The ENTIRE History of NVIDIA in 14 Minutes — Transcript

Full transcript

1. 0:00In 1993, three guys sat in a Denny's and
2. 0:03decided to start a company. 30 years
3. 0:05later, that company was worth more than
4. 0:07the entire stock market of most
5. 0:09countries. Not a phone company, not a
6. 0:11burger company, a company that makes
7. 0:13graphics cards. Yeah, the thing your
8. 0:16cousin uses to play Fortnite at 240
9. 0:18frames per second. So, how did a gaming
10. 0:21chip company become the engine behind
11. 0:23chat GPT, self-driving cars, and
12. 0:25basically the entire AI boom? And here's
13. 0:28the wild part. This company almost died,
14. 0:31more than once. One time it was
15. 0:33literally about a month away from going
16. 0:34broke. Stick around cuz the story of
17. 0:37Nvidia is the story of one guy making
18. 0:39one insanely risky bet while pretty much
19. 0:42everyone told him he was cooked. Let's
20. 0:44get into it.
21. 0:45Our main character is Jensen Huang. He
22. 0:47was born in Taiwan in 1963. And when he
23. 0:50was about nine, his parents sent him and
24. 0:52his older brother to the United States
25. 0:54to get a better education. Now, you'd
26. 0:56think he landed somewhere nice. Nope, he
27. 0:58ended up at a boarding school in rural
28. 1:00Kentucky, which was way tougher than he
29. 1:02expected. And his job there? Cleaning
30. 1:05the bathrooms, every day. Future
31. 1:07multi-billionaire scrubbing toilets at
32. 1:10age nine. Honestly, kind of an origin
33. 1:12story, right? He survived it, moved on,
34. 1:15and ended up studying electrical
35. 1:16engineering at Oregon State. Then he got
36. 1:19a job at a chip company called LSI
37. 1:21Logic, then at AMD. And somewhere in
38. 1:24there, he met two engineers who worked
39. 1:26at Sun Microsystems, Chris Malachowsky
40. 1:28and Curtis Priem.
41. 1:30These two were obsessed with one idea.
42. 1:33Computers were getting boring at
43. 1:34displaying things. Games, 3D graphics,
44. 1:37video, all of it was coming, and regular
45. 1:40processors just couldn't keep up. So, in
46. 1:431993, the three of them met at a Denny's
47. 1:46in San Jose. Fun fact, Jensen actually
48. 1:49worked at a Denny's as a teenager,
49. 1:51washing dishes.
50. 1:52The man really went from dishwasher to
51. 1:54running one of the the valuable
52. 1:55companies on Earth. Absolute glow up.
53. 1:58They started the company with around
54. 1:59$40,000. That's it. For a chip company,
55. 2:03that's like trying to start a space
56. 2:04program with your lunch money. They
57. 2:06needed a name. They called all their
58. 2:08files NV, short for next version. Then
59. 2:12they looked for a word with those
60. 2:13letters in it and landed on Nvidia, the
61. 2:16Latin word for envy. Yes, the company's
62. 2:19name literally means jealousy. The goal
63. 2:22was to make everybody else green with
64. 2:24envy. And look, it worked. Hold that
65. 2:27thought. In 1995, Nvidia launched its
66. 2:30first real product, the NV1.
67. 2:33And it was a disaster. Here's why. Every
68. 2:36other company was building graphics out
69. 2:37of triangles. Triangles were becoming
70. 2:40the standard. Nvidia decided to do it
71. 2:42with curved surfaces instead because
72. 2:44they thought it was smarter.
73. 2:46It was smarter. It was also completely
74. 2:48incompatible with everything. And then
75. 2:50Microsoft released its graphic standard,
76. 2:52DirectX, built around, you guessed it,
77. 2:55triangles. So, the NV1 was basically a
78. 2:57brilliant chip speaking a language
79. 2:59nobody else spoke. Sales tanked. Then it
80. 3:02got worse.
81. 3:03Nvidia had a deal with Sega, the video
82. 3:05game company, to build a chip for their
83. 3:07next console. But, the engineers
84. 3:09realized the approach was wrong. Jensen
85. 3:12had to fly to Sega and tell them,
86. 3:14basically, "Hey, so what we promised you
87. 3:17isn't going to work." Now, in a normal
88. 3:19world, Sega says, "See you in court."
