Medidata Solutions

The most important software company you can't name

Somewhere right now, a patient swallows a pill in a trial they hope will save their life. The moment that dose is logged, it lands in Medidata.

Medidata Solutions does not make medicine. It makes the system that medicine is tested on. Headquartered at 350 Hudson Street in New York, the company builds the cloud platform where pharmaceutical and biotech researchers design trials, capture data, engage patients, and prove - to regulators who do not accept "trust me" - that a new therapy works and is safe.

The scale is easy to miss precisely because the work is invisible. Medidata says its technology helped advance 80% of FDA novel drug launches in 2025, moved 70% of new molecular entities through its platform, and supported the design of more than half of all new medical devices. Eighteen of the top 25 pharmaceutical companies are customers. If a drug reached your pharmacy in the last decade, the odds are good its evidence passed through here first.

Medicine ran on paper, and paper lies slowly

In the late 1990s, a clinical trial was a logistics nightmare wearing a lab coat. Data lived on paper case report forms, faxed from hundreds of sites, re-keyed by hand, queried by mail, and reconciled months after the fact. Errors hid in the gaps. A trial could take years partly because the information about it was always out of date.

This was the tension that would define Medidata, and it has never fully gone away: the science of medicine had raced ahead, while the plumbing that carried its evidence was stuck in the era of the filing cabinet. Every day of delay was a day a working drug stayed off the shelf - and a day a failing one kept burning money.

The fix sounds obvious now: put the trial in software. In 1999 it was not obvious at all. Regulators were cautious, hospitals were offline, and "the cloud" was still mostly weather. Somebody had to bet that an entire risk-averse industry would move its most scrutinized records onto the internet.

Three people, one hand-coded app

The bet was placed by an unlikely trio. Glen de Vries and Ed Ikeguchi had been at it since 1994 with a precursor called OceanTek, building web applications for clinical research before most people had a browser. In 1999 they joined forces with Tarek Sherif, a former investment banker, who moved into a small office with de Vries while de Vries hand-coded the original software himself.

Their wager was specific. Not "the internet is big," but "regulated medicine, the most conservative buyer on earth, will trust the web with its data if the system is good enough." That is a strange thing to believe in 1999. It is also, conveniently, the kind of belief that turns out to be worth $5.8 billion if you are right.

Money followed conviction. A $10 million round with Insight Venture Partners arrived in 2004, with earlier backing from Milestone Venture Partners and Stonehenge Capital. In June 2009, Medidata went public on Nasdaq under the ticker MDSO. The filing cabinet had found its replacement, and Wall Street had noticed.

Rave, and everything that grew around it

The product that changed everything was Medidata Rave EDC, introduced in 2001. EDC stands for electronic data capture, which is a dry name for a quiet revolution: a single system where trial data is entered once, validated instantly, and visible to everyone who needs it. No fax. No re-keying. No waiting for the mail.

From that root, a platform grew. The Clinical Cloud arrived in 2013. Patient-facing tools - electronic consent, electronic outcome assessments, sensor data - let trials reach people at home instead of only in clinics. Analytics turned the exhaust of thousands of past trials into predictions about new ones, including synthetic control arms that can reduce how many patients get a placebo.

The receipts, by the numbers

Plenty of companies claim to be transforming an industry. Medidata has the unglamorous virtue of being measurable. Revenue reached roughly $636 million in 2018, the last full year before the acquisition, up from $4.1 million in 2003. The customer roster reads like a pharmaceutical hall of fame: Johnson & Johnson, AstraZeneca, Amgen, GlaxoSmithKline, Novartis.

Then there is the deal that settled the argument. In 2019, French industrial-software giant Dassault Systèmes bought Medidata for $5.8 billion in cash, at $92.25 a share. For a company that started in a small shared office with hand-coded software, that is a respectable exit. For Dassault, it was the anchor of a bet that drug development would become a data-and-simulation business - the same logic it already applied to aircraft and cars.

The partnerships keep the momentum honest. Worldwide Clinical Trials signed on to embed Medidata AI across the full trial lifecycle; the CRO Caidya adopted its AI-powered experiences. These are the buyers who would notice fastest if the software did not actually work.

"I want everything to go faster"

That line belongs to Anthony Costello, who became CEO in 2025, and it doubles as a company thesis. Medidata's stated mission - powering smarter treatments and healthier people - sounds like a poster until you remember what it is racing against. Every month a trial drags on is a month patients wait. The mission is, bluntly, about time.

That is why the 2026 product news is all about removing friction. Dot, the AI companion unveiled at NEXT New York, threads the study, data, and patient experiences into one journey. AI Study Build reads a protocol and configures the trial system from it. The next-gen imaging tool redacts protected health information 32% faster. None of these are flashy. All of them buy back days.

The boring infrastructure of being cured

Here is the case, and the skeptic's caveat in one breath: market-share figures like "80% of FDA novel drug launches" come from the company itself, and the right reaction is interest, not awe. But even discounted, the direction is unmistakable. Drug development is becoming a data problem, and Medidata has spent a quarter century building the place where that data lives.

If AI-designed trials, decentralized studies, and synthetic control arms become the default rather than the experiment, the company that already runs the underlying platform has an awkwardly good seat. The future of medicine looks less like a lone genius in a lab and more like a very fast feedback loop - and feedback loops need infrastructure.

Return to that patient, swallowing that pill, hoping. The dose is logged. It used to vanish into a paper form, faxed somewhere, re-keyed weeks later, true by the time it no longer mattered. Now it is in the system before the patient leaves the chair - visible, checked, counting toward an answer that arrives sooner. That is the entire point of Medidata. Not the software. The time it gives back.