What happens in the business world when artificial intelligence compresses a company's headcount to the extreme? Matthew Gallagher, 41, has provided an answer using two months, $20,000 in startup capital, and over a dozen AI tools: Medvi, a telehealth company with only himself and his brother as formal employees, yet projecting $1.8 billion in revenue by 2026. Founded in September 2024, the company focuses on GLP-1 combination weight-loss drugs and men's health products. Driven by AI, it achieved $401 million in revenue and $65 million in net profit during 2025. However, entering 2026, Medvi became embroiled in multiple controversies including false advertising, improper marketing, legal disputes, and warnings from the U.S. Food and Drug Administration (FDA). Medvi's rapid rise and sudden stumble reflect that the One-Person Company (OPC) model, while benefiting from the AI era's dividends, must also pay a high price. Matthew Gallagher is not a typical Silicon Valley elite. He taught himself programming early on and later founded a watch subscription company, Watch Gang, which struggled to become profitable long-term. Following the release of ChatGPT, he began intensively researching AI tools and quickly identified an opportunity in telehealth. He used AI to handle branding, marketing, website development, customer service, and data analysis, while outsourcing processes like doctor consultations, prescriptions, pharmacy services, delivery, and some compliance aspects to external platforms. Crucially, he targeted one of the hottest consumer healthcare sectors: GLP-1 weight-loss drugs. In its first month, Medvi acquired 300 customers, adding another 1,000 in the second month. For the full year 2025, sales reached $401 million, with a cumulative customer base of 250,000. The company achieved a net profit of $65 million, representing a net profit margin of 16.2%. For 2026, Medvi's revenue target is $1.8 billion. Reports indicate its daily revenue already exceeds $3 million. For comparison, a listed competitor in the same sector, Hims & Hers Health, employs over 2,400 people, generates approximately $2.4 billion in revenue, but maintains a net profit margin of only 5.5%. As business grew explosively, Matthew Gallagher found it difficult to manage alone and hired his first, and currently only, full-time employee: his brother Elliott, who handles information screening. Matthew Gallagher himself focuses on core business matters. Medvi's success story quickly captured attention in the tech world. OpenAI CEO Sam Altman stated in an email to media that he appeared to have won a bet with fellow tech CEOs about when a "billion-dollar one-person company" would emerge, expressing a desire to "meet this person." A venture capital partner focused on the tech sector commented that the most noteworthy aspect of Medvi is not "one person starting a business with AI," but the founder's inherent ability to select the right market and effectively utilize tools. The partner noted that Gallagher's prior experience in a highly distribution-focused watch subscription business meant "distribution is his forte." For him, AI served as a tool to amplify this distribution and front-end customer acquisition efficiency. The partner described the approach as an "AI stitched-together creature" – not relying on a single AI capability, but piecing together multiple tools and platforms to rapidly assemble a basic commercial engine. The partner emphasized that while AI can replace much repetitive labor, it cannot replace human "taste," which encompasses aesthetic judgment, business acumen, cognitive frameworks, decision-making ability, and long-formed ways of thinking. AI can greatly enhance efficiency in knowledge generation and repetitive tasks, but the human element remains critical for choosing the right market, judging direction, and making key decisions amidst chaos. The quality of business judgment often relies on information disparities not readily available online, existing within small circles, networks, and industry exchanges, making this exclusive experience and layered information the most valuable and hardest for AI to replicate. Just as Medvi garnered increasing attention, underlying risks quickly surfaced. Controversy first arose around its marketing methods. Reports indicated that approximately 30% of Medvi's advertising was conducted through affiliate marketing channels. Some affiliates ran ads on platforms like Meta featuring what appeared to be AI-generated "fake doctor" images promoting products, without the required prominent disclosures. These "doctor" accounts were riddled with inconsistencies. One account, "Matthew Anderson, M.D.," listed an Angolan phone number, with historical information suggesting it previously belonged to a gospel music singer. Another, "Spencer Lanford, M.D.," had records pointing to a clothing store in the Republic of the Congo. The company also faced direct pressure from regulators. On February 20, the FDA issued a warning letter to Medvi, directly citing "false or misleading" promotional content on its website. Consumer rights groups also questioned Medvi's model. The director of a food, nutrition, and obesity project at a national consumer league stated that as early as September 2025, the coalition had