash flow including CHF 40 million of cash proceeds from asset disposals, which had no material P&L impact. This was offset somewhat by the cash outflows related to our cost reduction program. For a closer look at the cash conversion, let me move on to the next slide. Here, we see the usual comparison of free cash flow conversion in the most recent quarter versus historical average. What is new in this presentation is a focus on history since the first full year of APAC consolidation, an event which changed the pattern of our free cash flow generation with a larger proportion of freighter-based, freight activities. The second quarter is typically the second weakest cash conversion quarter after Q1 with an average conversion rate of 54% since 2022. In the current second quarter, conversion was 28% or just above 30%, excluding the effect of the cost reduction program outflows. The weaker relative conversion in Q2 is a product of high growth in air freight volumes supported by dedicated freighters. Looking forward, we expect a continuation of normal underlying free cash flow conversion trends over the coming quarters, and would highlight a long-term average of 90-plus percent conversion rate heavily weighted to the second half of the year. Turning to our financial guidance for 2026. We are raising both the lower and upper ends of our recurring EBIT guidance by CHF 100 million and CHF 150 million, respectively. This results in a new recurring EBIT guidance range of CHF 1.35 billion to CHF 1.55 billion. Our upgraded guidance for the year implies a stronger recurring EBIT result in the second half of the year versus the first half. This is in step with the historical average distribution of earnings power, given that peak demand is typically weighted to the back half of the year. We continue to see global GDP growth but with persistent uncertainty across geopolitics, macroeconomic policy and trade. Our cost reduction program is on track. With a faster than anticipated start in Q1, I reported about that, and some incremental progress in Q2 to deliver approximately CHF 50 million of savings so far. We expect a run rate in Q3, similar to that of the second quarter with acceleration into Q4. The target is unchanged with an annualized run rate of more than CHF 200 million of gross savings and an impact of more than CHF 120 million in 2026 alone. Moving to currencies. In terms of currency translation headwinds, our financial guidance assumes no more than a 5% negative impact with the bulk of this headwind in the first half. Our expectation for a 25% effective tax rate is unchanged. And one final note regarding capital structure. Please remember that our long-term preference for a small net cash position remains unchanged. That leads me to the key takeaways, namely our unchanged strategic focus is on market beating growth in targeted attractive sectors. Air Logistics is currently a key driver of profit growth, thanks to market share gains and attractive mix development. Yields in Sea and Air Logistics remained stable respectively, significantly improved. Our cost reduction program is on track with continued confidence in targeted savings. We reiterate our expectation of material AI productivity gains emerging in 2027. Today, we provided you with further details around specific initiatives and business areas that will contribute to this initial traction with the scope of annualized EBIT impact by year-end 2027 in the range of CHF 100 million to CHF 150 million. Lastly, we are raising our full year earnings guidance range for 2026 to CHF 1.35 billion to CHF 1.55 billion. With this, I want to thank you for your attention and hand back to the operator to open the Q&A session.

[Operator Instructions] The first question comes from James Hollins from BNP Paribas.

I was just coming on the contract logistics side and the EBIT phasing there. Clearly, you flagged a lot of investments and new contracts starting in Q2. Is that kind of the peak headwinds were this year? Or maybe should we be thinking about Q3, Q4 seeing something similar, I guess, as a headwind to the overall divisional EBIT in 2026? And then secondly, I see relative to the Q1 presentation, you removed the comment sea and air freight market demand growth in line with GDP at best. Am I reading too much into that? And is that, I guess, the uncertainty around that is why you've widened the range on full year guidance?

Hi James, it's Markus. Let me talk about contract logistics quickly. Q1, I think we called out also a nonrecurring effect of CHF 36 million. That has obviously moved first quarter result quite substantially. Going forward, Stefan has alluded to contract wins. Contract logistics is a business that needs couple of weeks, if not months, to start implementing these businesses. Usually, what we see is that as we speak, back end of the second quarter and going into the third quarter, we have 300,000 square meters under implementation for various customers, so you should see a bit of a start-up cost situation, which goes into the third quarter and will see the first positive impact on the P&L in the fourth quarter. So it's really a matter of growing the business in a contract logistics way.

Yes. I take -- James, I'll take the second, Stefan speaking. I'll take the second question in terms of first of all, is the GDP growth, the 1.5x GDP expectation, still our focus, it's still our commitment. Yes, it is. We have seen a little bit of a dip in sea freight, but we remain confident that this is coming back. We have expectations on positive volume development for the third quarter and as well for the fourth quarter and overall most likely, we will end up the year positive. In air freight, there is a very strong demand still coming in, and that's the reason why we are rather confident to deliver the 1.5x GDP growth in terms of volume in air freight on a constant basis, so hopefully, that answers your question on the uncertainty around 1.5 GDP growth. Yes, no, yes, it's okay, yes.

Just confirm you're talking sea volumes expected to grow in Q3 and Q4?

Yes. Yes.

The next question comes from Alex Irving from Bernstein.

