Biotech ≠ tech: Why venture capital needs a new playbook for the therapeutics era

Silicon Valley investors have been conditioned by two decades of software: launch quickly, capture users faster, and hit the magic $1 million ARR mark within about three years,¹ the median journey for a SaaS start up today. The pattern is comforting because the inputs are clear: hosting bills are trivial, distribution is digital, and failure shows up early and cheaply.

Yet, the same template disintegrates when it meets a new molecule. In 2024 the average Big Pharma programme cost $2.23 billion² to shepherd from discovery to market, and industry analysts logged $7.7 billion of spend on candidates that were simply terminated last year.² Even after a drug survives a decade of trials, its historical likelihood of approval from Phase I remains a stubborn 7.9 percent.³ You can model ARR all day, but current drug development still carries a strong dose of biological roulette.

How we arrived at the one model monoculture

Ever since Genentech’s recombinant-insulin breakthrough in 1978, proof that technology could bend biology into a commercial product, biotech has chased one dominant business model: licensing. For most of the modern era, that made the most mathematical sense. A company pushed an asset through costly pre clinical and early clinical hoops, then sold rights to a pharma partner that could finish the marathon and write an eight figure upfront check. Because a single approved therapy can out earn a mid tier cloud platform, the product’s sheer value seemed to forbid any softer approach: why would you “rent” a molecule the way you rent CRM software?

Investors internalised that logic. Outside a handful of specialist life sciences funds, most mainstream VCs waited until a programme had actually entered Phase I human trials, by which time the science had already been financed mostly inside academic laboratories. Only when a molecule looked potent, patentable, and publishable did its inventors spin it out. The start-up thus functioned as a late stage derisking vehicle: data first, business second.

Data, algorithms, and the chimera that followed

Over the past two decades, two forces have scrambled that neat picture. First, data has exploded; genomic sequencing, single cell assays, high throughput screens, at rates that outstrip even Moore’s Law.⁴ Second, machine learning systems thrive whenever data is abundant, compressing wet lab timelines and spitting out candidates at industrial speed. The result is a new generation of start ups that look, from a distance, almost tech native: they model proteins on GPUs, deploy code daily, and talk about ‘platforms’. The global AI drug discovery market is already growing above 30% per year and will likely pass $5 billion before the decade ends.⁵

Yet, no amount of computation can override pharmacology (yet); every candidate must still prove its safety and efficacy in living systems, and there is no instant customer base waiting on the other side. Drug biotech is now a chimera, adjusting to a shifting ratio of software sprint and clinical marathon. Venture capital must treat it accordingly.

Three business archetypes, one stubborn reality

The post genomic biotech landscape now clusters around three primary business models, each molded by the convergence of data, computation, and automation, yet still constrained by biology’s hard rules.

1. SaaS like tool providers.

These firms sell subscriptions to modelling suites or data portals; revenue is recognisably recurring, but the upside is capped by market size (which is not counted in the millions), and every licence still demands deep scientific support.

Schrödinger would be the best exemplification of such an approach by providing predictive molecular modelling software on a subscription basis, though Schrödinger itself is actively diversifying into other business models including direct drug development.

2. Project based collaborations.

Platform companies run discovery projects for pharma, collecting research fees and, if lucky, downstream milestones. Cash arrives earlier than in pure licensing, but scaling is tethered to scientists, not servers; marginal costs do not vanish.

Exscientia and CytoReason are good examples of leading AI-driven drug discovery firms, frequently partnering with large pharmaceutical companies, leveraging AI to generate drug discovery insights and securing project-based revenues.

3. Traditional licensing pipelines.

Here lie the moon shots and the billion dollar exits, along with the longest timelines and the harshest failures. Early stage VCs recoil, as their money may disappear into a statistical funnel where nine of ten programmes die. AI-driven drug discovery companies like Atomwise exemplify this model, licensing drug candidates to pharma after extensive computational and experimental validation.

The SaaS dream stumbles on the value paradox: a curative therapy might be priced at $2 million per patient, so no CFO will “pay by the month”. Collaboration models may often deter generalist investors, as head count rises line by line with revenue. And licensing returns, while spectacular in theory, require patience most funds were never structured to provide.

What venture capital should concede

In software, the greatest asset is talent: world-class engineers can pivot a weak feature into a winning product, so investors often back people more than metrics. Traditional therapeutics invert that logic, the data is the asset, because no amount of charisma can rescue a molecule that fails toxicology. The new generation of AI-native biotech companies straddles both worlds, asking investors to trust exceptional talent and early signals, rather than mature data. Whether that talent premium can someday outweigh biological read-outs is an open, and fascinating, experiment.

If tech trained VCs want exposure to the therapeutic revolution, they may need to rethink their risk calculus:

• Time horizons that reflect the extended development cycles inherent in any biological model, rather than the shorter spans expected in tech.
• Milestone tranches aligned with each model’s natural inflection points, whether early‐revenue signals for SaaS-like tool providers, pilot milestones for project-based collaborations, or clinical‐stage achievements for licensing pipelines, instead of rigid calendar runways.
• Carry structures that share in royalties or sales, because IPO windows will not always oblige.
• Deep scientific expertise must be embedded in the investment team, as interpreting early signals in biotech may depend on deep understandings of biology, something that goes beyond traditional SaaS analogies.
• The business model mutates with the molecule, today’s service platform can pivot to asset-licensing (or vice-versa) as data and cash needs shift; capital must be ready to follow.

None of these adaptations undermine venture economics; they simply map capital deployment to biological reality.

What biotech founders must offer in return

The burden of adaptation is not one sided. Drug start-ups have to derisk earlier and speak a language investors grasp:

• Early read outs, even at an exploratory stage, that prove the engine works.
• Paid pilot projects or low friction collaborations that turn data into revenue within the first 18 months.
• Transparent metrics on throughput, hit rates, or cost per lead, shared as openly as SaaS companies report customer retention, making performance easy to evaluate.
• Support diverse, and not strictly biotech-standard, revenue channels: equip the venture with multiple monetisation options, so you can pivot as data, stage, or cash needs shift, while still giving investors familiar, lower-risk paths to value.

Some companies offer demos for free, others charge small fees, but all are forced to show meaningful early signals, whether scientific, technical, or commercial, well before any ultimate proof point. These signals are what connect biotech’s slower cycles to tech’s faster expectations.

A truce worth striking

Biotech will never scale like cloud software, and cloud software will never heal a sick child, yet, the space between them is dissolving into fertile common ground. Investors who accept that SaaS laws stop at the cell membrane, and founders who demonstrate that their biology can be engineered with the efficiency of technology, will meet in the middle.

Together, they can build companies that are not only financially rewarding but truly transformative for human health.