We've had that historically, and that's our commitment going forward. We invest organically in the business first to drive outside growth and margin expansion. And we'll talk briefly on the next page in terms of capital allocation. But we've been acquisitive over the years, and we intend to continue to be acquisitive. And then third, we're going through a massive, and I use the word massive, enterprise-wide business transformation. We're more than halfway complete. It's a six-year-plus journey. We're digitalizing every part of the business. I'll leave you with this: we're a Fortune 200 company. When we're done, brand new tech stack. Every system we're operating on, every system, will be new to the company. We're more than halfway complete through this journey. We launched this in earnest six months post the closing of acquiring Anixter back in the middle of 2020.
This is going to be absolutely transformational for our business. We think it'll position us as we unlock the power of our big data in one world-class data lake to accelerate our top-line growth, dramatically expand our margins, and also increase our speed to value for future acquisitions. Quickly on cash flow. If you look at the right-hand side of this page, you can see our average annual cash flow for 10 years pre-Anixter acquisition. We closed with them in 2020. We've averaged over all parts of the economic cycle roughly 100% of net income is converted to free cash flow. Our guide for this year is $800 million-$1 billion. To give you a few data points, the second half of last year, we generated $400 million of free cash flow. First half of this year, $500 million of free cash flow.
We just generated $280 million in our third quarter of this year, so well on track to deliver the guide for this year. We committed to a cumulative target of $3 billion over 2025 through 2027. The drivers are shown on the left of this page. Finally, in terms of capital allocation, we're going to direct 75% of our free cash flow, that $3 billion plus, to M&A. We've been acquisitive over the years. We're going to continue to be acquisitive. We've played a major role in industry consolidation. We expect that to continue, as well as focusing our acquisitions on services and adding services to our company portfolio, and I know we'll talk more about that. At least 25% of the allocation to a consistent buyback program, and we did initiate a dividend two years ago. We increased it 10% after the first year.
And we expect to increase that over time as well. To the extent that the M&A opportunities aren't manifesting themselves, because you can't always time those perfectly, we would direct the cash flow to buybacks and debt paydown. To give you a sense, through three quarters of this year, we've paid down $475 million of debt, and we bought back $375 million of stock through three quarters. So with that, Dave, I'll open it up.
All right. Session1@rwbaird.com. If you have any questions. Trimble, we don't want to do that. My first question for you, John, is we're asking about the interest rate situation. I'll reframe it a little bit to apply to your business. The economic environment, we're in this period I'm calling stall speed. I mean, when they put up GDP and it's looking at 2.8%, I'm asking myself, where is that? It's not on my coverage list, I guarantee you that. And with a, what is it, 40% of your business directly in sort of that EES, that more of an industrial angle, could you talk about the, as we sit here today, your outlook for 2025? And just thinking about relative to the overall economy, I mean, can we see some growth next year? New administration, potentially lower rates. Maybe we get a little bit of growth next year.
Can you talk about how you feel about it?
Yeah. We've not given our formal guidance for 2025 yet. We typically do that with our Q4 earnings release. With that said, we've outlined a bit of the framework on how we're thinking about it. And look, in this year, we clearly have seen what you said, Dave, which kind of stall speed. I mean, I guess the good news is we didn't really experience any deep recession. Soft landing appears to be underway. But with that said, we did not deliver the growth we expected this year. As we got through the third quarter, we have three business units. It's very good to see one returning to growth, our CSS business, on the back of very strong AI-driven data center growth. The utility business, obviously, has had some headwinds through the year, and broadband is part of that, coupled with utility and electrical as well.
Had a series of kind of minor tailwinds, but a lot of headwinds as well. As we look forward to 2025, we expect all three businesses to grow. Again, we're not going to give the outlook today, but I think what we'll see in 2025, as the first year of a multi-year process, I'd leave you with this: the secular growth trends we see in our value chain, our served markets, are going to begin to really rule the day. That was my view pre-election. I'll come back to that, but clearly, when these secular trends that many companies talk about, whether it's electrification, IoT applications, automation, AI-driven data centers, GenAI-driven data centers, nearshoring, reshoring, all of that requires one thing: increased power generation. And electricity demand in the U.S. has been flat in 2024, back to 2007 levels, so you're talking about 17 years we're at the same level.
We're now facing a rising power demand curve. And that's required for anyone, let alone collectively, all these secular growth trends. So I think those trends start to rule today. Now, when you factor in the election outcome, I think one thing can be said: this nearshoring, reshoring secular growth trend was underway pre-election. I'll digress for a moment. We sell to 90% of the Fortune 100 companies directly. They're all looking at reshoring, nearshoring to varying degrees. It's because of what happened in the pandemic. It put an electron microscope on the global extended supply chains. And in many cases, these companies have supply chains that got single-threaded. And no one ever thought about that black swan event happening. It's raised the priority of supply chain not only to a C-suite and board level, but to a country level.