Good afternoon. Two from me, please, both on AI. First of all, thanks for the helpful detail. Third question, why would you be able to hold on to any cost reduction from AI in the margin rather than using the cost advantage to take volume share? Second one, you have your advantages in deploying AI as you have before, workflow ownership, TMS ownership and clean data, what scope do you see for third-party software vendors or AI startups to help other forwarders match our capabilities. In other words, it is a lasting structural moat? Or is it nearly a head start as the rest of the market?

Alex, it's Markus. First question, the cost reduction. Spot on. You're absolutely right. I think I'm a believer in -- when the cost benefits are being created. Eventually, they will move on to the customer's benefit rightfully so over time. So you have benefits as a first mover. Hence, I think Stefan's comments were clear, our clear preference is to take on over proportionately more business with the existing well-educated expert workforce rather than not growing and -- or leaving some of the business opportunities on the table and just change cost or opportunities that might transfer to the customer. So the answer is very clear, preference of growing over-proportionally business with our experts then the second option.

Alex, this is Alireza. I'll take the second one. As I mentioned, one of our strongest mode is that we have the in-house development, as you rightly called out on the TMS data layer that we currently have. There's a couple of places that we do -- we're playing right now. One is that we're currently developing initiatives in-house. Obviously, given that this space is moving so fast, we also partner with external companies and start-ups to either cocreate together or utilize the latest technology and bake that entirely into our AI core stack. That is one of our biggest advantage so that we don't have to negotiate with third parties and can straight implement the effective of the latest AI models right into our systems. And when it comes to partnering up, it's obviously important for us to be more agnostic because, for example, it does not make sense to use a very expensive frontier to read a field in the database. Here, we're looking at the latest modes that are currently available and can either use a small language model to do that, which is cost efficient and lower cost or develop the technology in-house to exactly execute on that.

The next question comes from Marco Limite from Barclays.

I have a follow-up question on your estimate of CHF 100 million to CHF 150 million cost savings on AI. Shall we think about it as addition to EBIT, or in other words, how have you been factoring in your calculation also some pricing dilution, as most of the other freight orders, we also try to implement some cost savings. So is the CHF 100 million to CHF 150 million net EBIT increase or that's just a cost gross of some possible pricing dilution? Would be my first question. And second question is on OpEx. There have been some headlines out there where you're suggesting or headlines suggesting a potential disposal of a 20% stake or an IPO. Can you just clarify what's going on there? Because clearly, you got OpEx and you sold a minority stake, then the option was triggered and now this headline. So yes, just why you are considering a disposal, is there anything going wrong there? Or you just think that makes sense to crystallize value?

Marco, it's Markus. So first, the question on the CHF 150 million and how they arrive into the EBIT line. And I think, however, you want to look in asset I can confirm that this is a gross amount at the current stage. And I'm emphasizing the gross first, we are currently not 100% sure how the cost development of AI services will continue into 2027. But this is something we will see as we go into the year. But how it's going to get into the EBIT line can be different routes, right? One can be the straight out cost reductions, manpower reductions, FTE reductions or alike. And the other one is, again, our preference, growing the business, and we have assumed in our models that growing the business we would do at the current productivity and unit economics that we currently enjoy. So it can be through both ways, clear in our mind is that the CHF 100 million to CHF 150 million will arrive at EBIT level. So that's, I think, your view on -- you can probably look into this from a calculation of the old modeling perspective in both ways. Apex, I can only say Apex continues to be our strategic investment. It's a massive growth opportunity and growth organization within the network of Kuehne + Nagel. There I have also seen commentary out in the market, but there is nothing that we could comment or confirm or make any statements around that other than Apex is a part of Kuehne + Nagel organization and a very high valued growth machine for us.

Marco, let me add, Stefan speaking. And I think we have mentioned it already a couple of times since now almost 2 years, we are leveraging Apex as well as our carrier. So the Kuehne + Nagel legacy basically is using their expertise, their expertise in terms of the charter capacity and the operation, which is managed out of Hong Kong, and this is as well an enabler -- was an enabler, still is an enabler for our significant growth, in particular in the tech sector. So this is an integral part of our business offering.

The next question comes from Muneeba Kayani from Bank of America.

Yes. Just coming near term, I wanted to talk a little bit on what you're seeing in the ocean market right now? Do you think the strong demand we've seen was pull forward? And you talked about volumes higher -- growing year-on-year in ocean and then just want to make sure, how are you thinking about ocean yields in the third quarter at this point? And then secondly, just on contract logistics. So with this growth on the hyperscaler side, where do you see potential for the contract logistics business? And how much can it grow, not just this year, but as you think about the next couple of years?

