AI + Synthetic Biology: The Most Transformative Technology in Human History | Ben Lamm (Colossal)
Is it worth your time?
You get a detailed account of how a biotechnology company applies automated genetic engineering to commercialize products for environmental restoration and human medicine.
This conversation examines how artificial intelligence and synthetic biology intersect to address environmental degradation, mass extinction, and biological manufacturing. The discussion details the operational pipeline required to scale genetic editing and apply it across ecological and medical domains.
Synthetic biology and AI. Lamm argues that combining artificial intelligence with synthetic biology creates a powerful platform for engineering living products. He reports that his team has scaled from executing a few genetic edits at 40 percent efficiency to performing hundreds of edits at 90 percent efficiency.
Plastic degradation microbes. Lamm explains that the biological systems designed to bring back extinct species can also produce microbes that break the chemical bonds of plastics. He points out that humans currently absorb 5 grams of microplastics into their brains, making targeted enzymatic solutions urgent.
Artificial wombs and IVF. Lamm outlines the development of artificial wombs to productionize species development and bypass long gestations or genetic bottlenecks. He notes that the underlying hydrogel and microfluidics technology also improves human IVF by allowing embryos to grow healthier than standard morphological grading permits.
Invasive species and gene drives. Lamm warns that invasive species create a global economic problem valued at $5.4 trillion. He argues that engineered gene drives can humanely reduce destructive populations like screwworms and invasive cats over time.
Colossal’s Vision on De-Extinction and Rewilding
The idea of bringing back the woolly mammoth has been around for a long time and had been played around with in nonprofits. I met George Church, one of the greatest synthetic biologists, CRISPR gene editor, entrepreneur, and professor at Harvard Medical School. I asked him if he could work on one project for the rest of time with unlimited capital, and he instantly replied that he would work to bring back mammoths, rewild them into ecosystems, and build technology applicable to saving species and human health care. He didn't hesitate.
I was going to fund it as a side project at first, but then I thought it was just really interesting. Colossal is a parent company spinning out a dozen companies with massive potential. Our first biological products company, Breaking, focuses on plastic degradation. The same system that can bring back a mammoth can also make microbes that break the chemical bonds of plastic.
Every company should be or is an AI company; without AI, we would not be able to do anything we're doing. Ben is building the ability inside Colossal to design using AI and then build living products. That's a moonshot.
He loves woolly mammoths and wants Jurassic Park, which is the number one request we get. Megalodons are second, which is also weird, but they are just cool. They littered their teeth all over the ocean floor and are still scary—Jaws times 100.
You built a company that went from zero to a $10 billion valuation in four years. As the CEO, you built an extraordinary team.
We have 260 scientists, 200 in the US and 60 in Australia. While we don't always lead with AI, we would not be able to do anything we're doing without it. Ben is building a platform and engine inside Colossal for designing and creating living products using AI.
To build this end-to-end pipeline for synthetic biology, we must develop technologies across computational biology, cellular engineering, genetic engineering, cloning, somatic cell nuclear transfer, and eventually artificial wombs. By utilizing genotype to phenotype relationships, ancestral state reconstructions, and comparative genomics, we can build a system model applicable to all types of biological products. This is critical because we are facing a massive extinction crisis.
Colossal is a parent company spinning out a dozen companies with massive potential. Our first spin-out was Breaking, a plastic degradation company. The same system that can bring back a mammoth can also make microbes that break the chemical bonds of plastics. This is vital since we have 5g of plastic in our brains—the size of a credit card or plastic teaspoon—with 90% absorbed through the gut.
We originally thought an enzyme from a single microbe had been discovered. After taking this discovery from the Wyss Institute and analyzing it at Colossal, we found it is actually a consortia of microbes working together. This was even better, as we could understand the enzymes being made and edit each microbe to create different variants of enzymes to hit different types of plastics.
Advancements in Plastic Degradation Technologies
Most plastic treatment and degradation companies just make smaller plastics, but this discovery actually breaks the chemical bonds of the plastic.
Using directed evolution, our pipeline, and editing tools, we supercharged it so it can break down a larger breadth of plastics at a much faster rate per surface area.
Reintroducing Extinct Species and Bio Vaults
Imagine a supplement that breaks down the bonds of microplastics in your gut before they are absorbed. This is not just an environmental problem; it is in our food supply, our reproductive tissues, and in some cases, crosses the blood-brain barrier.
