Steven
Banerjee

Steven Banerjee arrived in Silicon Valley the way most ambitious immigrants do - with a degree, a plan, and an instinct that the plan was already outdated. The degree was in mechanical engineering from New Zealand's University of Canterbury. The plan was semiconductor research. What he found instead was a gap so obvious he couldn't unsee it: the most advanced field in modern science was running on chaos.

Life sciences researchers in the early 2010s had access to more raw data than ever. PubMed, ChEMBL, Ensembl, Google Scholar - each a siloed library, none of them talking to the others. Scientists were duct-taping workflows together with email, Slack, spreadsheets, and institutional memory. Drug discovery was costing $2.6 billion per approved therapy, taking 10-15 years, and failing more than 96% of the time. Banerjee saw a software problem dressed up as a science problem.

But he didn't build software first. He went deeper. He became a Doctoral Fellow at IBM Research Labs in San Jose, studying semiconductor technologies with applications in biotech. He wrote federal grants alongside Ron Davis - the Stanford pioneer who led the first complete sequencing of a eukaryotic genome. He conducted advanced research as a Visiting Scholar at UC Berkeley. By the time he founded his first company, he had worked at the intersection of physics, biology, and computer science long enough to know where each discipline broke down.

In 2017, Banerjee founded Mekonos, a Silicon Valley biotech company built on an insight he'd carried from his semiconductor research days. Cell and gene therapies were transforming medicine - but the delivery problem, the question of how to get biological payloads into specific cells reliably and safely, remained largely unsolved.

Mekonos set out to answer that with platform technology borrowed from the semiconductor world. The company raised over $40 million from top-tier investors and pioneered novel platforms in cell engineering and gene editing. Then it was acquired.

Most founders take a victory lap after an exit. Banerjee took notes. He'd spent years watching brilliant scientists lose months to information scatter - the same study rediscovered twice, the same failed pathway explored by three independent teams, the same data point buried three papers deep in a PubMed search nobody had time to finish. The problem wasn't the science. The problem was the infrastructure around it.

Those numbers are not abstract to Banerjee. They're the measurement of a system that has never been properly organized. They're the reason he started over in late 2020, building something harder to explain than a biotech company but potentially more consequential: a connected intelligence layer for the entire field of life sciences research.

The premise of Nextnet is deceptively simple: the world's biomedical knowledge exists, but it is not connected. PubMed has 35 million citations. Ensembl maps the genome. ChEMBL catalogs chemical compounds. Google Scholar indexes tens of millions more papers. None of them know what the others contain. A researcher studying a drug target has to cross-reference them manually, which means they almost never do it completely.

Banerjee's solution was to build an ontology - a purpose-built semantic structure that understands biological entities and the causal relationships between them. Not a database. Not a search index. An actual map of how genes connect to diseases, how drugs interact with pathways, how authors build on each other's work across decades and disciplines. Within the platform's first year, this map contained nearly 100 million machine-curated relationships.

On top of that map, Nextnet runs two tools. Nextnet Copilot delivers precise, evidence-backed answers grounded in verified scientific sources, without the hallucinations that make general AI tools unusable in research contexts. Nextnet Explorer visualizes the connections - a dynamic workspace where researchers can map ideas, find hidden links between genes and diseases, trace drug-target relationships, and share their work with collaborators.

The launch of Nextnet's MVP in early 2024 produced what most founders only dream about: organic growth. No paid acquisition campaign. No splashy launch event. Just researchers finding the product, using it, and telling other researchers. Within a year, that referral chain had produced 1,200% growth. The platform is now active in more than 100 countries. Harvard, MIT, UCSF, MD Anderson, Columbia, UC San Diego - all on the platform, none of them through enterprise sales.

"Software hasn't eaten life sciences," Banerjee has said. The field missed the productivity revolutions that transformed finance, legal, logistics, and even medicine's administrative layer. The scientists doing the work that actually generates new treatments are still, in many cases, running on 1990s-era research infrastructure. Nextnet is Banerjee's argument that this is a solvable problem, not a permanent condition.

Banerjee's stated goal is to democratize access to biomedical knowledge - to make it possible for any scientist, anywhere in the world, to work with the same completeness of information as the best-resourced research team on earth. That's a large claim. The architecture of Nextnet - a semantic web, not a database - is the mechanism he's betting on to make it real.

The platform's RAG (retrieval-augmented generation) approach with guardrails is a direct response to the hallucination problem that makes general AI tools unreliable for science. The ontology's 100 million relationships represent years of curation work that can't be replicated quickly. The 1,200% growth rate without paid acquisition suggests researchers are spreading it through word of mouth - the surest signal that it's actually useful, not just impressive in a demo.

Banerjee also serves as an advisor and board member to multiple Silicon Valley technology startups, carrying forward the pattern-recognition he's built across two companies and four very different research environments. His media presence extends across the US, Europe, New Zealand, Japan, and China - a reach that reflects both the global nature of scientific research and his own unusual trajectory from a Canterbury engineering program to the center of Silicon Valley's most ambitious bets on life sciences.