89. 3:22Instead, Sega's boss believed in Jensen
90. 3:25enough that Sega still invested $5
91. 3:27million in Nvidia. $5 million that,
92. 3:30according to Jensen himself, kept the
93. 3:32company alive because at that point,
94. 3:35Nvidia was about 30 days from going out
95. 3:37of business. They laid off most of the
96. 3:39staff. The company was down to roughly
97. 3:4135 people. So, the question is, how do
98. 3:44you come back from that? Easy. You admit
99. 3:46you were wrong and build the thing
100. 3:48everyone actually wants. In 1997, Nvidia
101. 3:51released the Riva 128. This time,
102. 3:54triangles, fully compatible with
103. 3:56DirectX, and it was fast. Over a million
104. 3:59units flew off the shelves in just a few
105. 4:01months. Jensen has said that the company
106. 4:03survived because it was willing to
107. 4:05quickly kill its own bad idea. Lesson
108. 4:07learned. Being stubborn is cool, but
109. 4:10being stubborn about a failing product
110. 4:11is just ego. Then the momentum kicked
111. 4:14in. Nvidia went public in January 1999
112. 4:17at $12 a share. And later that year came
113. 4:20the product that changed everything, the
114. 4:22GeForce 256. Nvidia marketed it as the
115. 4:26world's first GPU, the graphics
116. 4:28processing unit.
117. 4:30Was it technically the first chip ever
118. 4:31to do graphics? Not exactly. But Nvidia
119. 4:34invented the term, put it on the box,
120. 4:37and basically owned the word. That's not
121. 4:39just engineering. That's marketing
122. 4:41genius. And the funniest part? Today
123. 4:44everybody says GPU. Everybody. Nvidia
124. 4:47basically named an entire industry.
125. 4:50Unreal. But of course, there was
126. 4:51competition.
127. 4:53The big dog back then was a company
128. 4:55called 3dfx.
129. 4:57Their Voodoo cards were the cool thing
130. 4:59every gamer wanted. Nvidia's strategy
131. 5:01was simple, release new chips on a
132. 5:03brutal 6-month cycle. Faster, faster,
133. 5:06faster. 3dfx couldn't keep up. They
134. 5:09stumbled, made some bad decisions, and
135. 5:12by 2000, 3dfx was done. Nvidia bought a
136. 5:16big chunk of its assets. Another rival,
137. 5:18ATI, hung around for years. Then AMD
138. 5:21bought ATI in 2006. So the GPU world
139. 5:25became basically a two-player game,
140. 5:27Nvidia versus AMD, and it still is.
141. 5:30Also in 2000, Nvidia landed a massive
142. 5:33deal to provide the graphics chip for
143. 5:34Microsoft's first Xbox.
144. 5:36So gamers everywhere were suddenly
145. 5:38running Nvidia technology without even
146. 5:40knowing it.
147. 5:41Quick pause. If you're enjoying the
148. 5:43story so far, subscribe to CEO Zoned so
149. 5:46you don't miss the next company
150. 5:47breakdown.
151. 5:49Okay, now here's where the story gets
152. 5:51really interesting because everything so
153. 5:53far was just warm up.
154. 5:55By the mid-2000s, Nvidia was a
155. 5:57successful gaming company. Great
156. 5:59revenue, great products, happy
157. 6:01investors, but Jensen was staring at
158. 6:03something weird. A GPU is made of
159. 6:06thousands of tiny cores doing lots of
160. 6:08simple math at the same time. That's
161. 6:10perfect for graphics, but it's also
162. 6:12perfect for science, medicine, finance,
163. 6:16weather simulation, anything that needs
164. 6:18a mountain of calculations. The problem?
165. 6:21Nobody could easily use a GPU for that.
166. 6:23You had to trick it by disguising your
167. 6:25math as graphics. It was painful. So,
168. 6:28Nvidia built a tool called CUDA,
169. 6:30released around 2007. It let normal
170. 6:33programmers write code for GPUs in a
171. 6:35normal way. Suddenly, this gaming chip
172. 6:38was a general-purpose supercomputer.
173. 6:40Sounds amazing, right? Wall Street hated
174. 6:42it. Why? Because CUDA cost a fortune.
