Mark Cuban on Why AI Could Turn Solo Founders Into Trillionaires

Julie Yoo opens by saying Mark Cuban had the courage to enter healthcare’s “morass” and that Cost Plus Drugs changed incumbent behavior simply by existing. Because Cost Plus publishes its prices through an API, large companies can compare those prices with their vendors’ offers; Cuban says one company was told by a major insurer that Cost Plus was in its network, even though no such agreement existed. He argues that the company’s core product is not merely inexpensive medicine but trust: customers can see the cost-plus pricing model, and the company does not spend money on traditional marketing because customers tell others how much they saved. Cuban recounts launching costplusdrugs.com on January 19, 2022, with 111 medications. Demand immediately exceeded expectations, manufacturers expanded their portfolios, and the catalog grew to roughly 4,000 stock-keeping units. He says the business is growing every week, with the main challenge being operationally keeping up. Cost Plus also created the Team Cuban Card, a discount card that works at about 17,000 affiliated pharmacies for local pickup. Cost Plus pays those pharmacies $12 per prescription, making it less competitive for a very cheap drug but often cheaper for specialty generics, biosimilars, insulin, and medicines requiring cold shipping.

Cuban describes a second major part of the business: manufacturing drugs that appear on the FDA short-supply list, including pediatric cancer medicines, epinephrine, and other widely used hospital drugs. His co-founder, Dr. Alex Oshmyansky, originally contacted him with the manufacturing idea. Cuban says AI and robotics reduced a facility that might have required a football field to roughly 5,000 square feet, while allowing the production line to switch drugs every four hours. The company sells these medicines to hospitals and clinics. The same robotics and AI have been compressed into mobile manufacturing “pods” roughly the size of a tractor-trailer. A pod costs about $20 million, can be shipped anywhere, and can be configured for sterile injectables, tablets, or other generic-drug production. Multiple pods can be daisy-chained, and the government has already ordered one, with a second order following. Cuban says the pods could be sent to FEMA, the military, or disaster areas and reconfigured for local needs. He is especially proud that pods could be sent to research hospitals working on NF1 therapies, allowing them to manufacture and iterate quickly on specialized treatments. Cost Plus is also making ASO drugs at its Dallas facility, although Cuban jokes that he cannot pronounce what ASO stands for.

Cuban says Cost Plus created Cost Plus Marketplace to sell sterile injectables—both those it makes and those sourced from other manufacturers—to hospitals, clinics, and providers at cost plus roughly 10% or 12%. He believes the marketplace can save hospitals millions of dollars annually, but says the obstacle is often internal purchasing departments: CEOs may agree with the economics while purchasing staff resist changing established arrangements. Drawing on his experience as a self-insured employer and conversations with hospital CEOs and CFOs, Cuban is also pursuing direct contracts between employers and hospitals. He proposes trusting the physician and generally eliminating prior authorization and denials, except in extreme cases such as a fabricated illness claim. In return, Cost Plus would pay the contracted price on time and assume responsibility for the patient’s out-of-pocket cost, either paying it or taking the patient’s credit risk. He wants a lower price because the hospital would face less administrative work and achieve a higher yield. He also wants to publish each direct contract on Cost Plus Wellness so employers and providers can use the agreements as open-source templates for their own negotiations.

Julie says employer-sponsored insurance covers more than half the country—roughly 155 million or more people—and has historically been dominated by national carriers because startups needed their networks and could not overcome entrenched relationships. She says premium increases are now extreme, with some startups facing 60% year-over-year increases. Cuban responds that vertical integration makes the problem worse: UnitedHealth has more than 3,000 subsidiaries and about $150 billion in annual inter-company transfers, while CVS and Cigna are also heavily integrated. Employers may face higher premiums, deductibles, and out-of-pocket costs while being directed to companies the carrier already owns. Cuban focuses especially on self-insured employers, which he says account for about 60% of people receiving employer healthcare. These employers depend on third-party administrators, or carrier-owned administrative-services-only businesses, to process medical claims. He argues that medical-side spread pricing is even worse than the familiar pharmacy-benefit problem because medical spending is the larger market. Employers often cannot access the contracts, claims, or the 835 and 837 files needed to reconcile what was charged, what was paid, and what they ultimately owe. Even public transparency files may be inaccurate. Cuban calls the carriers’ long agreements “contractual terrorism,” citing a 1,335-page City of Dallas PBM contract that no human team can fully review.

