Hello everyone, Welcome to our Q2 Earnings Call. Before we start the Q&A session, we will have our Investor Relations Officer reporting the numbers of Q2.
Now we are going to review the financial numbers of Q2. Q2 revenue reached $124 billion, marking a record high quarterly result. This represents a 20% year-on-year growth and a 4% sequential increase. That said, the strong appreciation of anti-dollar against the U.S. dollar in Q2 somewhat weighed on the reported growth, making it a little bit lower than originally expected. Gross profit in Q2 was $44 billion, up 25% year-on-year and 17% quarter-on-quarter, driven by robust shipments from our data center-related and passive component businesses. Shipping margin in Q2 was 35.5% versus 34.1% a year ago and 31.8% in Q1. Both the dollar amount and the margin were record highs.
Q2 OpEx increased by 14% year-on-year and 7% quarter-on-quarter, with SG&A growing a little bit faster than R&D expenses in Q2. R&D expenses to sales was 9.5% compared to 10.2% a year ago and 9.3% in Q1. SG&A spending to sales was 10.9% compared to 11.2% a year ago and 10.7% in Q1. The OpEx ratio declined to 20.4% from 21.4% a year ago, but slightly increased from 20.0% in Q1. Thanks to the favorable shipping margin, operating margin in Q2 substantially improved to 15.1% from 11.8% in Q1 and 12.7% a year ago, reaching an all-time high. The operating profit in Q2 increased by 42% over a year and 33% over the previous quarter. In terms of the segmentation performance, year-over-year infrastructure delivered the strongest revenue growth, followed by power electronics.
Thanks to the robust data center demand, mobility, however, continued to struggle due to ongoing market weakness, while automation was impacted by broader macro-economic headwinds. Quarter- over- quarter, power electronics recorded the fairest growth, while mobility and infrastructure showed very modest sequential improvements. Automation, on the other hand, remained relatively soft. Earning-wise, except automation, all other segments recorded varying degrees of q-o-q improvements. On a year-on-year basis, infrastructure posted the strongest profit growth, followed by power electronics. In contrast, mobility swung to a loss amid ongoing market softness, while automation also came under pressure from the macro headwinds. Q2 non-operating profit was $900 million versus $1.6 billion in Q1 and $1.9 billion a year ago. The loss in foreign strategy. Income was driven by the strength of Thai baht over U.S. dollars, while the losses from the investment were mainly related to some minority holdings.
In Q2, we had $19.6 billion profit before tax, up 30% year-on-year and 25% quarter-on-quarter. Q2 EBITDA reached $26.8 billion, up 24% year-over-year and 17% quarter-over-quarter, setting another all-time high. Q2 tax expense was about $4.2 billion. The effective tax rate in Q2 was 21.6%. Net profit after tax was about $13.9 billion, up 40% year-on-year and 36% quarter-over-quarter. The Q2 EPS was 5.37, achieving a new historical high. Now we have a look at the accumulated numbers of the first half. The first half revenue was $242.9 billion, up 25% from a year ago. Gross profit was up 31% year-on-year, with the gross margin improving to 33.7% from 32.32% a year ago. The OpEx in the first half was up 18% year-on-year, with SG&A up 18% and R&D up 70%. Thanks to the improved economics of scale, SG&A as a percentage of sales dropped to 10.8% from 11.4%.
The OpEx ratio also shrank to 20.2% from 21.5% a year ago. As a result, operating margin in the first half was up 59% year-on-year, and the operating margin improved to 13.5% from 10.5%. In terms of the segmentation performance, infrastructure and power electronics delivered strong revenue growth, followed by a moderate pickup in automation. Mobility, however, continued to face soft demand. On the earnings side, infrastructure recorded the most notable profit expansion, with power electronics and automation also seeing some improvements. In contrast, mobility swung to a loss amid a continued weak market environment. The non-operating income came in at $2.5 billion, down from $3.3 billion a year ago. In the first half, we had $35.2 billion pre-tax income. The first half tax expense was around $7.9 billion, representing a 22.3% effective rate. The net profit after tax was $24.2 billion versus $15.7 billion a year ago. The EPS in the first half was 9.31, up 15.54% from a year ago.
First of all, congratulations on the very strong results this quarter. I have two follow-up questions to ask. First of all, we know that we have a very long-term ambition and long-term plans regarding the data center-related businesses. We also saw one of your peers, the American peer, just gave the market a pretty positive outlook for the next couple of quarters. Could you please give us some updates or maybe more color on your plans regarding the data center-related business, your product roadmap, and so on and so forth?
First of all, in response to your question, as I previously or repeatedly explained, the data centers, especially those hyperscalers, continue to invest aggressively into the data center, AI data center deployment. It seems to be a pretty clear trend right now.
Of course, going forward, especially if you're looking into the next couple of years, it's still very highly subject to whether this AI deployment can really translate into the profit contributions or the revenue contributions. If you look at the recent results just announced by the U.S. hyperscalers, I think if I remember correctly, at least some of those hyperscalers actually reported pretty decent results, which can somewhat underpin their continuous investments. That would be the preliminary view on the AI investments or data center investments by those hyperscalers customers.
