The second customer case study is around production rather than optimizing that enzyme. These are a customer that wanted to figure out how to produce a small molecule compound at higher titers, basically, like how much of it you get out of the fermentation putting in this big tank. And they had a goal here to do this over the next 3 to 5 years, which will be important in a second. And so here, we were able to deliver a better outcome than what the customer asked for. Originally, they just wanted us to try a couple of different host strains, and we did it faster and cheaper than they wanted. So on our first experiment, we were able to improve the titer by 12.5-fold. This is what the customer wanted us to get done by the end of the first year, and we did that in the first experiment, and it was building on knowledge that we already had in our code base.
So this is less about the AI and more of that we had genetic elements and we knew which ones worked well in this organism, and we could just take them off the shelf. Okay. That’s very powerful, again, strength of scale. Because we have that big code base, we could take something off the shelf that another company that didn’t have that wouldn’t have on the shelf. And then by the end of the program, which only took us 10 months instead of 3 to 5 years, we were able to deliver 50-fold titer improvements, which is almost double what the customer was originally working towards. The second round of improvement was driven by machine learning and driven enzyme improvement similar to the last case study. And again, this is something that makes Ginkgo unique.
We draw on a wide range of tools. So in this case, genetic elements off the shelf, as well as protein design enabled by AI, putting those together, gave that outsized outcome for the customer. So these are a couple of examples of how we’re integrating our AI tools into customer programs. I cannot emphasize we’re just getting started with this. I’m super excited about AI’s ability, like I said, in that equation, to improve efficiency of all those different steps, hopefully reducing cost and speeding time lines for customers. Okay. Second topic. So let’s take a look at the rest of the pipeline of customer programs. So again, as a reminder, I want to be very clear about this. Ginkgo does not have its own product pipeline. Like I do not have my own drug development pipeline here at Ginkgo.
So when I’m saying pipeline, this is a pipeline of customer programs. In other words, we had to negotiate and sign up a customer and get them to outsource this work for us — to us to get on this program list. In Q3, we had our highest number of active programs on our platform of all time, with pharma making up the largest percentage of those programs. But an additional piece of color I want to give you today is where those programs stand in terms of their maturity. In other words, how are they progressing through the technical work? All right. And again, I’m happy about that pharmaship. I don’t want to undersell it. I do think that’s really important, especially as industrial biotech has gotten tighter. It’s really great to see that. That is, again, a real strength of being a platform, right?
If a certain area gets thinner, we can move to areas that have more demand. As long as biotech in general is moving, we’ve got something to do. Okay. So this is a fun chart to me. As a traditional product-based biotech company would often show a pipeline like this for maybe 5 or 10 drug assets that are sort of moving through preclinical and clinical trials. Here, we have so many of these programs that we can’t fit them on 1 slide. This page is just a programs that are over 50% complete. And we’ll show the rest on the next slide. This is the point I was making earlier of why I’m still excited to see 21 programs being added in the quarter, even if it was less than we were hoping. It’s just a really great amount of scale on a relative basis in the biotech industry.
And to give you an overview of the chart, each horizontal bar here represents a program, and the dark portion of the bar, like on the right-hand side, represents the progress made on that program year-to-date as a portion of the total program. And I cannot tell you the number of times we get asked for this. So I’m very happy to be sharing it with all of you. We will try to do this again, if people ask for things, we try to clean it up and get it out there is a good example. So as you can see at the top, there are a number of programs that are at 100%, okay? So that means that Ginkgo’s program were concluded on that program in Q3. So we do this again for Q4, it would be gone, right? And on the next slide, you can start to see some of the shift in program mix.
