00:00Michael Hunstad the president of Northern Trust Asset Management says investors may still be underestimating the impact of AI and
00:07he argues the technology could drive a major profit margin expansion or should I say continue to drive a major
00:15profit margin expansion. Seven out of the 11 industry groups in the S&P 500 have seen margins expand. Michael
00:21joins us now here on set. And the timing is great Michael because Jamie Dimon just said that they have
00:2850 to 60
00:29significant use cases for AI. And that seems like a lot especially if they're discreetly different use cases. You know
00:36not a surprise at all. Every company is going through this process right now of trying to make a list
00:42and sort by return on investment of what you know AI can bring overall to the financial picture. We're doing
00:49the same thing at Northern Trust. And I would say that we've unearthed a large number of opportunities. What are
00:55some what are some specific use cases that you
00:58guys have for AI and I know that you beat your earnings expectations. You're outperforming most financial services peers in
01:06terms of earnings growth margin expansion capital return. So you're doing very well. How are you specifically putting AI to
01:12use there. Yeah. So kind of three ways. Number one is the yeah we automate PowerPoint and Excel. Everybody's doing
01:18that. Number two though is process redesign around AI as the core. Not just at the margin but putting it
01:28as part of the
01:28the centerpiece of the centerpiece of the centerpiece of your process whether that's the back office process the reconciliation matching
01:35cash flows anything on the portfolio management side. There's a lot of efficiency gains within that as well. But I
01:42would say for us the most interesting application of AI is actually an alpha generation. How do you actually buy
01:48securities stocks bonds whatever the case might be. It's applicable across the board based on alternative data and AI models
01:56that can potentially give you an edge.
01:58And I think that's really where the interest is. There does seem to be something that shifted because Ken Griffin
02:02had long said AI won't give us alpha. He hasn't fully backtracked on that but he has been talking more
02:07kind of similarly to what you're saying that there are more uses of it in the investment process. Has something
02:13changed in just the past year that AI is better at eking out sources of alpha. Yeah absolutely. So think
02:19about your traditional investment management process security selection.
02:23Whether you're a fundamental analyst or a quant analyst. We use about maybe two or three percent of the data
02:31that a company actually generates. We look at cash flows weekly balance sheets income statements.
02:35But that's a tiny fraction of the overall data that a company actually produces. So that remaining 97 percent let's
02:43call it that. If you can use that to your advantage if it has any edge in it then that
02:49gives you a significant differentiator in terms of generating alpha. And there absolutely is edge embedded in that data. I
02:56would say it's two dimensions though. One dimension is the algorithms that unlock that edge. And I would say you
03:01know Ken Griffin's comments probably more in that direction.
03:05The second is the second is the data. Getting access to this massive amount of information that we could only
03:11dream of a few years ago and exploit that for hey you know what do I want to buy. What
03:16do I want to sell. We're just on the cusp of doing that. We're doing some really really interesting things.
03:22But as an economy we have a whole new sandbox in which to play.
03:26It's not. I mean it's not the cheapest though not the talent but the tokens and silicon data has some
03:33data of just how much how much more expensive tokens are getting.
03:36There's this idea that leading with Google that they really subsidize it and price things really cheaply. But at some
03:42point this industry is going to have to start passing off the cost. According to their data the price of
03:46tokens has doubled in just the past year.
03:49How are this is the price per million tokens. How are you handling that the rising price of just how
03:55expensive compute and inference is.
03:57I would say it's a balance between on-prem and in the cloud so it's sort of token based. Now
04:03that being said you know from our perspective even at higher prices there's still a huge ROI to some of
04:09these AI projects especially in the alpha generation side.
04:12So you know we haven't really come close to a cautionary red flag for many of our projects to say
04:20hey it's too expensive but it is certainly a consideration.
04:23We see a lot of firms going back or maintaining their on-prem computing facilities as a result.
04:29I want to get your take on the mega IPOs. Obviously it's huge to your business. It's huge to your
04:35clients your investors. SpaceX open AI anthropic. We could be looking at what four trillion dollars worth of IPOs here.
04:47And David Rubenstein says that has unleashed animal spirits. It's a huge part of the reason we've run up to
04:53record high after record high. How do you. What's your take.
04:55Yeah I mean IPOs are always a positive sign to the economy as a whole. The fact that they will
05:01enter probably the larger mid cap or part of the index almost immediately garners a lot of interest as well.
05:08That even a passive investor will have to be invested in these IPOs as well at some at some point.
05:13So they're getting a lot of attention getting a lot of headlines.
05:16I would say they're reinvigorating also interest in AI in the smaller cap end of the spectrum as well as
05:23in the private markets. So that I don't want to call it euphoria. It's not a euphoria.
05:27I think it's a reality because the earnings are there behind it. But that enthusiasm. Well there's no earnings at
05:33SpaceX. Well there's billions of dollars of losses.
05:36But technology as a whole when you look at earnings expectations for the next year more than 40 percent in
05:43the S&P 500. That is a phenomenal number.
05:45That's something we haven't seen in a long time. There is plenty of research out there. BCA has a piece
05:49that just historically when the market has to absorb all of the supply of IPOs the S&P tends to
05:53underperform.
05:54And there's just uncomfortable historical corollaries of tops coming when you get all these IPO issuances. Does any do any
06:00of those aspects give you pause?
06:01I would say that they would except for the fact that the macro and earnings backdrop are so positive in
06:07our favor. From a macro perspective we have very good very solid real economic growth.
06:13Yes inflation is high but that's more of an external exogenous source. Earnings I already mentioned it. S&P forward
06:21for 2026 22 percent.
06:23And into next year more than 15 percent. So when you look at it at a sector level tech comm
06:29services very very healthy materials very very healthy.
06:33So there's a lot of reason to be optimistic about the equity markets going forward. We are risk on in
06:38our portfolios. We're tilted very heavily into equity especially into that tech space because of the fact that we just
06:46see such a great earnings picture and expanding margins as well.
06:49I finally want to ask you about the Fed because you mentioned inflation right and obviously with huge profit growth
06:55and economic growth typically we could get rate hikes in a market that had been until recently expecting cuts.
07:04Yeah I think the AI productivity consideration is one that is not well understood from a monetary policy perspective in
07:14that if you think about the potential for AI productivity gains.
07:17Now the numbers out there range from zero to 50 percent productivity gain.
07:23I'd say the most objective there was a paper released by the Federal Reserve last year that put that range
07:28somewhere between 15 and 30 percent.
07:30Now not all of that productivity is going to accrue to lower number of hours worked or higher wages but
07:38if even a fraction of that materializes and one of the other considerations I think is really important as we
07:44listen to all these earnings calls.
07:45We obviously had a great earnings calls. We obviously had a great earnings season last quarter. As we listen to
07:50all these earnings calls everyone is talking about the ROI of AI.
07:55Like what do they invest in and what is the potential productivity gains as a result of that. Again if
08:00even a fraction of that materializes that will have a very very stimulative effect on the economy.
08:06My perspective is that in the face of this significant positive supply shock. You know one of two things can
08:14happen. It can be deflationary if the economy thinks that it's more transitory, more temporary.
08:19My view is that it accrues positively to economic growth going forward. So I think that there is a big
08:27upward movement in growth as a result of this positive supply shock at some point.
08:32So what do you do from monetary policy perspective in that face. Well you tend to be more restrictive. There's
08:38going to be more stimulus behind it.
08:40You don't want to cut rates in the face of this positive supply shock. That never made any sense. But
08:45there's this tradeoff between short term inflation, longer term growth.
08:49I think AI productivity comes out on a longer term growth side.
Comments