But I’ll just point out. Now, it’s kind of the upper part of this graph. Most of the data in biology has yet to be created, okay? So we do have great starter assets. Ginkgo’s got bigger ones, I think, than most places. But what really we want to do is generate more data using the automation, so our customers and partners can do things like fine tune an AI model offered by, for example, another service company, without that company needing to develop their own automated lab. And so you’ll hear about that in a second with our partnership with Cradle. That I’ll talk about in the next section. But some companies will also want to generate their own huge data assets to build, for example, maybe a proprietary foundation model in a certain disease cell line that they’re interested in as an example.
We can do that too. This is similar to the business model of Scale AI in the tech industry. So on that tech chart, I showed OpenAI on top of Scale AI, right? So OpenAI pays Scale AI to generate a lot of the data they use to train things like ChatGPT, Tesla pays Scale AI to basically analyze images and highlight. That’s a dog. That’s a pedestrian right to help train their models for self driving cars. It’s a company in the business of generating labeled data to feed into other companies machine-learning and AI models. Absolutely happy to do that at Ginkgo at large scale for customers. Okay. So one of the best investments we’ve made was bringing in Zymergen’s proprietary software and automation technology to dramatically enhance both the scalability and flexibility of data generation at Ginkgo.
And you can see within a year of the acquisition, we had installed the technology at our site in Boston. This is one of the things we’re so excited about, because they had this automation software team that had been supporting Zymergen’s efforts, and so they could drop right in. And since then, we’ve actually evolved. We call these racks reconfigurable automation carts. And so without nerding out too much, what’s exciting about this is we can easily plug-in new equipment into a big centralized automated system without needing to do a whole new automation rebuild. And if you live in the world of lab automation, when you have a great idea for some new thing to automate, you can start the clock and you’ll have a big automated system ready six months to a year and a half later.
Okay. With the rack system, if you have a new piece of equipment, you want to add to an integrated automation workflow, we can pop that right into the system and then using software, add that equipment into the workflow relatively quickly. And importantly, we are now able to manufacture these racks much more expensively or inexpensively. We’ve increased our ability to manufacture them fivefold while keeping our hardware teamed flat. And with this increased production capability, we plan to deploy 3 times as many racks in 2024 as we did in 2023 at Ginkgo. And we aren’t the only ones that believe data generation is important. Just to give you example, on the government side, just a few weeks ago, we hosted the House Select Committee on the strategic competition between the United States and the Chinese Communist Party.
Congressman Mike Gallagher, who chairs the committee, noted that they came to Ginkgo, to talk to experts and figure out the right strategy so that we, and not the Chinese Communist Party, can dominate AI and set ethical rules of the road. So what did we talk to them about? Well, we talked to them about the facility you can see here, Biofab1. This is actually a render, but I can look out the window over here and see it. And I took a tour of it last week, and that tall top floor is actually filled with things like huge cooling units and other hardware that’s needed for running a building that is really meant to be filled largely with automated labs, with these racks I just showed you. This looks and feels like a data center, right? If you went and saw an Amazon data center and things like this, it is built to purpose buildings that are going to be filled with a certain set of hardware and have the infrastructure to enable that.
That’s very much what we’re doing here, except instead of a bunch of servers, what we’re going to have is a bunch of rack automation, hopefully generating the data that powers infrastructure services across the biotech industry. All right. Onto section two, okay, so I’m super excited about this, as we had the chance to announce our technology network with its 25 inaugural members just yesterday. We want to make infrastructure services the way that R&D work gets done in biotechnology. And Ginkgo cannot do that alone. Ideally, we want hundreds of new tools and service companies to flourish and to move that percent of R&D spending currently going to outsourced R&D services. And Mark showed some numbers of like the total R&D spending at biopharma companies compared to, for example, our revenues today.
And my estimate is across the entire sort of research services industry, we’re maybe at less than 5%. It could be even as low as less than 1% of total research spending going to outsourced infrastructure. It’s mainly going to on-prem work, right? People doing work by hand in labs, people buying reagents and things like that. And really, I would love to see that shift into infrastructure services like it has in tech. We are big believers that brilliant technology is being invented outside of Ginkgo. There’s not a big not invented here culture at Ginkgo. We partner deeply with many life science tools companies to integrate their tech in our workflow. To give a specific example, we were the first large DNA synthesis like big contract for Twist when they were a very small startup back in the day, and we’re still one of their largest customers today.
We have also conducted over 15 M&A transactions over the last several years, including three we announced yesterday to bring certain technologies in-house. And quick update about why we’re excited about those three Patch Biosciences, Proof Diagnostics and Reverie. So Patch is going to bring in large data sets ready to deploy ML models and downstream assays for promoter and RNA stability and expression. And so we do a lot of work in RNA right now. So that’s very exciting. And a quick and good fit. Proof offers massive libraries of obligate mobile element guided activity, omegas for short, with RNA programmable nucleases and nickases that will enhance our code base and overall offerings to pharma companies. This is more sort of microbiology technology, sort of in the kind of editing space, and then importantly, also a great data asset.
