TJ Parker: Hims' compounded GLP-1 pill is a 'financial hail Mary' with no evidence it works — and Trump Rx is the real deal

Summary

TJ Parker, co-founder of PillPack and former VP of Health and Pharmacy at Amazon, calls Hims & Hers' newly launched compounded GLP-1 pill a "financial hail Mary" with no clinical evidence it works. He distinguishes this move sharply from Hims' earlier compounding of injectables.

Hims announced it would mass-market an oral semaglutide pill using liposomal absorption technology, bypassing Novo Nordisk's patented SNAC technology that enables oral absorption. When Hims compounded injectable semaglutide during shortage periods, they used a formulation similar to the branded drug. This time they've invented a new delivery mechanism with no clinical data supporting it.

By Parker's analysis, Hims is selling a $50 drug that likely doesn't work but still carries side effects from the active ingredient, which won't be absorbed through the GI tract. The FDA and Novo responded immediately, with FDA Commissioner Marty Makary threatening swift enforcement and Novo signaling legal action.

Parker frames this as fundamentally a safety and efficacy problem, not an IP issue. The FDA moved faster and more explicitly than it did with the earlier injectable compounding because the agency enforces safety and efficacy, not IP rights.

Why Hims took the bet

Hims is a $5 billion market cap company where roughly half of revenue comes from compounded GLP-1s. That market is contracting. The company has done significant buybacks and CapEx investment. Hims' customer base bought the injectable because it was the cheapest option. Novo Nordisk's Wegovy pill is half the price and doesn't require weekly injections, eliminating Hims' value prop entirely. Novo has no shortage issues and pills are exponentially easier to manufacture than injectable pens. Parker expects significant churn as the compounded injectable business that built Hims faces direct price competition from an FDA-approved product with clinical evidence.

Hims conflated two different businesses and became addicted to the margin profile of one. As a digital pharmacy distributing branded drugs, the company makes $10–15 on a $150 product. As a compounder, it made $250 products with 90% gross margins. Hims shifted toward manufacturing without acquiring IP, R&D infrastructure, or the regulatory moats that come with being a real pharma company. Competitors like Ro managed the transition away from compounded drugs by ramping access to branded therapies. Hims didn't. An acquisition seems unlikely now given the regulatory overhang and retrospective liability.

Trump Rx and net pricing

The real story in pharmacy is Trump Rx. The executive order aims to eliminate the gap between list prices (often $1,000–$1,200 for GLP-1s) and net prices insurers and PBMs actually pay ($200–$300). Manufacturers have agreed to offer consumer-directed net prices on branded drugs. For out-of-pocket payers, that's roughly 70% savings compared to a year ago.

About 25 non-generic branded drugs currently have low net prices on the program. If manufacturers start launching new products at net pricing, it unwounds the value of the current supply chain and PBM middlemen. Parker sees this as potentially transformative to how drugs go to market and are priced.

AI doctors and medical history

Parker sees consumer AI health tools as net good. People are already using ChatGPT for health advice, and it's more efficient than Google searches. The real unlock happens when AI is fully integrated with a patient's complete medical history.

Providers can access much broader datasets to deliver care. Technology companies accessing data via HIPAA face severe constraints. Most healthcare providers don't contribute data to those channels. OpenAI and other non-provider AI companies are therefore constrained in their ability to provide care comparable to what a licensed provider can deliver.

Companies like Lotus, which employ actual doctors, can offer a different value proposition. The real magic happens when AI conversations distill into provider briefings before appointments, arming doctors with patient history and prior discussions rather than forcing patients to repeat themselves.

Parker shares a concrete example. A patient had plantar fasciitis on one foot. The AI system flagged that the same issue had appeared 20 years earlier on the other foot, something the patient didn't remember. Trend analysis across historical lab tests and screening recommendations based on what's been done versus missed illustrate why full medical history access matters.