The Graveyard of Brain Drugs Just Got a New Gravedigger

The failure points are everywhere. Screen failure rates in neurology studies average 57%, meaning more than half of patients who show up don't even qualify. In preclinical Alzheimer's trials, that number climbs to a staggering 88%. A single Phase III Alzheimer's study can cost around $370 million and drag on for up to eight years.

Then there's the placebo problem. In psychiatric and neurological trials, patients often improve simply because they're getting attention and care. That inflated placebo response crushes the statistical separation between drug and sugar pill, killing otherwise promising compounds. The high-profile failure of ulotaront in schizophrenia trials in 2023 was partly blamed on exactly this phenomenon.

Subjective endpoints make everything worse. When your primary measure of success is a clinician's interpretation of a cognitive test, you're introducing noise at every step. Different raters score differently. Patients have good days and bad days. The signal gets buried.

The Bakersfield Experiment

Kordata's first move is interesting and a little unexpected. Rather than setting up shop at a major academic medical center in Boston or San Francisco, the company planted its flagship trial site in Bakersfield, California. That's deliberate.

Kordata's thesis is that its AI platform can simplify trial operations enough to bring the vast majority of hospitals that don't currently participate in research into the game. If you can turn complex neural data workflows into repeatable, manageable processes, community hospitals become viable research sites. That means access to patient populations that are typically underrepresented in neurology studies.

The company claims its model can compress recruitment timelines to 3 to 6 months instead of the years that large CNS trials typically require. That's an ambitious promise, and sponsors will want proof before betting their programs on it.

Who's Running the Show

The leadership pairing is worth noting. Emil Hewage, who also serves as CEO of BIOS Health, brings the neurotech and AI pedigree. He built the underlying neural data platform that Kordata is commercializing for clinical trials.

On the operational side, Dawn McCollough serves as president. Her résumé carries serious weight: roughly 30 years in clinical trials, involvement in more than 15,000 studies, and 27 drugs successfully brought to market. She previously held executive roles at both Biogen and Novartis, two companies with deep neuroscience portfolios. If Kordata's AI is the engine, McCollough is the driver who actually knows the track.

The pre-seed round (amount undisclosed) was backed by MAVRK Celestia Fund, Kern Venture Group, and Digital Neural Infrastructure Holdings, with strategic support from BIOS Health.

The $26 Billion Question

Kordata points to a $26 billion backlog in clinical trials for neuromodulation and neural therapies. The initial target indications include Alzheimer's, Parkinson's, epilepsy, and drug-resistant forms of cardiovascular and autoimmune conditions where neural circuits play a central role.

The timing aligns with a broader regulatory tailwind. The FDA recently launched a pilot program for real-time clinical trial data monitoring, which Kordata explicitly references as validation of its approach. Industry analysts tracking 2026 trends have noted that "platformization" (moving from one-off AI tools to integrated operating models) is becoming a primary competitive differentiator in trial execution.

The Catch (Because There's Always a Catch)

For all its promise, Kordata faces the same question every AI-in-trials startup must answer: does it actually work in practice?

Regulatory acceptance of AI-driven dosing decisions and novel neural biomarkers as trial endpoints will require extensive prospective validation. The FDA doesn't take kindly to black-box algorithms making decisions about drug doses in vulnerable patient populations. Transparent models, peer-reviewed evidence, and clear performance metrics will be non-negotiable.

Then there's the community hospital challenge. Asking a facility with limited research infrastructure to suddenly manage continuous neural data streams, even with a simplified platform, is a tall order. Staffing, IT governance, and clinician training could all become bottlenecks.

And the competitive landscape is getting crowded. Companies like Unlearn.ai (digital twins for smaller control arms), Phesi (trial simulation with 100 million patient profiles), and TriNetX (real-world data networks) are all attacking different pieces of the same problem. Kordata's neural-data specialization is a genuine differentiator, but it only matters if sponsors are willing to bet on a brand-new platform with no track record of completed trials.

The Bottom Line

Kordata is making a fascinating, focused wager: that the biggest bottleneck in brain drug development isn't the science itself, but the clinical trial machinery surrounding it. If NeuroTune can deliver on the promise of real-time, biologically grounded trial execution in neurology, it could meaningfully improve those dismal 5 to 6% success rates.

But "if" is doing a lot of heavy lifting in that sentence. The next 12 to 24 months will tell us whether Kordata is a genuine breakthrough or another ambitious pitch that couldn't survive contact with the messy reality of running clinical trials in American hospitals.

A Tiny Korean Biotech Just Put Up MASH Numbers That Rival Novo Nordisk

A small Korean biotech just dropped MASH liver disease data that looks competitive with Novo Nordisk's blockbuster semaglutide. The catch? Only 35 patients were in the trial, and the competition is about to get fierce.