So it’s more to add to that big data pile I was mentioning earlier. And lastly, Reverie is an AI company focused on leveraging computational chemistry and machine-learning to accelerate drug discovery programs. And this acquisition will allow Ginkgo to significantly accelerate our own build out of AI program development. And I’m excited about the technology, but in all of these cases, I’m especially excited about the teams that are coming over in these acquisitions. It’s a real speed up in terms of our human capital build in those areas, and just awesome people. Okay. But the thing I want to highlight today is our technology network. And just as I highlighted several of our suppliers and acquisitions, we are now partnered with over 25 different companies that can benefit our network.
All of those companies bring something different to the table, whether that’s a focus on pharma, AI, enzymes, manufacturing, or biosecurity. And we believe that this network is just the start of driving a cultural change. Like I mentioned earlier in biotech R&D, from doing things in-house to doing them with outsourced infrastructure services. And I’ll give you a couple examples so you get a sense of what might be possible. So Cradle, this is an exciting company, developing AI models for DNA design in the area of enzyme engineering. So let’s say a customer comes to Ginkgo to improve the activity and titer of an enzyme, and then after they find it, they want to express it in high titers in a cell, for example. So that would be a big end-to-end project at Ginkgo take a little bit.
And so I’ll give you one part of that where we’d be integrating in Cradle. So we’d first go through our normal steps of designing and screening a metagenomic library to find an ideal candidate for our customer. So, remember that big 2.7 billion gene database I showed you earlier. Great. We source an interesting seed sequence from that library. Now, here’s where we can leverage Cradle’s generative AI platform to then computationally create a library of protein designs and their respective DNA sequences based on that seed sequence. And then the magic those designs could then be synthesized, probably at Twist, given how Ginkgo works, and then cloned and tested using Ginkgo’s automated labs and their foundry, and eventually in Biofab1.
And once we’re happy with the results, the leading candidates would be transferred to our customer. And we might go through that loop a few different times for a customer project. And importantly, we’ve been able to pull together, in that case, a couple of different providers. Someone who’s really providing computational horsepower and AI magic, Ginkgo providing a lot of lab infrastructure, Twist at the DNA synthesis side. Okay. And so we’re also excited to help Cradle’s customers be able to engineer proteins, really from the comfort of their browser, as Cradle likes to say. And we can do the let lab work for them at Ginkgo. I want to say, I really love this vision. Okay, so I did a PhD. I have – I spent five years holding a pipette, R&D scientists put down your pipettes, right?
In a model of outsourced infrastructure services, drug development scientists will spend their time designing experiments and workflows, understanding and reading about the biology, so they know what they want to try to attack next as they’re developing these biotech products, not spend their time manually moving liquids around a lab bench, right? It is just not the best use of the folks who have all this experience in both experimental design and biology and in a world of outsourced services. I love Cradle’s vision of doing that all from the comfort of your browser all that lab work, okay. So another partner I’d like to highlight with sort of an illustrative case study is bit.bio. So bit.bio has created what they call ioCells, which are human iPSC cells, differentiated to represent various wild type and disease models, which are needed to effectively study drug designs and better predict in vivo behaviors.
The bio cell lines are a terrific asset that we can incorporate into our programs. They’re customized for a particular disease state a customer may be looking at. And we can test a multitude of different drug modalities across these cell lines and also screen our large libraries of optimized promoters and gene editors, like you heard, for example, in the Patch acquisition, so you could integrate some of those technologies in with theirs. The magic here is that Ginkgo gets bit.bio cell lines running at high throughput, which is great for our customers. And bit.bio gets a distribution channel in Ginkgo, so that they could have more companies licensing their strains and paying them, right. And I think this is key, right? So we need to make it easier both for biopharma companies to access the latest new cell lines like those at bit.bio.
But it also needs to be easier for companies like bit.bio to exist and sell cell lines, right? And as I mentioned, Ginkgo has a 90 person commercial team and growing. So we’re hopeful this is a real asset to tool companies in the space to get their technology in the hands of these large biopharma R&D customers. And hopefully by Ginkgo going out and being able to show a multitude of technologies coming together in these programs, it also helps increase that 1% to 5% or whatever it is of research spending going out to infrastructure services make that number go up for the whole industry. So I’ll end with this. Biopharma customers repeatedly tell us what they care about is increasing probability of success, reducing the time to results, and reducing R&D costs.
Large AI models and big data sets in the hands of the scientists at our customers, we think can really increase the odds of success in drug development. And it’s worth like an extra point here. There’s a lot of noise about automation AI, like replacing scientists and robot scientists or whatever. That is not our belief at Ginkgo. We are not trying to take scientists out of the loop with the infrastructure services model for the industry. This is not like Combi-Cam. It is not like some paper about robotic scientists. This is super powering your scientists to let them design much larger experiments at enormous scale via automation and then be able to handle all the data that comes out of that, using models to give them biological insight back so that they can decide on the next round of experiments to design, right?
