P.S. For early readers, I tried to release this thread in 1 go but Twitter crashed and only released the first tweet -- forcing me to do every other one 1-by-1... sorry! 😢

END/ Check out Jensen's interview below for more details on his lightbulb realization moment in 2012 and his current views on generative AI. If you enjoyed learning about Nvidia's early days, others may too! Please retweet to help get the story out! 🙏

9/ Pivot to AI Jensen realized, “features weren't going to be engineered or created by humans, but somehow automatically  extracted out of data, refined out of data" — i.e. every company in the world was about to need GPUs to make sense of their data, even if they didn’t realize…

8/ AI’s lightbulb moment 💡 Over the next decade, “not much happened” — stock was flat, competitors rose up the graphics space — until 2012. One day, 3 computer vision scientists simultaneously approached Jensen. They were all gearing up for ImageNet, the biggest annual…

7/ IPO! On Jan 22, 1999, NVDA IPO’ed at $12/share and $625M valuation — Adjusted for stock splits, the stock is up almost 50,000x today! And this was just the beginning… it was still the pre-AI era.

6/ The 1st GPU NVDA in 1998 had clearly reached product market fit. Revenue climbed to $158 million and it was time to expand the product suite. The next chips built were: - RIVA TNT, the 1st multi texturing 3D processor - DirectX® 6.0, which it co-developed with Microsoft -…

5/ Getting to PMF (product market fit) The founders kept trying. In 1997, they released the RIVA 128. It was 400% faster than incumbent graphics processors, and instantly won deals with major enterprise OEM customers: Dell, Gateway, and Micron! aka the 1st big success!

4/ First Product = Failure In 1995, NVDA released its first chip, the NV1, and was a huge commercial bust. Performance sucked, raw materials were too pricey, and nobody wanted to pay for it. On the brink of bankruptcy, NVDA laid off 50% of its staff and went back to the drawing…

3/ Capex & First Fundraise In Q3 1994 NVDA closed its first deal with supplier and Euro chipmaker STMicroelectronics. STM agreed to manufacture NVDA’s single-chip GUI accelerator for a hefty price of several million. To fund this, Jensen raised $20M from Sequoia and Sutter…

2/ Cofounders take action So Jensen quit his director job at chipmaker LSI Logic (now Broadcom). And Chris and Curtis quit their engineering jobs at Sun Microsystems. Nvidia initially had no name and the co-founders named all their files NV for “next version.” When the founders…

1/ On Day 0 The idea came together over breakfast at Dennys — to bring 3D graphics computing to the burgeoning video game industry. The risk was clear—$10M+ initial capex needed to ship the first accelerator with no pre-committed customers, no funding, and huge technology &…

Nvidia is about to become the 1st trillion-dollar chipmaker, after surging $200B in valuation in a single day. But when cofounders Jensen, Chris, & Curtis started the company in 1993, they had only $40K in the bank. Here’s Nvidia’s founding story, from 0 to Taxman of AI. 👇 🧵/

56/ The nuts and bolts of a term sheet

Pls share with a retweet if this was helpful! So that others get educated and understand how to position themselves to not get taken-advantage of before an important raise! Start of thread here:

Summary/ Most people have never seen a term sheet until they need to. VCs on the other side of the table have seen 1000s! ⚖️ This 🧵covers the subtle clauses u need to know, which affect ur: - control - dilution - claim to profits - ability to say 👍/👎 to an acquisition offer!

Summary/ Most people have never seen a term sheet until they need to. VCs on the other side of the table have seen 1000s! ⚖️ Here are the subtle clauses you need to know, which affect your: - control - dilution - claim to profits - ability to say 👍/👎 to an acquisition offer!

Summary/ Most people have never seen a term sheet until they need to. VCs on the other side of the table have seen 1000s. Here are the subtle clauses you need to know, which affect your: - control - dilution - claim to profits Pls share if this was helpful so that others get…

9/ Pro-rata rights This gives VCs the right, but not obligation, to maintain their level of ownership throughout subsequent financing rounds. e.g. You have 100 shares and you sell 10 shares to a VC w/ pro-rata rights. You then issue 500. VC can buy 50 shares before anyone else.

10/ No-shop Exactly what it sounds like: can't shop this term sheet around to others for X days (norm: 30) Pretty standard to have, BUT since term sheets are non-binding, said VC can take the max days to do "diligence" then back out last minute. Try to cap the no-shop at 30d.

Weighted average anti-dilution Anti-dilution clause for prior investors that takes into account the number of shares issued in the new (lower) round, therefore reducing the dilution to founders vs. full-ratchet. See formula below.

What's a "founder friendly board"? At the A: "2+1," i.e. 2 founders, 1 investor (appointed by Lead) What's (slightly) dirty but kinda common arrangement? At the A: "2+2+1," i.e. 2 founders, 2 from investors, 1 independent (appointed by company)

8/ Anti-dilution Prevents company from selling/issuing stock at a lower price than what current VC is paying 2 types: - ratchet based (less common) i.e. if ABC sells new shares at a price below Series X, Series X price is reduced to match) - weighted average (more common)

6/ Drag-along rights Only relevant when there's a pending exit: gives majority shareholders the right to "drag" minority shareholders into an exit, i.e. force them to sell their shares on the same terms & conditions Why: to ensure minority shareholders cannot block an exit

7/ Tag-along ("co-sale") rights (similar-but-opposite to drag-along) i.e. upon liquidity event, allows minority to "tag along" with majority note: doesn't always apply to employees, so it's important to check if u want to sell shares alongside existing investors in secondaries

5/ Board seats Note: #1 clause leading to founder loss of control — choose your Lead investor carefully b/c they choose your Board additions! Here's what Boards typically vote on: - hiring/firing senior execs (can vote you out!) - exec comp Typical track: 1 at series A