AI Expectations: Nvidia Earnings & California Bill
Key Points
- Nvidia posted record‑breaking year‑over‑year revenue growth and beat its own earnings outlook, yet its shares fell because analysts had set even higher expectations for future chip demand.
- The market’s focus on Nvidia’s ability to exceed aggressive forecasts underscores how AI “expectation games” are driving stock valuations more than raw performance.
- California’s newly passed SB 1047 mandates transparency and safety reporting for frontier AI models costing over $100 million, aiming to prevent catastrophic “rogue AI” incidents that could cripple critical infrastructure.
- Major tech firms (Google, Meta, Microsoft, OpenAI) oppose the bill while Elon Musk and Anthropic support it after amendments, leaving its ultimate fate pending Governor Gavin Newsom’s decision.
Full Transcript
# AI Expectations: Nvidia Earnings & California Bill **Source:** [https://www.youtube.com/watch?v=4W-uIjWqurE](https://www.youtube.com/watch?v=4W-uIjWqurE) **Duration:** 00:05:31 ## Summary - Nvidia posted record‑breaking year‑over‑year revenue growth and beat its own earnings outlook, yet its shares fell because analysts had set even higher expectations for future chip demand. - The market’s focus on Nvidia’s ability to exceed aggressive forecasts underscores how AI “expectation games” are driving stock valuations more than raw performance. - California’s newly passed SB 1047 mandates transparency and safety reporting for frontier AI models costing over $100 million, aiming to prevent catastrophic “rogue AI” incidents that could cripple critical infrastructure. - Major tech firms (Google, Meta, Microsoft, OpenAI) oppose the bill while Elon Musk and Anthropic support it after amendments, leaving its ultimate fate pending Governor Gavin Newsom’s decision. ## Sections - [00:00:00](https://www.youtube.com/watch?v=4W-uIjWqurE&t=0s) **Nvidia Earnings Defy Expectations** - Despite beating forecasts and posting roughly 120% year‑over‑year revenue growth, Nvidia’s stock dropped because investors’ ultra‑high expectations for AI chip demand weren’t fully met, underscoring how market expectations shape perceived success. ## Full Transcript
real quick we have two pieces of AI news
that basically resolve down to
expectation games in AI number one is
nvidia's earnings yesterday so Nvidia
published earnings after the close of
the markets on August 28th and everyone
would think that you'd be happy about it
they beat on the earnings projection
they'd given they Advanced the earnings
projection for next quarter ahead of
what the street had expected they
delivered phenomenal year-over-year
growth I think it was 120% or close to
in Revenue absolutely amazing and the
streets still cut the stock price after
hours so was trading down like 4% after
hours why is that because expectations
are through the roof and the expectation
game is what matters sometimes on the
street it's your ability to Beat the
Streets forecast by a certain amount
that they care about and Nvidia did not
do that and why that matters is because
Nvidia is something that street is
watching to see if AI is continuing to
accelerate because they're looking for
whether chip demand is continuing to
grow and so even though the street
provided a very aggressive forecast for
upcoming growth so we can still see sort
of that they are expecting more chips to
be
sold the street wanted more and more and
more and didn't get it expectations are
shaping the financial reality now
nvidia's employees are still doing just
fine I read a report that one and three
of them is worth $20 million so you know
let's not let's not rush to to weep just
yet number two contentious California
bill so s SP
1047 is a bill for AI safety that just
passed the California state legislature
and is now headed to Governor nome's
desk the goal here it's kind of like
sock 2 if you're familiar with B2B uh
software and compliance you basically
need to be very clear and public and
transparent if were building Frontier
models spending more than $100 million
on a Model what you are going to do
about safety for that model but what's
interesting to me in the same
expectations game setting space is that
at the end of the
day what they are trying to protect from
is a lot like Terminator is the Skynet
scenario where AI goes Rogue AI takes
over critical infrastructure systems AI
controls Something That Matters to
theity or the county or the state and
does hundreds of millions of dollars in
Damages you can imagine AI shutting down
an airport AI shutting down a hospital
those are the kinds of things that this
bill is worried about and trying to
preclude through essentially sock 2
style process and Reporting
standards now this has been widely
opposed so Google opposes it meta
opposes it Microsoft and open AI oppose
it uh Elon Musk has supported it and
anthropic has has supported it after
some amendments to relax it a little
bit we will see where it ends up it's
not clear if Gavin Nome will approve it
or not we shall find out the thing that
matters though is that they are only
writing the Bill as far as I can tell
about a particular disaster scenario and
trying to preclude that and we have not
yet seen evidence that we can
actually see at attack vectors from llms
for the disaster scenario they're
describing and they are saying as
advocates of the bill like the state
legislature uh members who support the
bill are saying we don't have to see it
we want to preclude it because it's so
disastrous if it happens I would
argue that we need to be thinking
holistically about AI risk and not just
assume that it's going to look like
Skynet we need to think about AI risk in
terms of disinformation for example
because we are starting to see that in
the wild today people imitating other
people's voices with AI people who are
able to gain access or unauthorized
access to accounts or property with AI
those kinds of things we are seeing
happen in the wild disinformation we're
seeing happen in the wild
today but we're not actually seeing the
attack vectors that this law is designed
to to go after and so you know part of
me wants to say maybe we should focus on
the harms that we're seeing right in
front of us and making sure that we have
really clear standards and guard rails
around that and maybe we should think
carefully about where we expect llms to
go before we rush to craft
laws about imagined harms that we've
seen in Sci-Fi movies but haven't really
seen evidence of here in
reality because the llms we're getting
are not really like the llms in sci-fi
they're not these mechanical sideboard
creatures they're very fallible they're
very flawed they remind us a lot of
people and
I just don't know if that mental model
is correct so what do you think those
are the two pieces I thought both of
them you know the Nvidia one and the S
SP 10471 ties into this mental
expectation game we have in AI right now
and expectations shape our reality when
we collectively act on our expectations
we are shaping what actually happens
that bill could become law the Nvidia
market cap is you know 3 4% lower right
now what are your expectations and are
they in line with reality