It is – again, think of the change over the last 20 years in product development tools in tech, right? If you were a software developer 20 years ago and a software developer today, oh my gosh. The software developer today, they’re like Tony Stark or something, it’s unbelievable, the capability difference in developing products. And they’re still there. The software developer is still there. In fact, there’s more software engineering jobs, right? Because the market grew so much for software with better product development tools. That’s what we’re really talking about here. We’re not talking about removing scientists. We do see programs completing faster. We want to talk about speed now every year as we build bigger scale and as various approaches at Ginkgo get more mature.
And one reason to think of how infrastructure services will be better than traditional by hand work on speed is you can try many approaches in parallel, rather than guessing one, seeing the result of a small batch of experiments and then serially moving on to the next approach. As you can increase your scale of experiments, you can try many in parallel. Finally, cost. Another myth I hear a lot of time is, biopharma doesn’t care about costs in R&D, they just want more success or whatever. If you have talked to an R&D Head, you will find out that they do, in fact, think a lot about the research budget at their company. And we have – as we have scaled at Ginkgo, we are seeing a 40% to 50% average year-over-year drop in the cost of our campaigns.
And a campaign is basically a cycle of designing, building, and then testing genetic designs. The fact of the matter is, in house by hand, R&D work at the lab bench does not get cheaper with scale. It really doesn’t. But our automation does. And infrastructure services will keep getting cheaper with scale every year if we keep growing, just as they did in tech. And I think that’s going to be an important engine for the industry if we want to change how research is done. All right, finally, I want to cover our last strategic topic for the day, which is about how Ginkgo is leading as a systems integrator in global biosecurity infrastructure, alongside our work in cell engineering. Now, I’ve shown you this slide in the past, but I want to reiterate how important it is to plug in global gaps in bio threat data.
We put a ton of effort into response measures globally, but they tend to be slower and less effective than they could be, because we’re so often flying bind or gathering information in a reactive mode. And to give you how I get this in my head, the way that we approach biosecurity today, it would be roughly equivalent to the way we approached hurricane tracking was letting everybody know a hurricane was coming once it had hit the coast of Florida. I grew up in Florida, okay? And that would not be effective. And that is what we do. We wait until where do we track infectious disease? At hospitals. People arrive in hospitals when they are starting to get sick en masse, okay? The hurricane has hit the shore, right? Instead, we should be monitoring for infectious disease at animal husbandry facilities, at airports, at places where people congregate, anywhere we can most cheaply look at a lot of infectious disease genetic data in one shot to get a baseline picture, again, think like satellites, monitoring all the time.
That’s the vision. And within the past two months, and really in the past few days, we’ve made major announcements surrounding our partnerships in this space that will boost our data and intelligence capabilities. Our partnership with Illumina will allow us to more rapidly scale pathogen monitoring programs to new countries. The agreement leverages Illumina’s leading next gen sequencing tools and Ginkgo’s end-to-end, we call these bioradar per my example services. And together, we aim to increase the scale and scope of pathogen genomic surveillance globally. So in other words, remember things like variants in COVID monitoring, the genetic sequence of different diseases as they are spreading around the globe. So we have more visibility and we empower in particular countries with local capacity.
And that’s what I want to talk about next. Just this week, we announced that we’ll be launching the first Center for Unified Biosecurity Excellence in Doha, we call it CUBE-D, in partnership with Doha Venture Capital and the Qatar Free Zone. And when complete, we expect CUBE-D to support major bioradar programs both in Qatar and across the broader region, serving as a key hub in Ginkgo’s global network. And CUBE-D will be a foundational piece of biosecurity infrastructure to advance pathogen monitoring and biosecurity analytics, enable the development of biological intelligence, and support the development of next generation biosecurity leaders globally. That’s also very important. We’re excited and gratified to see our international partners leaning into the long-term growth of regional biosecurity ecosystems.
It is a great complement to the work we’re doing in the U.S. here with the U.S. CDC. So we do have a program in airports here that’s similar to what we do in Doha. These initiatives demonstrate the growing momentum in the new market for global biosecurity and the huge potential for rapid progress in building resilience to biological threats. And this is going to be important if you want your kids in high school designing flowers. And we want that world to happen safely. We need to also build out biosecurity. And that’s why these two parts of Ginkgo’s business are so complementary. So, okay, in summary, I’m really excited about the great work we’ve done in 2023 and what we’ve already accomplished in 2024 thus far, especially the work that was done to enable Ginkgo to be a one stop shop for all types of biotech infrastructure services.
I couldn’t be more excited for the year to come as we continue to expand our offerings and growth of the business. All right, now I’ll hand it back to Megan for Q&A.