sent a joint letter to the Federal Trade Commission (FTC) requesting an investigation into six telehealth companies, including Medvi. She argued that Medvi's use of terms like "expert trusted" and "doctor approved" on its website violated the FTC Act. Furthermore, over the past 11 months, Medvi has been involved in at least three lawsuits where plaintiffs accused the company and its affiliate marketers of violating laws by sending unsolicited spam marketing text messages and emails. In response to the controversies, Medvi issued an official statement on April 8. Regarding the FDA warning letter, Matthew Gallagher argued that the mentioned URL (medvi.io) was not the company's official website (medvi.org), but an unauthorized page used by an affiliate marketer. He claimed Medvi never directly received the FDA correspondence and had demanded the partner remove the non-compliant content and respond directly to the FDA. Concerning the "fake doctor" ads, Medvi's statement acknowledged awareness of the situation, stated it had updated its marketing policy to explicitly prohibit ads using AI-generated doctor images, and was initiating corrective actions. The venture partner suggested that the issues facing Medvi reflect not just one company's marketing失控, but the systemic challenge of governing AI-generated content. Images, text, and personal stories can be mass-produced by AI, posing a huge challenge for platforms and regulators to identify, manage, and distinguish human-created from AI content. Especially in a highly sensitive, serious scientific field like healthcare, AI was used not only by Medvi for efficiency but also to create numerous front-end marketing "hooks," some of which entered high-risk territory. Medvi's journey from rapid fame to controversy highlights the inherent contradiction of the OPC model: extreme efficiency often accompanies extreme fragility. Reports described instances where Medvi's customer service chatbot sometimes fabricated drug prices, forcing Gallagher to honor the erroneous quotes. At times, the AI suffered "hallucinations," claiming the company sold non-existent products. If customers insisted on speaking to a human, calls were forwarded directly to Gallagher's personal phone. He reportedly answered over 1,000 customer service calls in one day. On another occasion, the Medvi website crashed, halting orders for an hour. With no one else available to fix it, Gallagher had to run home from a mountainside, reportedly losing around 200 potential customers. The OPC model also means the founder is completely tied to the company. Gallagher admitted in media interviews that he works all the time except when sleeping, showering, or spending time with his children. To free up more time for company matters, he even used an AI clone of his own voice to handle private affairs. This total dedication yielded remarkable business growth but also left him feeling "lonely" and contemplating hiring more people. Domestically, OPCs are also emerging. One solo operator of an FPGA video transmission service noted that OPCs aren't an entirely new "myth," having previously been called "independent entrepreneurs." The difference now is that tasks once requiring outsourcing can first be handled by AI, "like having an extra employee," before being refined by the individual. However, he cautioned that the most practical issue for an OPC isn't "whether you can do the work," but whether customers are willing to buy from you. The biggest difficulty is gaining trust from new clients. "Unless you have a compelling, irreplaceable core advantage, clients typically won't choose a solo operator." While many tout AI's role in boosting efficiency for OPCs, the key question is whether anyone wants the increased output. Another serial entrepreneur focused on AI transformation consulting and personal AI efficiency services pointed out that, theoretically, a super-individual using a dozen or more AI agents to complete a business loop is feasible. However, AI is an "amplifier," not an "engine." He stressed that based on his experience, entrepreneurs should first validate whether they can complete a basic business loop, even with low profitability, *without* AI tools. If that foundational capability is missing, AI merely accelerates failure. Regarding OPC risks, he highlighted that the limitations of single-person decision-making are amplified. If AI is used incorrectly, it can further迎合 your existing思路, acting as a "confidence amplifier." This means if an idea is fundamentally flawed, AI won't provide a warning but may instead推进 the logic forward, ultimately accelerating the failure. Beyond decision-making risks, he identified the psychological burden of loneliness and lack of positive feedback as the hardest aspect for an OPC to bear. A true OPC often involves working alone at a computer without colleagues or the immediate feedback from other people. While traditional employment involves negative feedback like management and evaluation, it also provides a stable stream of positive feedback through colleague interaction, organizational structure, and the certainty of earning a salary. In contrast, for an OPC, both positive and negative feedback largely come from oneself, and the entire process's uncertainty is borne alone.
Comments