Publicly, we've announced the woolly mammoth, the Tasmanian tiger, the dodo, and the moa. We have also made dire wolves, with more coming that we haven't announced. Here are the woolly mammoth mice, the woolly mammoth, and the dire wolves, Romulus in the front and Remus in the back. We did a Zoom with George R.R. Martin, who wrote A Song of Ice and Fire, which became Game of Thrones and popularized dire wolves. Most people, including some members of the Game of Thrones cast I won't call out, thought dire wolves were just mythical creatures. It was pretty cool to show that we could take a 73,000-year-old skull and make puppies in 18 months, which is remarkable.
We are looking at educational and STEM-related content, as well as ancillary effects. About 12.5 percent of global consumers buy something extinct every year, which turns out to be 1.7 trillion net new dollars. Part of our de-extinction model is to subsidize the platform and help countries preserve their species, which is a lucrative business model and provides phenomenal educational feedback.
We announced the world's first bio vault. After building reference genomes for every species we work on, we decided this should be a global project and partnered with the UAE as our first partner. The region has incredibly diverse fauna, much of which is going extinct. We need to sequence it, build digital backups, and ensure they are shared with the global scientific community, subsidized by governments. Over a year, we educated them on the importance of biodiversity for national pride, technology, and the data from these animals.
Artificial Wombs and Species Production
Even if you don't like animals, you should support this because the applications could be helpful to humans. Fundamentally, we got them to agree to put hundreds of millions of dollars into the world's first bio vaults. If you're going to spend X dollars, spend X plus Y and wrap educational content around it. This is a nine-figure initiative for us and for the country, building capabilities for countries to protect their biodiversity in a completely new way while sharing data globally. I think of countries as your customer for saving their endangered species.
Besides biobanking, we are giving birth—pun intended—to an artificial womb company so mammals and birds can give birth ex utero. We have three mini moonshots involving artificial wombs for different animal clades. Our vision is that by using biobanking, synthetic biology, automation, robotic process automation with AI and computer vision, and artificial wombs, we could productionize species development.
When you have a genetic bottleneck or long gestations, like with the northern white rhino, this is vital. There are two females left; they are functionally extinct because they and their 18 embryos are related. If you can engineer genetic diversity synthetically and from lost specimens, then productionize it through artificial wombs, you could use a piece of the $25 million spent annually keeping two animals alive to productionize the species, and the rest could go to water or education for the country. Productionizing endangered species and helping them adapt at the same curve we are changing environments will be needed because evolution is not fast unless it's directed.
We also purchased the world's top two cloning companies. Most people think of celebrity cloning—we did clone Tom Brady's dog, and he is an investor. But while only 18 species have ever been cloned, 15 were cloned by ViaGen, the main company we bought. Most cloning efficiencies are only about 2%, but ViaGen's was consistently 78%. The only endangered species ever cloned on the planet, such as the black-footed ferret, were cloned by ViaGen, which took old cells to reanimate and clone them. We are now taking those technologies into countries to productionize the cloning of critically endangered species. I want you to think about the platform Colossal is using to leverage AI and synthetic biology to specify the desired phenotype, genes, and gene copies.
Using AI to Engineer Desired Phenotypes
We want to be able to understand how to make the snout or teeth longer, using AI to change which genes and enhancer sequences allow us to design the living animal we want.
When I said we could probably create Pikachu, that became the first question I was asked at South by Southwest a few years ago. The rest of the entire panel was about Pokémon, though I had been hopeful to talk about our vaccine development for elephants and others.
Invasive Species Control and Gene Drives
It is fascinating that if you can engineer life this way, you can build companies to create disease-resistant or drought-resistant plants. We are spending a lot of time on this because extinction rates are not just human-caused, but some are supercharged by humans, such as diseases.
I am very passionate about our project on chytrid, the leading extinction driver for frogs and amphibians. We can solve the current problem and create chytrid-resistant frogs, salamanders, and other amphibians, which has huge applications. After dinosaurs, corals, and perhaps Pokémon, we aren't working on dragons, Pokémon, or corals yet. We now have an entire group focusing on the biggest issues, from livestock to critically endangered species.
I believe synthetic biology, especially paired with AI, will be the most transformative technology humanity ever has. This market is already worth hundreds of billions today, though it is not well tracked. Consider when a swine flu or bird flu wipes out a population.
When we brought back the dire wolves, some feedback suggested people dislike wolves because they kill cattle, but the way we raise cattle means wolves won't go near them. People also feared GMO corn would change your genome, but as I discussed with Rob on Zoom today, GMOs have saved so many lives.
There was a season when seat belts were scary, and car manufacturers argued against them because they feared it would make people think cars were dangerous.
