The argument over whether America can build its own machines again has dragged on for years, and the price tags tossed around have often sounded absurd.
Back in May 2025, Wedbush analyst Dan Ives projected more than $3,000 as a possible price for an iPhone made in the U.S., as reported by Business Insider, compared with $799 for a standard iPhone 16.
Mark Cuban thinks that framing misses the point. In his view, the real contest is machine against machine, and he is not just speaking in theory.
His drug company opened a Dallas plant that uses “advanced robotic and AI computer vision technology,” which gives his argument a little more weight than a typical social media hot take.
Mark Cuban’s drug plant puts his automation argument to the test
Cost Plus Drugs was built to strip out the middlemen. Cuban co-founded the company in 2022, and it sells generic medicines at the manufacturer’s cost plus a flat 15% markup, along with pharmacy and shipping fees, skipping pharmacy benefit managers entirely, as reported by Fierce Healthcare.
The company then pushed into making drugs itself. Its 22,000-square-foot Dallas facility, built for $11 million, started with commercial batches of epinephrine and norepinephrine for intensive care patients, with pediatric chemotherapy drugs to follow. Those are cheap, essential medicines that often run short.
Also read: Mark Cuban has strong words on taxes and wealthy Americans
Automation is the selling point. CEO Alex Oshmyansky said the plant could switch from one drug to another “in principle within four hours,” letting it make whatever product is running short, and he described the generic drug supply chain as extremely brittle, as reported by Fierce Healthcare.
That kind of flexibility is exactly what Cuban means when he says better robots and software could tilt the math back toward American factories.
Skeptics see bigger hurdles than fees. Not everyone thinks robots solve the problem. Rosenblatt analyst Barton Crockett has said U.S. iPhone production is a long way off, naming construction lead times, Apple’s exacting quality bar, and weak American demand for that kind of work.
Cuban’s read is that smarter machines eventually close most of that distance.
China already has a head start on robots
The trouble with Cuban’s plan is that China is running the same race, and faster. Morgan Stanley analysts expect humanoids and robots to become the next key engine of China’s export machine over the coming five to 10 years, comparing the push to the rise of its electric vehicle industry, as reported by Bloomberg.
The early numbers back that up. Morgan Stanley estimated Chinese manufacturers made up roughly nine in 10 of the humanoid robots shipped worldwide last year, while American and Japanese rivals were still mostly building prototypes.
That is the gap Cuban says America has to close.
Tesla is America’s loudest bet on catching up in robotics. The carmaker raised its 2026 capital spending guidance to more than $25 billion, funding a major push into AI automation, with a significant share going to convert Model S and Model X capacity into Optimus robot lines, according to The Motley Fool. Analysts expect Optimus revenue to remain limited until 2029.
China’s robot boom has its own cracks, though. Unitree, the country’s best-known robotics company, has dropped 45% from its all-time high since listing in Shanghai, and regulators have reportedly made it harder for humanoid makers to go public.
At least six more Chinese humanoid makers are still said to be lining up listings. But a head start is not the same as a finished race.
America’s AI chip edge is getting squeezed
Cuban’s other condition, staying ahead on AI, looks shakier inside China. Nvidia’s share of the country’s data center AI chip market slid to 34% in the first half of 2026, while Huawei climbed to 28%, with U.S. export controls limiting Nvidia’s supply, according to Reuters.
That erosion started before this year. In 2025 Nvidia still commanded about 55% of China’s AI accelerator market. Bernstein now puts its share at roughly 8% by year-end, according to TheStreet, a collapse driven in large part by Beijing’s push to replace foreign chips with homegrown ones.
More Artificial Intelligence:
- AI’s biggest business opportunity may not look like AI
- AI racks are outgrowing old power: 2 chipmakers have a strategy
- BofA splits software into winners and losers as AI agents arrive
Memory is the next front. CXMT says it has made key progress on a new memory cell design and plans to launch a server memory product using it by the end of 2026. The layout squeezes each cell into about a third less space than mainstream chips at the same line width.
Wall Street noticed. Micron shares fell about 3% on Oct. 12 and were headed for a weekly drop of more than 10%, partly on worries about competition from CXMT.
Some analysts argued the selloff was overdone, pointing out that CXMT has not yet matched the high-bandwidth memory used in AI servers.
Washington plans to guard its AI lead
The White House sees the race the same way Cuban does. Speaking to reporters on Sept. 26, Trump said he would rather not combine American and Chinese AI efforts “because we’re leading by a lot,” adding that “when you’re leading, you don’t open it up to each other,” according to Al Jazeera.
That leaves the tougher half of Cuban’s formula for companies to solve. Building factories in America still takes patience.
When work began on Cost Plus Drugs’ Deep Ellum facility in 2021, it was scheduled to be finished by late 2022. Its first drugs did not come off the line until March 2024.
Still, his core idea is easy to follow. If American robots, software, and AI can match or beat what China puts on its factory floors, labor costs matter far less. Cuban has even predicted that within five years there would be no more shortages of sterile injectable drugs.
Whether the rest of corporate America follows his lead is the question that will decide the race.
Related: Mark Cuban just sent a strong message to America’s workers