Where biology and tech unite: Atomico’s take on building a Biology 2.0 company
As we are starting this new decade with many global challenges ahead of us, a better, healthier, more sustainable future can be hard to envision. A future where the products we consume — whether it’s food, material, or fuel — are created without exhausting our land, forests and oceans. A future where our medicines are generated to fit individual biology, or where health decisions are based on a complete biological picture and knowledge systems that constantly learn. A future where our processors are built on biology not the energy-hungry silicon used today.
However, that future may not all be science fiction, as one of our oldest ‘tools’, biology, is being transformed through technology. We at Atomico believe this represents one of the greatest opportunities of our century.
Biology and engineering have historically been two worlds apart. The traditional approach to understanding and using biology has been to try to link the (few) known parts of this complex puzzle, derive a limited set of hypotheses and rules, then test those through trial-and-error exploration and manual experiments. Breakthroughs in biology have been limited by our incomplete understanding of many of the other pieces of the puzzle and the links between them, by the quality and quantity of data, by the speed and scale at which we could run experiments.
Consequently, research productivity has declined despite more money going into the field. We spend $150 billion on drug R&D, yet the costs to develop a drug double every decade (as summarised by Eroom law): it now takes $2.5B and over 10 years to bring a drug to market. 60% of our science findings are not reproducible, and this is estimated to cost $28B in US only. The materials we are using are decades old, petroleum based.
The past years have seen a change emerging. As technology is entering this new frontier, biology is starting to shift towards a computation-driven, rule-free “search” and towards engineered, generative design, enabled by better and more integrated data, and more automated experimentation.
We call this new field Biology 2.0: applying engineering and computer science principles and tools to biological problems, bringing the physical and digital worlds together. Biology 2.0 is understood, programmable and replicable.
We believe we are at the beginning of a deep shift that will change multiple trillion-dollar industries, and create unique opportunities for entrepreneurs, investors, and society as a whole.
There was never a better time
Over the past years, several developments have enabled the convergence between biology and engineering, and enhanced our ability to explore biology at scale through technology.
‘DIGITISATION’: We can ‘read’ and ‘write’ our biological code cheaper and faster. The cost of genetic sequencing has dropped by 10-fold in the past five years allowing us to “read” DNA, our own four base code, at a much higher rate. There were over 500,000 genomes sequenced in 2017 vs. less than 10,000 genomes in 2012. We also made significant progress in our ability to “write” DNA, whether it’s DNA synthesis, or gene editing.
DATA & COMPUTING: We have better processing power and tools, underpinned by larger data sets. Alongside sequencing, other tools such as sensors, wearables, or high throughput lab assays have allowed us to generate multidimensional data series at higher scale and speed. Machine learning, enabled by advances in computation power and data, is achieving capabilities not thought possible before, with one well known example in AlphaGoZero proving the ability to extract principles without human knowledge and to become self-taught. Our friends at Obvious have written more about this here.
LAB INFRASTRUCTURE: We can scale and automate more of the manual research work. Lab work hasn’t changed much in this century, but in the past years, higher throughput equipment, robotics, precision automation and software tools in the research lab are making an impact, improving speed of experimentation, reproducibility, and accuracy.
TALENT: The science and engineering talent base with ambition to build Biology 2.0 businesses is increasing. With increasing accessibility and decreasing cost, biological work can now happen outside big academic or corporate labs. It is easier and cheaper to start a business in this space, and a viable career option for the new generation of PhDs.
Biology 2.0 is already here
Biology 2.0 creates multiple sectors with applications across fundamental industries. Below are some non-exhaustive examples where we’ve seen new companies created in the past years:
- Computation-driven drug R&D: a new, hybrid (tech and bio) type of companies employing computation to better search the solution space for drug discovery, and create “asset” engines more efficient than traditional pharma (e.g. our portfolio company Healx, or Recursion Pharma, both of which already advanced assets in the clinic at a fraction of time and cost to pharma). On the “D” side: AI/ML-enabled integration and analysis of rich, heterogenous real-world data while preserving privacy, to improve outcomes and speed of clinical development through predictions and new biomarkers (e.g. Owkin)
- Synthetic Biology :“writing DNA” to design new biological parts or products, or redesign existing ones, such as foods (e.g. our portfolio company Memphis Meats), protein-based drugs (e.g. LabGenius), microbes and organisms for industrial processes (e.g. Zymergen, Gingko Bio), advanced materials (e.g. Modern Meadow).
- AI/ML for diagnostics, monitoring and prediction: image recognition in radiology and pathology, where AI/ML has shown it can improve efficacy, speed and cost of diagnosis (such as our portfolio company Kheiron Medical in breast cancer, or Viz.ai in stroke, Caption Health and Ultromics for ultrasound); non-invasive alternatives for disease tracking and screening like liquid biopsy (e.g. Freenome, Grail); new biomarkers to detect disease early, non-invasively, and more consistently such as saliva, voice and smell; and longitudinal “omnics” approaches which in the long run can enable complex predictions and guide early interventions (e.g. Q.bio)
- Next-gen medical hardware & robotics: increasing miniaturization, connectivity and software intelligence creating a new wave of devices that are more portable, flexible, easier to train on and use, and often cheaper (e.g. Auris Health in endoscopy, Butterfly Network in ultrasound, CMR Surgical in surgery), or micro/nanorobots for applications like drug delivery.