175. 6:45Nvidia was putting this extra hardware
176. 6:48into every single chip, including the
177. 6:50cheap gaming ones, just so the software
178. 6:52would work everywhere.
179. 6:54That ate into profits. Investors were
180. 6:56basically like, "Why are you spending
181. 6:58billions on a hobby project?" Then the
182. 7:012008 financial crisis hit, and the stock
183. 7:04lost most of its value. Analysts said
184. 7:06Nvidia was wasting money.
185. 7:08Critics said it was distracted. And
186. 7:10Jensen, he doubled down
187. 7:13for years, even though almost nobody
188. 7:15could say what CUDA was actually for.
189. 7:18That's the part that makes you go,
190. 7:19"Wait. Why would anyone keep pouring
191. 7:22money into something that doesn't pay
192. 7:23off for a decade?" Because he saw
193. 7:25something nobody else did. He was
194. 7:27building a moat. Think about it. Every
195. 7:30developer who learned CUDA, every
196. 7:32university that taught it, every
197. 7:34research lab that built software on it,
198. 7:37all of them were now locked in. Even if
199. 7:39a competitor made a faster chip, you'd
200. 7:41have to rewrite everything. Nvidia
201. 7:43wasn't just selling chips, it was
202. 7:45selling a whole ecosystem.
203. 7:47That ecosystem is basically the reason
204. 7:50Nvidia is still untouchable today. Fast
205. 7:52forward to 2012. There's an annual
206. 7:55competition called ImageNet, where
207. 7:57researchers try to build software that
208. 7:59can recognize what's in a photo. A cat,
209. 8:01a car, a dog, whatever. For years,
210. 8:05progress was slow.
211. 8:06Then a team from the University of
212. 8:08Toronto showed up. Alex Krizhevsky, Ilya
213. 8:11Sutskever, and their professor Geoffrey
214. 8:13Hinton. They built a neural network
215. 8:15called AlexNet, and here's the part that
216. 8:17matters for our story. They trained it
217. 8:20on two Nvidia gaming graphics cards.
218. 8:22Two.
219. 8:23In a bedroom-level setup, AlexNet didn't
220. 8:26just win. It crushed the competition. It
221. 8:29beat the runner-up by a massive margin,
222. 8:31and the whole AI world went, "Wait, what
223. 8:34just happened?" What happened is that
224. 8:36neural networks needed insane amounts of
225. 8:38parallel math, and GPUs were built for
226. 8:41exactly that.
227. 8:43Gamers had accidentally funded the
228. 8:44hardware that AI needed. Jensen noticed
229. 8:47immediately. Nvidia pivoted hard.
230. 8:50Suddenly, the company wasn't just about
231. 8:52games anymore. Deep learning became the
232. 8:54main character. In 2016, Jensen
233. 8:57personally delivered Nvidia's first AI
234. 8:59supercomputer, the DGX-1, to a brand new
235. 9:03startup called OpenAI. Elon Musk was
236. 9:06there for it. Yes, that OpenAI. Remember
237. 9:09that for later.
238. 9:10Now, not everything was smooth, let's be
239. 9:12honest. Around 2017 and 2018, crypto
240. 9:16miners bought up GPUs like crazy,
241. 9:18because mining coins needed the same
242. 9:20kind of math. Gamers couldn't find
243. 9:22cards. Prices went insane. Nvidia made a
244. 9:25ton of money, and then crypto crashed,
245. 9:28and Nvidia was stuck with piles of
246. 9:29unsold inventory.
247. 9:31The stock took a beating. In 2018,
248. 9:33Nvidia launched RTX, graphics cards with
249. 9:36real-time ray tracing, which makes
250. 9:38lighting games look realistic. People
251. 9:41called it overpriced. They also called
252. 9:43it the future.
253. 9:44Both were kind of true. Then in 2020,
254. 9:47Nvidia announced it would buy ARM, the
255. 9:49chip design company behind basically
256. 9:51every smartphone for 40 billion dollars.
257. 9:54Regulators around the world said,
258. 9:56"Absolutely not." The deal collapsed in
259. 9:592022. So, by 2022, Nvidia was big, rich,
260. 10:04and successful. But, it was still mostly
261. 10:06known as the gaming company. And then
262. 10:08chat GPT dropped.
