Avoid shorting quantum-computing stocks solely on skepticism: the technology may be overhyped, but short positions can incur large losses if timing is wrong. Treat photonic-computing startups such as Neurophos, OptiCore, and Lumai as speculative watchlist names; technical hurdles remain, and no investment targets or commercialization timelines are provided. The insights offer no clear buy or sell signal for NVIDIA (NVDA).

Quantum-Computing Companies (Industry)

• Martin Shkreli argued that quantum computing is being oversold as a general-purpose computing breakthrough. He said current machines are slow, limited, and suited to a narrow set of problems rather than ordinary computing tasks.
• He said quantum computers could make certain factoring and cryptography-related calculations dramatically faster, but argued that this advantage has limited business value unless those specific problems matter to a company.
• He criticized the investment case at Google, IBM, and Microsoft, saying their spending depends on the hope that useful applications emerge someday. He described Google as having a few prototypes and IBM as offering a small cloud-based system.
• Shkreli said he and his wife had tested quantum computers and could not use them to multiply 300 by 200. He also said they had shorted quantum-computing companies and lost millions of dollars, while acknowledging that timing a short is difficult.
• As a cautionary analogy, he compared quantum hardware to earlier computing technologies that disappeared or went bankrupt, such as punch cards, vacuum tubes, and core memory.
• He said OpenAI had reportedly rejected quantum-computing pitches.

Takeaways

• The discussion presents a bearish, skeptical view of quantum-computing businesses, especially where valuations rely on broad future applications rather than demonstrated commercial use.
• The speaker’s own losses underline a key risk: even if an investor’s technological thesis seems persuasive, shorting can be costly if the timing is wrong.
• The transcript gives no specific stock tickers, price targets, or timeline for when quantum computing might become commercially useful.

Photonic and Analog Optical Computing

• Shkreli highlighted analog optical computing as a potentially more practical alternative for certain workloads. He said matrix multiplication—the type of operation he associated with much AI and GPU computing—could happen very quickly in light.
• He named Neurophos, OptiCore, and Lumai as startups working on photonic matrix-multiplication cores, which he suggested could outperform NVIDIA on some computations.
• He cautioned that this is not the whole story, citing challenges involving non-linearity and storage. He said he and his wife were investigating the area and might start a company.

Takeaways

• Photonic computing is presented as an early-stage, speculative opportunity tied to AI and computing infrastructure, not as an established investment recommendation.
• The potential advantage described is faster matrix multiplication; the stated challenges include storage and non-linearity. The transcript provides no performance data, financial details, price targets, or commercialization timeline.

NVIDIA (NVDA)

• Shkreli argued that current quantum computers are not positioned to outperform NVIDIA machines on general computing tasks.
• Conversely, he suggested that photonic matrix-multiplication cores could eventually “run circles around” NVIDIA for certain workloads, while noting that important technical challenges remain.

Takeaways

• NVIDIA is discussed as a benchmark for conventional GPU computing, not as a specific buy or sell recommendation.
• The transcript’s comparison is focused on potential computing performance; it provides no valuation analysis, price target, or investment timeline.