00:00All eyes on tech companies, a strong earnings and AI spending continue to drive optimism here on Wall Street.
00:06But is NVIDIA's data surge sustainable or are we near as some are concerned about near peak AI chip demand?
00:13Longtime friend of ours, Gil Luria, joins us now, Managing Director at DA Davidson.
00:18Gil, thanks as always for your time.
00:20What have you seen so far among the tech giants here in this current earnings quarter?
00:27Well, the results themselves were as strong as you can imagine.
00:31So if we're talking about companies like Microsoft, Amazon, Google, they're having very strong results.
00:38Their core cloud business is accelerating.
00:41It's accelerating because of the demand for AI compute.
00:45And they're matching that with huge increases to investment in building out data center capacity so they can provide more
00:52AI compute.
00:53And of course, they're doing that mostly with NVIDIA, which reported more mind-blowing results this quarter and took a
01:02step further by guiding for the rest of the year, which they have not previously done.
01:07So if you look at the results at their face, they're very strong.
01:12Can't tell that from the stock prices, but the results being reported were very strong.
01:19Quarterly results are one thing.
01:21Guidance and forecasts for the future are another.
01:24Now, you just alluded to guidance for one.
01:27Overall, what has the forward-looking picture looked like for the tech giants?
01:31And might that explain some of the more erratic name that we've seen and more erratic ways we've seen the
01:36stocks perform?
01:37No, that's not quite it either.
01:39The company's guidance was also very healthy.
01:42Again, they have accelerated growth to these giant businesses.
01:47So Amazon Web Services, $140 billion business.
01:51Microsoft Azure, $120 billion business.
01:55Google Cloud, $70 billion business.
01:57Growing respectively at 24%, 39%, and 48% and still accelerating.
02:03Guidance was great.
02:04The issue is that investors are concerned about whether this is sustainable or not.
02:11All of these companies are now using all of their cash flow to build out data centers.
02:16And while they're telling us this is a good investment, we're selling these data center capacity at a markup, they're
02:24already pre-sold, investors are cautious because they want to see that happen before they believe the guidance.
02:33They want to make sure that, in fact, this massive investment in data center capacity and the AI compute capacity
02:40translates to strong returns.
02:44We have good early signs with the acceleration we described, but investors are skeptical because these companies have never spent
02:51all their cash flow on an investment.
02:53And so investors want to see that happen before they give these companies credit for that investment.
02:59And the way that translates to NVIDIA is the investment is on NVIDIA.
03:03So if investors don't think that investing in data centers is a sustainable use of cash flow for the large
03:11tech companies, they don't believe NVIDIA will continue to generate strong revenue.
03:17Bill, do you think we are near peak AI chip demand or still relatively early in the cycle?
03:25That is the concern.
03:27Now, we don't think that.
03:30When we look at what AI is doing and the progress that it's making and the compute resources that it
03:36consumes, we think these companies will have to continue to invest past this year.
03:42They won't be able to grow mathematically at the rate that they're growing investment this year, but they'll spend at
03:49least as much, if not more, next year and probably the year after that.
03:53And again, the reason is the demand is there.
03:56The demand for our use of ChatGPT and Gemini and Anthropic Cloud is there.
04:02The demand from all other companies to use their own AI tools, which comes across, again, just through these three
04:12companies, Microsoft, Amazon, and Google.
04:14That demand is so great.
04:16And the reason it's growing is our use of AI is changing.
04:19If a year ago, we just made an AI query that would consume one unit of compute, a year ago,
04:28we shifted to reasoning models for many of our compute resources, which then takes 10, sometimes 100 units of compute
04:37for one query.
04:38And now we're shifting to agentic tools, to bots, and those may be fire and forget multiple queries from now
04:50on for each one of those being 10 to 100 units of compute.
04:53So it's exponential growth for our need for compute, and we're not building data centers fast enough.
05:00That's why we believe that as long as the AI models keep getting better and they appear to be getting
05:05better, we're going to need a lot more compute.
05:09How are you viewing right now, the general valuations, conversations?
05:12I mean, obviously, your PE multiples have fallen for a lot of these names, Gil, but Microsoft is down more
05:18than 17% here in 2026.
05:21Tesla's kind of in its own category.
05:23Amazon, of course, had been the big underperformer of the big tech stocks last year, and it's not off to
05:29a great start, down about 11%.
05:30What has this done to at least maybe make these names a bit cheaper?
05:34Is there a better maybe entry for people who may have missed out on the boom the last few years
05:39to find some entry points into these names?
05:42Yeah, there's a very big divergence in performance, right?
05:46So Apple is trading at the highest multiple on the lowest growth rates because it's perceived as a safe bet,
05:54as a defensive bet.
05:56Then you have Google, which is perceived to be a winner in AI, which is why it's trading at the
06:01high end of its multiple range, in spite of the fact that it's not growing any faster than Microsoft, Amazon,
06:07or Meta.
06:08And then you have those three.
06:11Amazon's valuation is declining.
06:13It's still in the medium and higher end.
06:16But companies like Meta and Microsoft and NVIDIA are now trading at the very low end of their historic multiple
06:23range, in spite of the fact that their growth has accelerated.
06:28And they may be, especially Microsoft and NVIDIA, sitting on the best business they've ever had.
06:32So I would argue that that's where the opportunity is, companies like Microsoft, NVIDIA, and to some extent Meta, where
06:39Google and Apple may be on the higher end of the valuation and maybe the opportunity there isn't as big.
06:46Although with Google, I would argue they're going to do very well so they can justify that higher valuation.
06:53Gil, before I let you go, there's a lot of things working, a lot of cross-currents in these markets
06:58right now.
06:58As you and I are recording the conversation, the broader markets are down multiple percent as the issues in the
07:05Middle East continue to expand.
07:07That puts some drag on the tech stocks.
07:09You've got pricing pressures for discretionary spend.
07:12You've got the rotating AI concerns that work its way through software, commercial real estate, et cetera.
07:18What is on your radar that you'll be paying attention to for these big tech stocks, which in so many
07:22ways, these are the names responsible for getting us to these all-time highs that we've all enjoyed across the
07:27markets.
07:28The last few years, but a kind of a choppy terrain, maybe, and a guide map for the rest of
07:322026.
07:33What will you follow?
07:35Yeah, so there are a lot of cross-currents, right?
07:38The geopolitical situation, concerns about the overall growth of the economy.
07:43There's concerns about what a hugely successful AI would mean for the job market.
07:48There's a lot of worry in the market.
07:50And we obviously keep track of all those things.
07:53But the one thing we're singularly focused is actually on the capabilities of the AI models.
08:00Because as the capabilities of the AI models progress, again, we need people will use them more.
08:06We'll get a lot more value out of them.
08:08It means we'll need more compute.
08:10But it also means we will become more productive.
08:13The better the AI tools, the more productive people will get.
08:16And I know people hear that as, oh, if people are more productive, we need less people.
08:21That's not how the world has worked so far.
08:24Maybe this time is different.
08:26Usually that's a dangerous sentence to utter, right?
08:29So I believe that as people get more productive, we're actually going to need more people.
08:35Because if every person generates more revenue and more profit, we want to employ more people, not less.
08:41So that's why for us, the core of this is how good are the AI tools getting and how much
08:46do they make us more productive?
08:49All right.
08:50Gil Luria, longtime friend of the show, head of technology research at DA Davidson.
08:53Come back anytime.
08:54Gil, thanks for your expertise, as always.
Comments