263. 10:10November 2022, OpenAI releases chat GPT.
264. 10:15Within days, the whole world is talking
265. 10:17about it. Within 2 months, 100 million
266. 10:19people are using it. And here's the
267. 10:21thing about AI chatbots, they run on
268. 10:24GPUs.
269. 10:25Thousands and thousands of them. Not
270. 10:27just to train the model, but every
271. 10:29single time someone types a question.
272. 10:31Who made basically all of those GPUs?
273. 10:34Say it with me, Nvidia. Their data
274. 10:37center chips, especially the H100,
275. 10:39became the most wanted piece of
276. 10:41technology on the planet. Tech giants
277. 10:43were begging for them. Startups were
278. 10:45using them as collateral for loans.
279. 10:47Countries wanted them. It was like the
280. 10:49entire planet was in a line outside one
281. 10:51single shop. In May 2023, Nvidia
282. 10:55reported earnings that blew everyone's
283. 10:56minds, and the stock jumped around 24%
284. 10:59in a day. By the end of that month, it
285. 11:01hit 1 trillion dollars in value. By
286. 11:04early 2024,
287. 11:052 trillion. Mid-2024, 3 trillion. And at
288. 11:09points, it was bigger than Microsoft and
289. 11:10Apple. And in October 2025, Nvidia
290. 11:14became the first company in history to
291. 11:16hit 5 trillion dollars. 5 trillion.
292. 11:19Remember, 40,000 bucks at that Denny's
293. 11:22table, and it's not even over. As of
294. 11:24October 2026, Nvidia's market cap is
295. 11:27hovering close to 6 trillion dollars,
296. 11:29with Nvidia saying it expects around a
297. 11:31trillion dollars in sales from its
298. 11:33newest chip families across 2026 and
299. 11:362027.
300. 11:37Jensen, meanwhile, is out here in the
301. 11:39same leather jacket being treated like a
302. 11:42rock star. He gets mobbed at events.
303. 11:44Fans ask him to sign things. For a chip
304. 11:47CEO, that's bonkers.
305. 11:49Here's the part nobody saw coming.
306. 11:52Gaming isn't even the main business
307. 11:53anymore. The data center side, the AI
308. 11:56side, is now the vast majority of
309. 11:58Nvidia's revenue.
310. 12:00Gamers built the company and AI made it
311. 12:02a giant. And Nvidia doesn't even make
312. 12:04its chips. It designs them. And a
313. 12:07company in Taiwan called TSMC
314. 12:09manufactures them. So, the most valuable
315. 12:11company on the planet depends on a few
316. 12:13factories on one island. That's a big
317. 12:15risk and it's part of why chip politics
318. 12:18is such a mess right now with export
319. 12:20rules and restrictions on what can be
320. 12:21sold to China becoming one of Nvidia's
321. 12:24biggest headaches. Plus, there are
322. 12:25rivals, AMD and even Nvidia's own
323. 12:29customers, like Google and Amazon,
324. 12:31building their own AI chips. Nobody
325. 12:34wants to depend on one company forever.
326. 12:36So, the big question is, can Nvidia stay
327. 12:39on top? Look at the whole timeline.
328. 12:41Nvidia started with a failed first
329. 12:43product. It nearly went bankrupt. It was
330. 12:45hated by Wall Street for betting on
331. 12:47software nobody understood. It got
332. 12:49mocked for being a gaming company for
333. 12:51years.
334. 12:52And every single time, the move that
335. 12:54saved it was the same one. Jensen made a
336. 12:57huge bet early and then stayed stubborn
337. 12:59about the vision while staying flexible
338. 13:01about the details. Killed the bad chip,
339. 13:04kept the company. Spent a decade on
340. 13:06CUDA, built a moat. Noticed AI before
341. 13:09everyone else. Owned the market.
342. 13:12So, the real story isn't a chip company
343. 13:14got lucky. It's that a team survived
344. 13:16long enough for luck to find them.
345. 13:19That's the story of Nvidia. The company
346. 13:21named after envy that ended up making
347. 13:23the whole world a little jealous.
348. 13:26If you want to see how another giant
349. 13:27went from nothing to everything, check
350. 13:29out the next video on screen. See you
351. 13:32there.