I may have some points to add on. First of all, I do believe that in terms of the product footprint, we do have a pretty complete spectrum in terms of the product footprint to the data center market, including our power products, power solutions, and the cooling solutions, such as the liquid-to-air cooling solutions and also the liquid-to-liquid cooling solutions. Speaking of the liquid-to-air solutions, we do believe that we actually enjoy pretty high market share in the liquid-to-air liquid cooling market. Speaking of the adoption rate of liquid-to-liquid systems, I think it's something we are not entirely certain about. While it does offer better energy efficiency than liquid-to-air and significantly lower CDU costs, the infrastructure complexity costs and long deployment time make the adoption challenging. Ultimately, the penetration rate will depend on data center operators' strategic considerations, which remain unclear at this point. In contrast, liquid-to-air systems are gaining traction due to their faster deployment.
The second question is related to your service revenue. If we look at the service revenue accounts for the total revenues, for your U.S. competitor, say, Vertiv, they actually account for quite a big part of their revenues. When will we see this service revenue become a big part of your revenues?
In terms of the service revenue, currently, it indeed accounts for a relatively minimal part of our revenues. I think it's going to take quite some time before we see it become more meaningful in terms of the service revenues.
I also have some points to add on in terms of their service revenues. As a head of the M&A activities at Delta , we actually have been looking at, we are also looking at, the opportunities in the market and seeing if there are any opportunities that we may acquire some of the service target companies that offer the services to those data center customers. Also, speaking of the service, if you look at the data center market, you can actually further separate it into the white space and the gray space. It's actually more complicated than we thought. We actually keep an eye on the market in terms of the potential acquisition opportunities in this area. In terms of your question related to how we view the AI outlook in the second half of this year, I just checked the data for the CapEx spending, CapEx plan of Meta this year. It should be $72 billion. I think in terms of the outlook for the second half, it's probably not going to see any significant turbulence. That would be my point of view.
My question, how much was the impact of tariffs in Q2, and has it been passed on to customers?
Most customers have agreed to absorb the additional U.S. tariffs, though we covered some upfront, which was put as operating expenses. However, considering the scale of Delta , the overall impact was not very significant.
I assume that your modern expansion, strong modern expansion in Q2, was mainly related to the very fast growth of your AI products?
If you look at our Q2 shipping margin, I think it's just about 1% higher than the previous year. Of course, the overall server power or overall data center related businesses, the margins are higher than other product lines. We think, of course they are higher, but just, yeah.
What is your current adoption rate of Power Rex? Will large scale adoption only happen after Kyber platform launches? If not, under what conditions might CSPs adopt earlier?
As AI servers become increasingly complex, the trend of moving power supplies to a centralized rack is becoming more apparent. We already have some designs ready with customers, and once they decide to adopt, we can quickly enter mass production. The Kyber platform requires not just a standalone Power Rex, but also high voltage DC specs due to its extremely high power consumption, which requires higher voltage to manage current and heat within controllable limits.
In response to your questions, in terms of the deployment of liquid-to-liquid solutions or systems, as we just explained, if you want to really adopt or deploy the full liquid-to-liquid solutions, for data centers, they actually require very high requirements in terms of the infrastructure or the facility level. At least, the liquid-to-air solutions are going to be probably the mainstream for a while, considering the reasons we just mentioned. For those reasons just explained, if you really want to adopt the liquid-to-liquid solutions, you may, of course, either build up new data centers or you have to retrofit the existing ones. Considering all those reasons, we may probably not really see that happen in the short run.
In terms of our overall production plans for our data center related business, considering the cost, the manufacturing cost, and the supply chain, I think it's more likely that we will continue to produce the components in Asia. Of course, if you look at the tariff of Thailand, if it remains high at the 36% level, we may not continue to expand or to move or shift our productions to Delta Thailand, to Thailand, if the tariff remains at this current level. In terms of the capacity in the U.S., we already have some capacity in the U.S., and we are also building a new plant. The current expansion plan actually has not been adjusted during tariff changes. Our primary purpose for expanding in the U.S. is to be closer to customers and offer better system solutions and services. However, manufacturing costs in the U.S. are much higher than in Asia, and the lack of a complete supply chain makes full component manufacturing very difficult. System assembly is visible, but replicating Asia's supply chain is very challenging.
My question is related to the humanoid robotics. For example, if Delta is going to, or maybe has the plan to, participate or to penetrate into this market, where you may consider to produce the components or the products of the humanoid robotics?
We do believe that the market for this humanoid robotics is enormous in the future. The key, in terms of the service robots, is not really related to the components or the hardware you produce, but it's more related to the software, especially the AI embedded in the robots. You also have to separate the fields, the application fields, so whether these robots are for the industrial purpose or the robots are going to be in the service field, especially to be cowork or co-live with human beings. In terms of the specs, in terms of the designs, these could be hugely different between or among those different applications. In terms of our plans, this robotic market, we are going to actually launch Delta Robotic Research Center in August. Maybe we will have more insights to share after we dig more deeply into this market.