Julie argues that AI could make it possible to build a full-stack health plan quickly and reduce administrative waste. Cuban agrees that AI matters clinically and operationally, but says its most immediate value is contractual. His warning is that AI cannot fix a broken financial system once the underlying incentives and agreements are already wrong. If an employer signs a contract containing terms it does not understand, no later AI system can fully protect it. His practical advice is blunt: every employer should run its benefits, broker, consultant, TPA, and PBM contracts through an AI model and ask, “Where are they fucking me over and what should I do about it?” Cuban says the model can identify problems within 30 seconds and help negotiate before signing. He uses AI agents himself to find public-records opportunities, request contracts from governments and other entities, draft letters, and analyze documents stored in Google Drive. He says this has revealed how badly school districts, cities such as Denver, Dallas, and Fort Worth, and unions are being treated. If the employers representing roughly 67% of Americans did this before signing, he believes the major carriers and PBMs would lose their leverage and healthcare costs would “drop like a rock.” He is skeptical that AI-driven claims battles alone will transform the system. Hospitals already have strong revenue-cycle operations, while insurers’ acquisition-built structure leaves them with siloed data. In his view, the real intervention must happen before the contract is signed.

Cuban says the ultimate mission is to disrupt healthcare until it becomes affordable again. He supports universal healthcare in principle, but says it cannot work properly until the country has accurate, accessible data that lets people make informed decisions. In the near term, Cost Plus intends to defend its position through transparent pricing and trust rather than advertising. He says a competitor could offer cost plus 1%, but transparency and reliability are what make customers stay. He expects the manufacturing pods to become a substantial business. Cost Plus hopes to sell a couple hundred pods annually within two years and a couple thousand annually within five years. At $20 million each, a thousand pods would represent $20 billion in annual sales. Cuban says pods could eventually manufacture generic tablets in the United States for less than the company pays manufacturers in India, though Cost Plus chooses partners for quality rather than the absolute lowest price. The major obstacle is regulatory: each generic requires an abbreviated new drug application costing about $365,000 plus other fees, so 1,000 applications would cost at least $365 million before pod costs. He expects American-made products to appear on the website with a small American flag, but says that is probably two years away because of FDA work. Cuban says technology alone will not overcome “scale capture” and regulatory capture. Large companies talk about technology improving their earnings and shareholder returns, not lowering employer or patient costs.

Cuban says AI will not eliminate the need for entry-level workers. Models have latency, and organizations still need people who understand context, monitor consequences, and remain in the loop as systems change. His formulation is that AI does not replace jobs; people who know how to use AI replace people who do not. Executives must learn how agents work, the difference between open and closed models, and the limits of current systems. He illustrates the point with a consumer-food executive: Cuban asked ChatGPT to design a cheaper, GLP-1-friendly competitor to the company’s leading cookie, including materials, sources, and manufacturing requirements, and the executive realized how quickly a serious competitor could be sketched. For healthcare consumers, Cuban sees immediate value in using AI to understand bills and financial-assistance options. He recommends asking an AI system to identify nonprofit hospitals and their assistance programs before a crisis. Those programs may use federal poverty levels and, in some cases, extend to families at 600% of the federal poverty level—potentially more than $100,000 in some states. He says hospitals should make these programs easier to find rather than relying on insurers to send patients. Looking further ahead, Cuban predicts that the cost of designing a drug personalized to an individual’s DNA will approach zero as algorithms improve. Over-the-counter medicines such as ibuprofen or Tylenol could be personalized instead of forcing people to experiment with standardized products. He expects personalized versions of drugs such as Eliquis, Xarelto, or warfarin to follow. Doctors will remain necessary for their senses, judgment, and ability to interpret novel individual data, but medical education will shift toward using models and translating their conclusions.

If he were advising a president, Cuban would begin with transparency. He would require every contract involving taxpayer money to be public and unredacted, allow patients to apply a cheaper price they find toward their deductible, and require nonprofit hospitals to publish detailed general ledgers and procurement records. He also proposes replacing conventional insurance with individually owned healthcare accounts: people would deposit an amount comparable to an ACA silver-plan premium, retain the money and interest, borrow from a bank when medical costs exceed the balance, and receive federal or state-backed stop-loss protection. If they remain healthy until Medicare age, they would keep the accumulated money. The goal would be to eliminate insurers’ arbitrage while ensuring providers are paid. On AI’s public image, Cuban says frontier companies are failing to use the social-media algorithms that shape public opinion. Because each person sees a different feed, he argues that AI companies need targeted education and marketing rather than generic position papers. He also says communities’ opposition to data centers often reflects poor contracts and a lack of understanding, not just opposition to AI itself. Asked where entrepreneurs should build, Cuban answers “all of them.” He distinguishes people who use AI to learn anything from people who use it to avoid learning. Entrepreneurs should ask AI what problems remain unsolved, cross-check its output with other models, and tolerate interactions that feel strange or uncomfortable. He calls AI the greatest technological transition he has seen and predicts that a “trillionaire in the basement” could emerge. The episode closes with his view that the opportunity is not a single white-space market but the willingness to use AI confidently to discover one.