Key Takeaways

  • Warp Sciences raised $215 million in a round led by Lightspeed, with backing from Lux Capital, Menlo Ventures, IVP, Mayfield, and Scribble Ventures.
  • The pharmaceutical industry routinely spends three to five years modifying initial molecule hits to improve solubility and reduce toxicity, benchwork that US developers almost entirely outsource to Chinese contract research organizations.
  • Kevin Weil argues that treating physical lab experimentation as a commodity repeats the policy errors that stripped the United States of domestic solar panel, commercial drone, and hardware manufacturing.
  • Warp Sciences is constructing domestic, 24/7 robotic wet laboratories to compress preclinical testing cycles and generate clean biological data for artificial intelligence models.

Outsourcing the Grunt Work Surrenders the Future Industry

When a biotechnology company identifies an initial target molecule, the longest phase of early drug development begins. Teams spend three to five years synthesizing variations of that compound. They test each iteration to verify that it dissolves in the bloodstream, limits toxic side effects, and hits its biological target.

Over the past two decades, American pharmaceutical companies handed that optimization phase to Chinese contract research organizations. The math looked obvious on a spreadsheet: wet lab testing was treated as cheap, repetitive manual labor. Leaving it to overseas facilities kept headcounts low while domestic teams focused on computational design and clinical trials.

Weil argues that this playbook creates an existential trap. Decades ago, Intel leader Andy Grove issued a warning about treating physical production as low-value overhead: “There's this amazing quote from Andy Grove from like years ago when he was running Intel. He said abandoning today's commodity manufacturing can lock you out of tomorrow's emerging industry. Yeah, it happened exactly in manufacturing and it is on the way to happening in pharma and we can't let it.”

The United States followed that path in consumer electronics, commercial drones, and solar power. Domestic firms kept the intellectual property and sent the physical assembly overseas. Over time, the engineering knowledge, specialized tooling, and operational improvements migrated to foreign factories. Weil warned: “The second thing we need to reshore scientific capabilities from China so that we don't lose our pharma industry the way that we did solar drones and manufacturing.”

Automated Domestic Labs Create the Feedback Loop

Treating physical chemistry as cheap overhead ignores how modern machine learning systems learn. An algorithm designed to design biological compounds requires a tight experimental loop. If your experiments run on clipboards in a third-party laboratory across the Pacific, your data pipeline suffers from batch delays, manual recording errors, and opaque protocols.

Weil's thesis at Warp Sciences centers on building automated wet lab facilities on domestic soil: “Warp uses AI and robotics to help the world discover new medicines faster. Right. We're building robotic labs to do science at scale and we're doing it here in America.”

Building that infrastructure requires heavy capital expenditure: “We raised we raised 215 million. Uh Lightseed LED. We had folks like IVP, Mayfield, Lux, Menllo Ventures, Scribble Ventures, a bunch of amazing folks.”

Running automated labs continuously changes the economics of discovery. When robotic systems handle the pipetting, liquid transfers, and cell assays, every trial produces structured digital telemetry. The software updates against physical reality within hours rather than waiting months for overseas batches. Keeping that operational feedback loop inside the country protects domestic supply lines while giving local models the data density needed to outperform manual competitors.

What to Do With This

Audit your operational pipeline this week to find any physical process or data collection step you treat as an outsourced commodity. Check whether that manual work generates the training data your core software needs to improve. Bring that step in-house or automate it before your third-party vendor acquires the data to replace your core product.