When we met with the Australian government about reintroducing the Tasmanian tiger, we found that their law technically classifies them as GMOs, even if they are 100% genetically identical, as an amalgamation of 53 different tigers over 300 years. We had to educate the government that an anti-GMO narrative prevents the rewilding of this incredible species because they see it as what they feared in the '80s.
There is another multi-trillion dollar market involving gene drives, which is a 40% technology problem and a 60% marketing problem. The invasive species problem is a global issue currently measured at $5.4 trillion. As the world gets smaller from a commerce perspective, invasive species become more prevalent, from cane toads, cats, and carps in Australia to mosquitoes. In the US, Texas has declared a national emergency over the screwworm coming through Honduras and Mexico into southern Texas, which will decimate our cattle and bison industry.
You can create vaccines for the animals, but that requires working through the USDA and facing a persisting anti-GMO movement. Separately, you could create genetically modified screwworms and release them so the next generation are all male. In New Zealand, Australia, and parts of Africa, people are killing invasive cats and possums because they decimate local populations of small mammals, super meals in Australia, or birds in New Zealand. If you engineer the right gene drives and bio control, animals and insects can live out their normal lives while the population decreases humanely over time. People released gene drives around mosquitoes in Africa, but everyone freaked out and stopped it because they thought it was bad.
Future of Genetic Engineering, Biodiversity & Synthetic Biology Capabilities
We're working with our government and international governments. That's the magic of AI combined with synthetic biology. Two years ago, I wouldn't have worked on bio control, bio containment, and gene drives because I thought they didn't work when distributed around mosquitoes. Now it's a $5 trillion problem.
We have an interesting model and proprietary technologies that make it safer than anything ever dispersed in the wild, and we have the ability to roll it back. In the US, the cost is over $500 billion a year. I don't know the exact spend against it, but it's mostly combatted with archaic methods, like poisoning the environment to get rid of invasive species—similar to archaic ways of treating cancer. For larger animals, they are killing, poisoning, and trapping them.
The good news is that people are focusing on compute power, AI, and anthropomorphic robots. I think 99% of people fluent in synthetic biology and genome engineering focus only on human healthcare. While that's great, applying the same technologies to other use cases offers even larger economic opportunities and a bigger chance to help us.
Regarding creating healthier embryos, reinventing IVF, and advanced gene editing technologies: our artificial wombs don't work yet, and we have three projects around that.
We went through the IVF process, which is emotional and archaic. You look at something so precious through a morphological grading scale to choose embryos, which is archaic compared to where technology currently is.
For us to be successful long-term with our mammalian-based artificial wombs, we have to innovate in a couple of categories, as there are four core but nine different core placental types. One of those is keeping embryos healthier longer. Modern-day IVF clinics have worked the same way for a long time, basing decisions on morphological grades. We found that in some model species like mice, and in non-model species, embryos that wouldn't be the winner of the race morphologically at the stage where most humans make their decision actually become the healthiest if you go a little bit longer.
Humans are making one of the most important decisions of their lives using an archaic system based on one moment in time and bad imaging. To succeed with the artificial womb, we built a hydrogel and microfluidics device that makes embryos healthier. We have taken embryos in mice and non-model species much further than anyone else in the world. That same technology could be applied to embryos using a different grading scale that is proving more efficient and accurate in both model and non-model species. That innovation could be transformative to IVF, and as far as I know, no other company is close to what you have built in terms of production pipeline.
Focus and Future Directions of Colossal
Two or three years ago, we were doing victory laps for a couple of edits at 40% efficiency, which we thought was good since most people were doing 15%. Now, we are doing hundreds of edits at 90% efficiency, and in the coming years, that will be thousands. These are not linear repeats; they are all over the genome and so precise that I would feel comfortable applying this technology to human healthcare.
While we are myopically focused on biodiversity and de-extinction, we thought we were the best by every standard even two years ago. We have already surpassed the largest delivery by 5x and expect to be at 20x before the end of this year. Our DNA synthesis large heart delivery and clustering models are superior because we have taken a product and systems model approach to synthetic biology leveraging AI, whereas most people seek one-off point solutions for human healthcare.
Understanding genotype to phenotype expression across the genome presents a different set of challenges. We believe de-extinction and species preservation is a $10 trillion opportunity and one of the most important things to focus on because it solves the hardest things in biology. If we make a great discovery on plastics, we invent it here, spin it out, and assign incredible women and men to work on it through intercompany agreements where we share our developed editing efficiencies.