Thank you. Thank you. Welcome everyone to our Innovate at the Core event. I'm very happy to see you all, especially here in Taipei. To our partners, our customers, and all our friends from over the world, thank you all for joining us today.
The AI revolution has only just begun. Every day we see many new AI applications and new adoptions to improve almost every aspect of our work and our lives. Aggressive partners are coming to Super Micro and they all are asking us for help. Help to develop systems and solutions that can accelerate their AI cloud, enterprise AI, sovereign AI, or other AI applications.
As the AI revolution continues to grow everywhere, for every application, and looks never-ending, I'm super excited to share our latest experience, our new innovations, and our total solution with you. Most importantly, I want to invite you to grow at our speed, at our pace, or even much faster.
You know, in this incredible AI era, most of the people are racing as fast as they can. Every company, every country, every data center wants to ride this AI wave. But how can you best do it? Not just do it, but best do it. Optimize the chance. If your data center infrastructure, for example, is delayed for one month, three months, sometimes six months, you are losing massive opportunities. And we all know that time is money, and in AI, time is much more than money.
How fast can your technology investment be online, start to make money, and start to generate profitability? Think about that. Even just a few months or a quarter today, your business may get hurt. Not just millions or even billions of dollars of income loss, but the whole business opportunity may disappear. Simply put, every single day your system and your data center sitting idle is a direct hurt to your bottom line.
That's why we are focusing so much to help you, our key partners, to improve this one key metric: time to online, time to operation, time to make money. The good news is Super Micro can help. We are here.
Look at this equation behind me. Time to market. New technology, new product, time to market. Plus time to deploy. Not just send the product to your warehouse, to your data center, but to deploy and make them plug and play online quickly. Time to online. Once online, then you can start to generate profitability.
We can ship new generation AI platforms to you earlier. Make your AI data center online faster. How about three months faster? How much impact? How about a year or six months ahead? I'm going to show you how we can work together to achieve that.
How about that? This is only a small fraction of our famous building block solution. As you can see, especially old partners, right? You know Super Micro has the most optimized system, exactly optimized for your business. Because in most cases, we are able to engineer new generation systems quickly by leveraging some or existing subsystems, like some other, we can share some enclosure, power supply, and so much more subsystems.
The new systems created by this building block don't just work fine, it's fully optimized. Optimized for performance. Optimized for performance per dollar, for energy efficiency, and even for low cost for you. Optimization based on our building block solution. This building block DNA is why we are always the first to market since 33 years ago. Even now, we always have a faster time to market advantage, optimize AI infrastructure quicker.
This is the reality. Over the past couple of years, Super Micro has successfully deployed the oldest, absolute largest AI data centers. We delivered this massive, cutting-edge data center on time, tighter timing, and achieved incredible tier-one customer satisfaction. And let me tell you, these customers are some of the most picky, most demanding technology operators in the whole world. They demand perfection. They demand extreme efficiency. And we deliver exactly what they need with close cooperation with our partner, vendor partner, and customer.
We have built our orders for the Jigawatt Nikkuri DLC AI data center in less than five months two years ago. And recently, we just indeed finished another Jigawatt data center in about three months. So faster and faster because we leverage our experience, leverage our building block solution, leverage our data center building block solution.
Okay. For many long-term partners here, you know that we have been one of our original server, storage solution providers on the market since the early 1990s. So we don't chase the movement. We have been doing this for over 33 years, always time to market. Today, we continue to have the broadest system-level building block solution in the industry ready for you.
We have our X14, H14 portfolio in volume deployment since many years ago. And now we have a X15 Intel CPU-based, H15 AMD processor-based, and N15 Nvidia CPU-GPU-based solution, and a new one, A15 ARM-based system and rack solution. So all those are ready for OEM order today.
We also have those brand new high-density platforms that I'm super excited about. They are incredible platforms for cloud, enterprise, edge compute, and especially for agentic AI. With the highest core density, massive memory bandwidth, and flexible configuration, we will be able to ship them very soon, either system level or cluster level.
And look at right here on the stage, our massive X-GPU solution. We are one of the largest providers of Nvidia Blackwell GB300, B300 HGX solution providers. Indeed, in the last nine months, last three quarters, we shipped more than 8,000 racks so far. And we ship also very high volume B300 HGX because, as you know, we are building block solution needed. We always ship a product to customers earlier than others.
And we are getting ready for the upcoming Vera Rubin NVL72. Indeed, we already shipped some samples to certain customers, and so far their experience was great. And we are preparing for high-volume production for you to order today. And then Rubin HGX, right? Rubin HGX, we support a non-64 GPU or can be with another different configuration. Right? Kind of like this configuration, we can support 64, 128, or 144 R200 systems to you. And samples are about now, and production just in the next few months.