- Microbiome: computational platforms to develop microbiome-interventions to fight disease (e.g. Pendulum); proprietary data collection and analysis tools at scale to advance understanding (e.g. Biomesense); consumer propositions with medical back-end (e.g. Zoe for precision nutrition, Prime Discoveries)
- Research infrastructure: higher performance tools for capture and analysis of biological information (e.g. ONI); software for biological analysis and research (e.g Benchling, Synthace); cloud-based automated labs (e.g. Radix Labs, Arctoris)
- Neurotechnology: brain machine interfaces integrating hardware and software, from non-invasive, consumer-focused (e.g. CTRL-Labs, Halo Neuroscience) to invasive medical applications (e.g Paradromics, Bios Health)
- Biocomputation: using biological systems to create new computational infrastructure, such as encoding digital information into DNA for more efficient data storage (e.g. Catalog DNA)
Building a Biology 2.0 business is not a classical tech play
Biology 2.0 companies are a new kind. The classical tech playbooks, KPIs, frameworks, and standard milestones may not always apply. The revenue, operating and commercial models are different, though current industry and financing structures haven’t always kept up.
We are still learning how to best build and support a new generation of companies that will take advantage of these possibilities, but we laid out below considerations for some of the most critical decisions founders make when building in this space.
- Monetisation models
Many Biology 2.0 companies start with a service-based revenue model, with the intention to translate that into a SaaS-like, license-based model. Easier said than done. Getting a small service contract at a business unit level does not easily and quickly translate into licence revenue at a procurement-driven, enterprise level. In addition, a SaaS model is not always the best way to capture the value from the downstream IP that the respective Biology 2.0 company helps create for the incumbent.
We favour monetisation models where the value is owned in house (“full stack”), or where there is a value-share in the end outcome. Examples would be a royalty and milestone-based model in an asset-based co-development partnerships or risk-sharing agreements based on cost savings. It helps if the end product is high value, and if there is already a market for monetizing and trading the value/IP, such as in biopharma. The challenge is these models often come with discontinuous value creation curves that may not align well with the 18 to 24-month financing windows of traditional tech investing. These models also have a different profile to traditional tech in that the value increases (often significantly) at a specific milestone vs. gradually. However, we do recognise that for “picks and shovels” infrastructure businesses a SaaS like model can be more suitable. Our friends at Air Street Capital have also written on this.
2. Transition from tech to product to sales
Product development and early commercialisation of Biology 2.0 technologies often require a deep understanding of the ecosystem and the incentives — it’s easier to build a product that is aligned with, not against, many of these incentives (we often ask “who will this product upset?”). It also requires collaboration and partnerships with traditional industries, with long business development cycles — navigating these is an expertise in and of itself, reliant on a network into these incumbents, and we advise teams to get this on board early.
While it is possible that in the long run data becomes increasingly commoditised, today access to a sufficiently large, high-quality, hard-to-get dataset for a specific type of problem can still provide a competitive advantage. Detailed subpopulation-level data, multi-parameter longitudinal home-to-hospital observational data, or ethnically diverse data are some examples. Privacy concerns can also restrict using data off-premise and impact companies’ ability to pool data from multiple settings. We like to see businesses that find a creative way to access “hard-to-get” data early and quickly, whether it’s through partnership, proprietary data collection and generation, or federated learning.
4. Regulatory strategy
Regulatory hurdles are a reality in this space. Getting good quality advice is never too early as there are often multiple options for a regulatory approval strategy, and the choice will impact data and product roadmaps. In some cases, there may even be options to build smaller “non clinical” products that don’t require an approval, while developing the more regulated ones; this enables founders to start developing customer relationships earlier in the company’s life, and to put the product in the hands of users to test and learn early.
5. Team construct and culture
At the core of Biology 2.0 businesses are truly interdisciplinary teams — from biology, chemistry, regulatory, data science, computer science, hardware engineering, as well as operational and commercial hires from within traditional industries. Ensuring that these teams speak the same language and continue to operate in sync as the company scales is critical for long term success, but not trivial to achieve. Success requires management to ensure all staff have an equal voice around the table, and build an operational team structure and a culture that reinforces the value of cross-discipline collaboration.
6. Long term capital journey
Many Biology 2.0 companies will need significant capital early in their journeys, before they can generate any revenue. It is helpful to take this long term view into account when planning milestones and thinking about valuation, to ensure that these support future raises. In addition, we encourage founders to develop a network of later stage investors early on and getting them ‘ready’ well in advance, especially when building outside traditional hubs like Silicon Valley.
7. Conscious scaling and PR
Biology 2.0 inherently touches highly-sensitive areas, such as the genetic engineering of foods, materials, and humans, or the use of machine intelligence in high-stakes tasks like diagnosing and treating disease. The technical nature of the businesses also make them prone to exaggeration by the media and the general public, misunderstanding and thus the risk of public misunderstanding or scepticism. It is important that founders own the messaging early on even if the product might not be ready for market for a long time and invest in making sure their story and the potential of the technology is told to general audiences in an accurate way. We also work with founders to think about the potential unintended consequences that their businesses might have on society and the environment.
We are only just beginning
We are only just beginning to comprehend the possibilities presented by Biology 2.0. We will likely see some hype and some unfulfilled promises on the way, but Biology 2.0 is one of our most promising tools to address critical issues, such as sustainability and health, and in doing that, create outsized value.
At the start of this new decade, we are embarking on a quest for solutions that make our industries more sustainable, our lives healthier, and in fact, our own human potential greater. When biology and technology meet, when brains from across these disciplines collaborate, supported by improved access to capital, talent and tools, we expect to see radical innovation and new definitions of what is possible.
If you are an ambitious founder laying the foundations of a biological future, I would love to hear from you at firstname.lastname@example.org.