Martin Shkreli: Big Pharma's Blind Spot, Betting Against Quantum & What Prison Taught Him

Summary · 3 min read instead of 25 min

Martin Shkreli started at 16 at Jim Cramer's hedge fund, later worked at a Tiger Cub and started his own fund before building companies that bought and developed drug rights. He became known for raising the price of a life-saving drug 5,000% and served seven years in prison.

Shkreli argues quantum computing has essentially one business-relevant trick, fast integer factorization, and says he has shorted quantum companies despite losing millions doing so. He sees photonic matrix multiplication as a better bet than ever-larger GPU build-outs, though he stays long semiconductors. Drawing on his pharma career, he says big drug companies sell undervalued molecules cheaply, yet he does not expect a single AI-relevant drug in the next 10 years. He calls Anthropic's Dario Amodei a fear monger, champions effective accelerationism, and says the gifted-kid pressure to live up to an image put him in prison.

Minute by minute

1. 1:43

   Quantum computing has one real trick. Shkreli credits his wife Chelsea, an early OpenAI employee who taught quantum computing with Scott Aronson, for his thesis. He says quantum machines run at hundreds or thousands of hertz, are single core, have no quantum RAM and only excel at problems like integer factorization. He says he cannot multiply 300 by 200 on any quantum computer and lost millions shorting quantum companies, comparing the bet to going long punch cards and vacuum tubes.

2. 6:01

   Photonic chips could beat GPUs at matmul. Shkreli says a GPU spends 99.9% of its time on matrix multiplication, which takes a billion operations for two thousand-element square matrices. He cites a Microsoft analog optical computing paper in Nature showing matrix multiplication in light is effectively free. He names startups Neurofoss, OptiCore and Lumai and says he may start a company in the space.

3. 7:45

   Big pharma sells drugs too cheap. Shkreli says Cramer made 23% a year net of fees by trading constantly, but he found he was a better entrepreneur than trader. He says he bought a drug for $2 million from a company now worth $6 billion and that Pfizer would sell a drug worth a billion dollars for a million. He repositioned Sparsantan from hypertension to FSGS, where it won FDA approval for about 30,000 patients who had no drug before.

4. 11:26

   Today's GPUs will be scrapped. Shkreli predicts H100s, B100s and B200s will be in the scrap heap in 20 years and calls plans for space and nuclear-powered compute insane. He is long semiconductors on Jevons paradox, but warns a transformer a thousand times more efficient would be bad for Nvidia. He adds he probably would have guessed Yahoo and Cisco were longs in 2000.

5. 12:58

   Shkreli calls Dario Amodei a fear monger. Shkreli says he bought the sell-off tied to Dario Amodei's essay and calls Amodei a fear monger, comparing his exit from OpenAI to a partner taking Pandora's box down the street. He says he helped start the effective accelerationism movement and that AI is just software. He says he expects no AI-relevant drug in the next 10 years because computers have long been used in drug work without much gain.

6. 16:28

   Demand for intelligence is unlimited. Shkreli argues limiting AI's market to knowledge workers undercuts it, citing a friend's startup, Core Automation, which aims to have AI start businesses. He says McDonald's could rationally spend $10 billion on 10 million GPUs to pick its next sandwich. He compares AI to the internet, where consumers and businesses capture most of the value, and names Meta as a likely winner thanks to Muse.

7. 21:15

   The gifted-kid trap led to prison. Asked about Sam Bankman-Fried, Shkreli says intellectually gifted people carry pressure to live up to an image, citing Elizabeth Holmes and Raj Rajatnam too. He says he was petrified of not being seen as a great trader, though he was a wonderful entrepreneur, and that refusing to admit this put him in prison.

8. 24:13

   Prison shows people share the same needs. Shkreli says prison taught him that inmates of every background share common features, such as being upset when friends lose touch or missing someone. He found it refreshing to remember that, whatever heights people reach, most want to be the best parent they can.

Published Oct 12, 2026.