And also Vera CPU-based, right? Rubin-based and Vera-based, right? Vera-based, you know, CPU, Nvidia CPU-based. Again, kind of lots of different configurations already about ready. Now it's time to play, right? The first come, first served, right? And for sure, you are a first priority because you are our close partner, right?
So all of those are the best optimization for large-scale, agentic AI, HPC, and agentic AI. Right? Why is this modularized design, building block solution approach so critical, so good for the market? Because global IT build-out are changing very fast. As you know, data center and enterprise now moving from single high-performance workload to more complex AI, solving AI, agentic AI, and many kinds of AI applications.
So if you try to optimize your server and data center from scratch, you will run into a massive engineering bottleneck, and those are time-to-market delays. That's why based on modularized building block solution, direct scale, plug and play, that's the best way to go. And Super Micro, you know, we are the expert for those approaches for 20 years, more than 20 years, right?
So, and more importantly, I mean, how quickly can you deploy and make the product available, right? And how can we have high-volume deploy for you, right? So this brings us to time to deploy. In the old days, deploy data center meant receive thousands of boxes and you unbox them and integrate to a rack and then testing everything in data center and cabling and then kind of take lots of time, right? So now instead, we ship you in a kind of completed rack. Everything is pre-assembled, pre-validated. So when we ship to your data center, it's ready to run, a true plug and play.
We have some customers say, hey, they buy from others, they plug and cannot play. They plug and have to fix and another fix and finally play. So take too long. So we have been working together for many years. You know our plug and play is pretty much true plug and play with everything pre-designed, pre-optimized, pre-tested. So when we ship to you, it's fully ready for a plug and play.
So we save lots of labor-intensive jobs for you and minimize the order risk. We ship everything to you with all pre-validated data, and that's why when installed together, it is just ready to work, right? So that's TTD.
Also, because the event is getting bigger, right? I mean, as you may know, two years ago our revenue was just $15 billion, and then $22 billion, and then this year can be $40 billion. Next year can be much, much more than $40 billion. Right? So indeed, we have been expanding our manufacturing capacity very fast.
For example, in the USA, we just added a building in the last three quarters. We added 1.4 million square feet, all in San Jose, because not enough time to extend to other states yet, right? So we are extending very fast in Silicon Valley and now have a much higher capacity for you. In Taipei, in Taiwan, we also added two more buildings to exactly enlarge our capacity so that we can do huge-volume production with best quality and ship to you with scale, plug and play, and ready to run.
So even three years ago, our capacity was only about two to three million square feet, but today, globally, we have six million square feet. So our capacity is fully ready for your big demand. We help you by consolidating compute, storage, networking, power, and cooling into complete, pre-tested, again, plug-and-play solutions. When it leaves our shipping bay, it's ready to go straight to your data center floor and ready for plug and play. Right? Look at this beautiful integrated rack sitting here. Right?
I guess this is Vera Rubin 72, right? With Vera Rubin 72, NVL72, with fully optimized battery power shelf. And some customers, if you want BBU, battery backup system, we have that for you. Again, everything kind of pre-validated and fully optimized for you. And even this rack, you can see this rack is so strong. I guess this rack can carry 5,000 pounds of equipment, right? And the reason why we design our own rack is to optimize everything. Make sure the mechanical is fine-tuned and is easy for your installation, easy for your maintenance, and also the rack is strong enough even for a large earthquake. The system makes sure it's stably running your data center without interruption.
So, no downtime means a chance to win, to make more money. Now look at how we handle power. Again, we design our own power shelf, including a BBU, and kind of we provide a complete solution, including a water tower, dry cooler. Right? For some cities, the water is not just sufficient. People, the city may want you to save power, and you like to save power, right? So you can go for dry cooler, water cooler, kind of an outdoor CDU, liquid-to-air. When your data center is a traditional air-cooled data center, but you want to run a DLC, liquid cooling, right? You just have a liquid-to-air cycle, 250 kW, 500 kW, and then you can run your DLC direct in a traditional data center right away without spending months of time to convert your data center to install DLC equipment and to spend the money, right? Especially for enterprise applications, AI, or kind of small-size AI, this ready-to-go, ready-to-install will be a bigger help.
