On the All‑In podcast, former hedge‑fund manager Martin Shkreli highlighted three private photonic‑chip startups—Neurofos, Opticore and Lumi—claiming they could outpace Nvidia’s AI hardware. Nvidia, however, remains dominant, reporting record Q2 FY2027 revenues and expanding its own optical roadmap.
During the All‑In podcast released on October 12, 2026, Martin Shkreli said he is exploring the photonic‑computing field and mentioned three private startups—Neurofos, Opticore and Lumi—that are developing optical matrix‑multiplication cores. He argued that these designs “would run circles around what Nvidia does,” noting that light‑based calculations happen in “instant time.”
Photonic chips encode numbers in beams of light, allowing matrix‑multiplication operations to be performed as the light traverses the chip, which promises faster processing and lower energy consumption compared with traditional electronic GPUs. Shkreli cautioned, however, that neural‑network models also require nonlinear functions and substantial memory, both of which remain challenging for purely optical solutions and still rely on conventional silicon components.
Despite the hype around photonic computing, Nvidia’s own results underscore its current market strength. In its Q2 FY2027 earnings release, the company reported revenue of $96.22 billion, a 105.8 % year‑over‑year increase, with data‑center revenue of $89.02 billion and data‑center networking revenue of $40.313 billion. The firm is also integrating optics into its product line, shipping Spectrum‑6 switch systems with pluggable and co‑packaged optics and partnering with Marvell, Coherent, Corning and Lumentum on silicon‑photonic initiatives. Chief executive Jensen Huang summed up the outlook, saying, “Now, compute is revenue.”
Investor sentiment reflects both optimism and caution. Nvidia shares fell 6.39 % over the past week amid a broader tech rout, and on October 17, 2026, the stock traded at $214.13, down 1.43 % on the Pluang platform. The market’s focus remains on Nvidia’s ability to sustain AI demand while monitoring the progress of emerging photonic competitors.
Shkreli’s comments add a speculative layer to the conversation, suggesting that if startups like Neurofos, Opticore and Lumi can overcome the technical hurdles of nonlinearity and memory integration, they could pose a long‑term challenge to Nvidia’s dominance in AI chips.