I actually noticed you showcased your containerized AI data centers during the Computex event this year. Could you please share more details or give us more colors related to, in terms of your plans on this containerized AI data centers?
If you look at the current AI data centers, they are mostly operated or built by those hyperscale customers. In the future, when the colocations or maybe other enterprises, they are also going to build up their own AI data centers. Actually, the solutions for those customers, providing an integrated solution and system to those customers, and especially it actually can be fast deployed. I think the customers, the target customers, are different.
My question is related to the tariff. Currently, we still don't know how high the Taiwan tariff is going to be, but do you have any scenario plannings if our tariff is going to be higher or maybe just on par with, or maybe lower than Japanese and the Koreans' tariff rate? Do you have any scenario plannings for the tariffs?
First of all, in terms of the tariff, because still, at this point, we don't know the tariff, what level of the tariff it's going to be for Taiwan. For now, as we just explained, our customers have absorbed most of the tariffs by themselves. It actually hasn't really had any significant impact on us. In terms of the manufacturing footprint, we actually also have a pretty diversified manufacturing footprint. We don't just have batteries in Taiwan, but also elsewhere.
The tariff is something that is totally out of our control. As a company, we do have to be more agile and more flexible to fulfill customers' needs. Actually, since 2015, our strategy of being more diversified, not just in terms of the product lines, but also in terms of the production capacity, the geographic locations of our capacity.
As you can see, right now, because at this point, we are having this meeting, we still don't know the tariff level. There is no way for us to really plan your capacity based on the pretty unclear and uncertain tariff levels. In terms of the capacity, it's not just for this year or for next year, it's actually for the long-term plan. In terms of our long-term capacity planning, it's not going to really be adjusted by short-term changes. What we can actually do is be more agile and more flexible and then quickly respond to any crisis. That is what we can do. This kind of short-term change is not going to really impact the long-term strategic direction and planning.
My first question is, because as we know in Q2, some of your customers actually had some pulling orders because of the tariffs. Should we still expect the normal seasonality, which means that we will see the high season and peak season in Q3?
My answer is, I think what I can say is I think the second half, if things go smoothly, the second half of this year is still going to be better than the first half of the year. For next year, it's still too early to comment. In terms of your question related to our sampling plan, final sampling plan in the future, as I just planned earlier, as I said, for the U.S. market, I think it's very likely that we will have the final assembly in the U.S., but still, for the components manufacturing, to manufacture or to produce the components in the U.S., considering the cost and the supply chain, it's not really feasible. It's not really a feasible thing to do.
I have a question related to your capacity planning in Thailand. Considering the current higher level tariff to Thailand, is there any ongoing expansion planned in Thailand?
Yes, we are actually still in the process of building up our new factories in Thailand. As I said, those are for the long-term development and long-term planning. It's not going to really be adjusted by the near-term changes.
In terms of your AI power, sorry, AI server powers, can you give us a split between whether they are for the GPUs or for the ASICs?
Sorry, we don't have the split.
Okay, I have a question regarding the R&D expenses as a percentage of your total revenues. As we can see, the R&D as a percentage, as a percentage of your sales, was about 9.5% in Q2. Should we expect this 9.5% is going to be the new norm in terms of the R&D percentage level going forward?
I think, going forward, we will continue to invest a lot into R&D. It's not just for existing businesses, but also for new businesses. Our plan is more like to maybe maintain or keep the R&D expense ratio maybe around 9%. Of course, it's still very much subject to the revenues, the top line growth. We actually continue to always keep an eye in the market, and we always try to maybe get into new businesses.
We will continue to invest into the new areas. Of course, we have to consider manufacturers very carefully, for example, whether those new areas or new businesses are going to be related to our core competencies, like maybe they are the natural extensions of our core businesses. That would be a very critical criteria when we are assessing the opportunities of the new areas or new business.
My next question is related to the solar panels. As we can see, recently, the typhoon in Taiwan, the southern part of Taiwan actually has hugely damaged many solar panels in Tainan. Do you think that is going to be a big issue to Delta ? Also, as we can see, in the U.S., the government seems they are phasing out some of the subsidies for the solar industries. Is that going to be a big issue to Delta ?
In terms of the damage, the recent damage in Tainan, actually before we launch any renewable energy plants, we actually always make sure to conduct very strict due diligence. Even though recently we have seen very serious damage in the southern part of Taiwan, to Delta , we actually only had very minimal damage in terms of the solar panels set up in our own factories. We don't actually see that as a very big issue to us. Speaking of the second question related to the U.S. government, it seems like they are kind of phasing out some of the subsidy for the renewable energy, especially maybe the solar systems. I think we are not in a position to really comment on the policy, but still we do believe that, for the enterprises, there is still pretty high demand for the renewable energies in order to meet their decarbonization target.
We will have the final questions from the audience.
My question is still related to the tariff issue. The Thailand tariff currently is 36%. I think if the tariff of Thailand does remain 36%, it's definitely going to be a big impact on the business or the orders in Thailand for next year. What do you do? Thank you.