I had a decent amount I got into at 100 something a share back in the day, but I have to decided to go all in at 125 a share post split.
I sold everything - the amazon, the voo, every single asset I have was sold and dumped into nvidia. I will either retire in a few years or move under a bridge.
Jump in the discussion.
No email address required.
I bet against NVIDIA for Q2, thinking that surely the rest of the market would catch up because a 90% market share in as publicized/scrutinized a field as AI chips wasn't sustainable, especially with the fed investing so deeply into Intel.
Anyway that's why I'm an economist and not an investor, so my kid doesn't get ballet lessons this year.
Jump in the discussion.
No email address required.
I'm also bearish on NVIDIA. No way they keep their market share 5 years down the line. China is just starting to get its pants on but when that happens expect their market share to go down significantly, even if chinese chips are 30% slower and use 50% more energy as long as they are cheap enough compared to NVIDIA the developing world is just going to shift over to them.
Jump in the discussion.
No email address required.
Maybe if the chinks give the cards away for free
Jump in the discussion.
No email address required.
More options
Context
No AI company/startup is going to use anything other than Nvidia. Betting against Nvidia at least short to midterm is plainly uninformed.
Jump in the discussion.
No email address required.
More options
Context
More options
Context
Theyre completely entrenched right now due to an absurd software lead.
It will take at least 10 years for anyone to catch up.
Jump in the discussion.
No email address required.
How have you come to this 10 year software bet?
Jump in the discussion.
No email address required.
because cuda is leaps and bounds ahead of anything anyone else has lol. A friend of mine specifically works with this shit and says it takes 10 years to build something like cuda, and AMD isn't even close and they're the only ones trying to compete.
This isn't even getting into how entrenched they are. I don't see anyone dethroning them anytime soon.
Jump in the discussion.
No email address required.
Cuda is certainly better than opencl, but there are competitors. The real advantage is the overwhelming majority of data science software cowtools are very tightly coupled with cuda eg Pytorch. I pretty strongly disagree with anyone who says cuda isn't replaceable within 10 years.
Jump in the discussion.
No email address required.
More options
Context
More options
Context
More options
Context
More options
Context
Turns out you have to be pretty smart to make a quality GPU. How is anyone going to buy a better engineer with their 1% market share?
Jump in the discussion.
No email address required.
Talent is chasing big hardware. That's a massive fixed cost investment if you're looking to break into most AI markets, especially when your firm's relying on breakthroughs to justify the massive upfront investment. AI is similar to pharmaceuticals that way, and we can expect similar trajectories, with plenty of Elizabeth Smarts or Purdues cropping up.
However, AI markets aren't super vertical right now, so that big hardware is going to make its way into VC darlings and established giants in tangential tech sooner rather than later.
Jump in the discussion.
No email address required.
More options
Context
More options
Context
More options
Context