Skip to playerSkip to main content
  • 15 hours ago
Transcript
00:00You have exposure to AMD. It's added more than 17 billion dollars in market cap on one day alone.
00:05What does this deal signal? Yeah well I think it's a really exciting deal here for AMD. I think it
00:11validates their roadmap and their silicon and software strategy and I think it gives them a
00:17really big lead customer to really scale and I think that this will probably attract other
00:23customers to their roadmap and continue to build out Rockham. And I've always had this like thesis
00:28that the TAM is really big and if companies can deliver on that compute that there's there's a
00:36lot of companies that can win in the overall compute TAM. So overall very exciting and you know I think
00:43it's got a good setup for the next few years for the company's runway. I mean Lisa Sue just telling
00:48us that maybe she's being conservative when she puts the AI accelerator TAM at 500 billion dollars
00:52Tony. I'm interested though at OpenAI telling us Greg saying that we're looking at all kinds of
00:57ways of financing this. How are you feeling about this knitting together of purchasers and creators
01:05designers of chips? Yeah well I think that taking a step back over the last kind of 10 years everybody's
01:12kind of underestimated this TAM and I think what's exciting about AI is that there are a lot of
01:18productivity use cases here and we're starting to see AI agents really form and then I think physical AI.
01:24So I think as long as like the end use cases are delivering a lot of value and there's promise
01:29there I think that's what matters in terms of the ROI and I think a lot of these kind of contracts I
01:34mean they are backstopped by really blue chip companies like Microsoft or Oracle. So in terms of
01:40you know how I'm feeling about the financing I think it's it continues to be probably a good attractive
01:45way to do as long as the use cases are intact. Tony the mechanics of the deal are that OpenAI has the
01:52right to buy up to 160 million shares of AMD at a penny a piece but that right contingent on the
01:59operational and financial milestones that we discussed it is not circular financing it's something
02:06different but how comfortable are you with that mechanism? Well I think it can make a lot of sense
02:12right like I mean it puts the two companies kind of in co-development and partnership and I think
02:19there's shared economics it's a win-win for both companies I think some of the prices the warrants
02:24are struck at a pretty you know high price on the AMD stock price so I think it's good for both sides
02:31and if they win they win together so to me I think it can make sense. Does this alter AMD's position
02:40in the AI accelerator market in other words they're closing the gap a little with NVIDIA?
02:47Yeah well I think that both companies are doing really well and they're doing
02:51things a little bit differently but I think that you're alluding to kind of the market narrative
02:57of AMD was kind of in between kind of NVIDIA on the GPU side and then Broadcom on the custom side
03:04I think this can help lift the narrative to go from like a second place GPU player to a co-developed
03:11partner of OpenAI on the AI compute side so I think it positions them well gives them credibility scale
03:18and additional kind of co-development R&D to improve their roadmap so to me it definitely is a big positive
03:26and I don't think it's a zero-sum game I mean I think it just further signifies the demand for AI
03:33computing it's good for the country to build this much capacity because I do think the end payoffs are
03:39immense. I mean Tony we're at a new record high for the Nasdaq 100 at one point AMD was at a record high
03:45the valuations get ever higher when I'm looking at your holdings top holding Apple Alphabet NVIDIA
03:50Meta. Do you feel confident that these are going to continue to gain in value with this AI story?
03:57I think there's a good long-term story behind all our holdings and I think a lot of them are
04:03beneficiaries of AI they're strong companies that are great platforms and as a result I think that
04:10you know we are entering this inflection and AI driving a ton of productivity you think about
04:15you know output being productivity and labor I think productivity is going up and labor is going
04:21to be relatively uncapped I think as a limiting factor so to me I think it's exciting it's never
04:29been a better time to be a tech investor I think there's just so much change in dynamism
04:33so I'm excited for the multi-year here. Tony if you'll humor me please my column today was about
04:40all of the debt deals that are underpinning the AI infrastructure build out that we're seeing you
04:46know everyone seems to have a different comfort level with that as well how sanguine are you about
04:51the role of debt in what we're seeing? Yeah I think that you know the key signal here is you know
04:58what what is this AI capacity this AI infrastructure capacity worth and what's the useful life I think
05:05a few years ago we were debating is is the useful life of a GPU data center three or four years and
05:11I think that was just too short um you look at a lot of the GPUs you know from seven eight years ago
05:17they're still being fully utilized um because the AI workloads continue to evolve and then when you
05:22fully depreciate these assets like they're at a price point that could make sense for other uh LLMs
05:28be running on them and so I think that you know the useful life of these GPUs are definitely
05:33more than four or five years and um you know there's just more and more use cases you think
05:39about robotics like you know diagnostics um simulation uh there's just a ton more than we
05:46thought there were three years ago and I think that's what will keep the utilization high of these
05:51data centers as long as that you know continues I think it makes a lot of sense
05:55so
05:57you
05:58you
05:59you
06:00you
06:02you
06:03you
Be the first to comment
Add your comment

Recommended