Muneeba, thank you very much, it's Stefan. First of all, as we all know, we have seen strong demand coming into the second quarter, especially on the transpac business. A lot of demand in Asia, especially China to the U.S. and the benefit from -- and this is the second question, from a yield perspective will be seen in the second quarter, so the yield will go up. I have mentioned, I think, in the last two calls that we have started to focus pretty much on the new giants in China. And on the prepaid market, we see already decent success, and that's the reason why my statement was made a couple of minutes ago that we expect a positive volume development for us, low single digit, of course, but positive development supported by the Chinese prepaid market into Europe, in particular, the U.S. when it comes to the sea freight. So overall, rather or slightly positive in terms of the volume outlook driven by these trade lanes and the yield will be higher going into the third quarter. The hyperscaler question on the contract logistics side, we have mentioned or maybe take it a little bit different spin here, we have started 1.5, 2 years ago with air freight first with the inbound [indiscernible] to the U.S. then followed by the last mile, the installation activities, leveraging our road business unit and then on some of the power lanes as well as sea freight. And now last but not least, but a very important aspect is the nomination of some of our customers, large customers, very large customers for the contract logistics business. We have an implementation roughly 300,000 square meters in the U.S., which is brand new to us. your question was, is that to be replicated and what is ahead of us in terms of how many contracts we believe we can gain in this marketplace, and I think this is the beginning. We expect that looking at the pipeline in the U.S. and other markets for this tech hyperscaler marketplace that the stickiness of our contract logistics organization and the capability now, including vendor management has been proven, and I have no doubt that we see more contracts coming in our way.

The next question comes from Alexia Dogani from JPMorgan.

Just firstly, Markus, very helpfully, you mentioned about the cost development of these AI tools, we're hearing kind of increased cost inflation for cloud, for tokens. Can you give us a sense of what the basically cost base that is relevant for this kind of inflation trends? And how do you see kind of the evolution forward, that's one. And then secondly, if possible, can you give us a little bit of a sense of the CHF 100 million to CHF 150 million annualized gross savings by the end of '27 how they phase in? And do this give you confidence that basically, the organization has enough levers to offset any potential kind of pressure on the sea logistic vanilla brokerage yields?

I can take the first question on the token side. This is Alireza. We can all see an ongoing public debate around the future cost of compute, especially around AI deployment. But given the rapid pace of technology development, we emphasize the flexibility built in our AI stack and that our approach is model agnostic. This affords us to have the freedom to deploy white range tools with varying capabilities cost that are best suited to the task. And I double-click on the example I mentioned before, if you consider that you want to utilize AI for small task is not necessary to use the most expensive frontier models and said you would focus on a so-called smart language model that could give you the sound results at a lower cost base.

And I think the -- Alexia, I'll take the second question on the phasing. Practically what we're looking at is where processes have been already standardized, automated to a certain extent. I think this would be the -- or these are the first areas where we can see quick wins, I would call it, low-hanging fruits. Typically, these are already situations where we took benefit from wage leverages from higher-cost countries in the lower-cost countries, so I would look at the phasing of -- when you look at the two pie charts that we have provided that maybe from a pure FTE perspective, things are going to move a little bit earlier than on the cost side because the reading -- the two, I would say -- the GC opportunity we have might actually only get into full scale in the second half of 2027. That's why we have a phasing that is, I would call it, backloaded in a very traditional sense.

The next question comes from Kulwinder Rajpal from Europe.

So two-part question on Road Logistics. Firstly, I wanted to understand how much of the EBIT growth in Road Logistics stem from you going from sea to road in the Middle East? Because if I remember in the last quarter, we had 90% to 95% of EBIT coming from the base business in Europe, plus some growth in U.S. So I wanted to understand how that's changed sequentially. And then when you talk about the emerging demand from tech sector within road logistics, is there any carry-through from your other businesses that is translating into that demand on the road side?

Yes. Stefan speaking. So no change to my statement during the Q1 call. the business in the Middle East is good and developing quite nicely, but it is far too small, so 95% of the business is still, again, Europe, U.S. and Asia. And even if we see growth based on the current situation that it's not moving the needle significantly on EBIT level. That's the first one. The second one is I think what we need to understand and what we really push is the cross-selling between the different business units. As I mentioned before, it's not only the inbound for the tech sector when it comes to air freight is as well certain power lanes where we come in with our seed offering, in particular the last mile opportunities, leveraging our road business unit. And now last but not least, but very important vendor management and contract logistics capabilities with large warehouse operations coming in for us in the U.S. So here, you clearly see that we leverage all of 4 business units in order to support the customer end-to-end, and we cross-sell as much as we can between the different units in order to get a higher share from this industry.

And just a quick clarification on volumes in air. Was there any element of preordering that you saw in your trade volumes? And do you see that happening in your current discussions for customers?

This is difficult. This is -- you mean in tech, right? So the ordering. There was a -- so the second quarter was definitely strong. And we had a discussion with some of you already -- is that front loading? Or is it sustainable in terms of restocking I would exclude the tax sector on that because the dynamic here is completely driven. So I believe we do not see that from a tech sector hyperscale, there is any shift in demand or softening in demand coming into the next couple of quarters. So I would say this is an ongoing high demand request from customers on the inbound flows.