So let's talk about cooling. Again, we provide all different kinds of CDUs, which different sizes: 1 megawatt, 3 megawatts, or 5 megawatts, or kind of like an indirect CDU. Indeed, Super Micro is one of the first companies to design and manufacture liquid CDUs, right? No CDU, kind of 125 kW, 250 kW, or even a little bit larger size is coming, right? So either indirect rack or indoor CDU, kind of trio, Super Micro has all different kinds of trio. Again, all trio are pre-designed, pre-validated. So when we ship to you, ship trio to you, ship CDU to you, everything already well-tested, well-validated. Even the tower, when we ship everything to your data center, it's kind of ready to plug and play, so you can make your cluster go for online right away.
To give our customers total peace of mind, we also developed a revolutionary breakthrough in cooling chemistry. Our new coolant, we have some new coolant formula that features up to 1,000 times higher electrical impedance than standard coolant. For example, here. Why the 1,000 times higher impedance can be a great help? Well, I mean, first and foremost, in case you are wondering, our DLC system don't really leak. Indeed, our DLC system never leaks. But in case there are small leaks, when you have a high electrical impedance CDU, the system keeps running. Not like a regular coolant. Unfortunately, you have leakage, your system will be shut down and you have to send for RMA. But with unique high electrical impedance CDU, your system keeps running. And for maintenance, you just shut down your system, evaporate the cooling, and then your system can continue to run, right? And for some data centers, if people really want a very safe concern for your staff, this high-impedance cooling also prevents electric shock and short circuits. Right? So anyway, I mean, the customized cooling that can help you to build a really reliable, unique data center or cluster.
Okay, time to online, right? When the physical racks are sitting on your data center floor, you reach the next critical phase: time to online. How can you have a quick time to online? Loading racks onto the data center floor doesn't make you money, right? You have to make sure it's ready for operation. You must push it to production. You need to configure and run the network, provision the operating system, install and validate the management software, test thermal, and get the workload running. And Super Micro provides you all of those support, all of those software.
Recently, we also introduced our 1.2-terabit CPO optical switch, engineered specifically to remove communication bottlenecks in massive-scale computer clusters while delivering ultra-low power, lowest latency, and unmatched port density. Furthermore, to support the thermal realities of next-generation dense AI workloads, we are delivering a full liquid-cooled 800G and 1.6-terabit switch platform, all almost ready for sampling, and we can start production by the end of Q3.
To eliminate traditional third-party integration delays, we pair our hardware with our own enhanced industry-standard SONiC networking operating system, supporting both the community contribution and our full-supported enterprise version. So after your choice, on top of all the networking layer, we provide an intelligent software stack, our SuperCluster software suite, that includes our SuperCloud Composer, SuperCloud Director, SuperCloud Automation Center, SuperCloud Developer Console, and other software that gives you a unified, single-pane management platform across your data center. Capable of tracking real-time telemetry, power usage, and advanced NIC detection across tens of thousands of systems or racks simultaneously.
This software unifies our traditional system management to data center infrastructure management to solve complex power, cooling, scaling, and other challenges. Our DCBBS software products are critical one-stop solutions to help you build your data center with faster time to online, again, time to make money. Optimize quality and performance at the most efficient cost. Not just quick, better quality, but again, we provide the most cost-efficient solution for our partners.
We know hardware delivery and software support are not enough for lots of data center customers who want to achieve the best time to operation. What else to guarantee immediate revenue for your business? We don't just ship product. We will help you on the path to activation with our expanded global on-site service and life cycle support. All you need to have is what? A good plan, with great passion for your business, and of course, money, right? And we will do the rest for you.
So it's kind of a one-stop shop, one contact window, and then your data center, your cloud, AI cloud, ready to run. So Super Micro's service and deployment team are on site the moment when your racks arrive to handle complex physical, secondary fluid loop, and precision DLC pumping connections. The team will also build an automatic, real-time validation protocol to execute thermal, networking, and compute stress testing directly on the data center floor, eliminating any guesswork.
And most importantly, we are vastly expanding our post-sale support to ensure continuous and uninterrupted uptime. We know that in the AI era, even a moment of downtime costs a lot of money, maybe millions, maybe even billions of dollars. So to guarantee total peace of mind for your operation, in addition to having on-site technical service experts, we have established local depots to provide all the components, the spare components that you need.
So this means we can deliver critical replacement parts exactly when you need them. Our comprehensive life cycle support ensures your IT team can 100% focus on your core strengths, knowing Super Micro has your back. So DCBBS is how we help you achieve the best time to operation, again, equal to time to make money. Our DCBBS did not stop here. Super Micro has a few other features we will open to you when it's ready.
Okay, we talked about lots of, covered lots of ground here. We all know that the agentic AI opportunity is huge today. And today's special guest will help us win more of those. Before I invite him on stage, let me first introduce our CRO, Matt. Are you here?
