does oai ventures (not specifically controlled by oai but by sam altman) count?

assuming oai itself, i think 20% 2024, 60% 2025.

@CampbellHutcheson i think it should count yes, @Jacy what do you think?

@Soli I'm not sure I understand. Does action taken by OpenAI Startup Fund, an Investment Limited Partnership whose only general partner is Ian Hathaway and does not even include OpenAI as a limited partner (i.e., investor), count as an action taken by OpenAI? That seems like a clear no.

FYI Sam Altman is no longer its general partner, as reported in an April 25 SEC filing.

@Jacy then it doesn’t count haha

OpenAI would like to just have a working AI Chip company. There are no real competitors to NVIDIA currently on the market OpenAI could just buy (at least none OpenAI would have enough monetary resources to buy) and OpenAI does not have the capacity/the human resources/the know how to start a new hardware company.

Apparently, Groq only got 370 million in funding since their founding which leads me to believe that starting a new AI chip company doesn’t require as much capital as I thought.

@Soli i realized talking to the founder of one of the new asic ai chip companies that it’s very doable because most of the standardized parts are provided by tsmc (not sure about tooling). you can pre-order $100m of product with commitments from a couple of customers.

Will this happen? "Sam Altman Seeks Trillions of Dollars to Reshape Business of Chips and AI"

still trying

@Soli To clarify: We need verification (4) in 2024 (5) for (1-3) to count? So some news about an investment in October but verification in January would not resolve Yes?

@Primer good question - did not think deeply about it but i guess this is the most straightforward way to read the resolution criteria although it feels weird

@Soli I'm not sure how you're distinguishing "semiconductor" from "chips" in that sentence, but for what it's worth, I'm betting NO on the basis that I think the market overestimates the probability of something like what that news article describes, where OpenAI raises the billions it would take to create a new significant player in semiconductors or to acquire an existing one. State-of-the-art semiconductors are extremely hard to produce well and have little need for a "project" like this to continue scaling up as much as they can. Many have tried, but if anyone can do it, it's probably OpenAI, the U.S., or China during the current hype cycle, so I think this is an important and interesting market.

@Jacy I have little knowledge here honestly but I thought maybe AI chips are made from special semiconductors but also have other parts that go into them. To avoid confusion I thought it would be best to state that either would suffice to conclude the market as yes.

Edit: I think we can ignore my comment. AI chips are specialized semiconductors according to the brief research I did now.

I would be surprised if this doesn’t happen in 2024 - I think it makes sense for them on so many levels but maybe I am underestimating the challenge of building chips that can seriously compete with nvidia

@Soli which companies would it possibly fund/acquire if it did? I can't think of any promising AI chip startup/labs

edit: Most funded artificial intelligence (AI) chip startup companies worldwide 2021 based on Statista

Nothing reported for 2022 or 2023

@trixwit good question but i have little knowledge of the industry to give an educated answer here