And so nearshoring, reshoring was absolutely underway, even though you don't hear a lot of companies talking about it. They don't admit it. They don't want to talk about it. But this is absolutely a secular trend that started pre-election. With the administration that won, I think you're going to see an amplification of that. And so clearly, there'll be a focus on USMCA markets with U.S. at the top of that, which should result in even more support for, I'll call it, captive U.S. investment, which should bode well for those companies that have a large infrastructure base in the U.S. So I think, again, I'll end on that note. The secular trends will rule the day. If they are secular, they're multi-year in nature, and it's not one or two years. I think we start to see that. I think we're already seeing it in data centers.
But I think you see that spread to other applications as we move into 2025.
Yeah. We have heard some of the companies at this conference start to talk a little bit more about the onshoring that they are seeing some. We talk about chip fabs and battery plants and things like that. We are starting to see some.
New semiconductor wafer fab plants, that hadn't been done in three decades in the U.S. And you have more than one now, right? Again, needs power. But I think, again, I think it's just there's a lot of examples. Look, it's interesting. I've been in B2B distribution for two decades, all with WESCO. Prior to that, in different end markets at engineering manufacturing companies. My whole career, 40-plus years, the businesses I worked in, again, seven or eight different end markets, everything was being outsourced south of the border to Southeast Asia. And just think about that over the last several decades. So this didn't happen overnight. I think there's a fundamental shift. The knife-edge switch has been thrown. And again, that's independent of the election outcome. And I think, look, here we go. There's an opportunity for a reindustrialization of America.
It'll have a totally different mix, higher electronic content, but we're at the front end of that potentially, clearly.
Right. Some analysts at Baird gauged it at $15 billion over 10 years. Maybe that's right, maybe not. It's a good guess anyway. One other thing I wanted to ask you, John, was that we're hearing a lot about is, of course, tariffs. It's related to what we're talking about. But in the near term to medium term, if there are any tariffs enacted, and maybe reflect back on what happened last time, last time around, you think, for the company.
I think, I mean, obviously, this is a hot topic now because it's expected that there will be increased tariffs, or at least it'll be used as a lever. As a first derivative effect, it's a rounding error. We don't have a meaningful, in terms of size, private label business. As a distributor, if you've got a meaningful private label business, you're going to have to deal with this directly. So where do we see it? We'll see it with our supplier partners. Again, we're a business-to-business distributor between customers and suppliers, and for most of our supplier partners, we're their biggest customer, a channel to market, so clearly, they're going to, depending on their mix of their products and where they manufacture, they're going to have to deal with that.
Now, I'd say the good news in this is, at least in core electrical, and even for, I'll call it, fiber optics and data com, like a Corning, many of our supplier partners have been investing in the U.S. and stepped up those investments meaningfully as we came through the pandemic. And so there's some investments that are already made with increased capacity in the U.S., and there's other in-flight investments underway. So I think that gives them the opportunity, Dave, depending on the supplier, what the tariff hits for them to continue to leverage and pivot even more to kind of a U.S. or a USMCA manufacturing base. But clearly, look, again, we'll have to see the mix is going to matter. What's addressed to what degree? The devil's in the details.
And I guess if you're saying it's a secondary or tertiary effect, that means that the manufacturer would incorporate whatever that tariff is into the price of their product, which leads to inflation, which has obviously historically been a positive for distributors.
So to your point, I mean, if we have a reignition of inflation, we're not wishing that on anyone. But to the extent inflation occurs and it's broad-based, that's good for distribution. And I think what might be interesting this time is, to the extent that happens, how does that play against the demand environment? And if you believe these secular trends are really starting to build steam, it'll be interesting to see that dynamic. It's going to vary by value chain. But net, net, Dave, to your point, right now our view, because you can't do anything else, I mean, we saw about 1%-2% of price this year. We get a good insight into our suppliers' anticipated or planned price increases. We get a three- to six-plus-month lookout. Number of SKUs and magnitude of price increase relatively stable with our forward-looking view.
We're 1%-2% kind of realization this year. That's our kind of base case. But to the extent it starts stepping up, Dave, to your point, we know how to manage in that environment. It's our suppliers that will want to pass that through. We'd work with them. And hopefully, we'll see what the market bears. Some markets will bear it more than others, again, based on demand profile.
All right. That all sounds very good. Thanks for sharing that information. Maybe we could talk about, you mentioned capital deployment and M&A that you discussed at the investor day. One of the things I talked with Dave, your CFO there at the conference, was it sounds like you're looking at more like the Ascent deal or like a Rahi, where you've got maybe more of a service element, or you're trying to skew the business toward higher margin opportunities. But I think with that would typically come higher acquisition multiples. So my question is, how are you going to strike the balance there? And can you continue to move returns on invested capital higher with higher margins if you're paying more for the deals? Can you talk about how you bridge that gap?
Scott just framed up the page from our investor day. Thank you for that, Scott. As we think about our M&A priorities going forward, first, I'll say we historically have been very acquisitive. I think some of you know Wesco. Some of you don't. But Wesco spun out of Westinghouse in 1994. We're talking 30 years ago. It was $1.5 billion in sales and losing money. We've done over 50 acquisitions since 1994, over 30 under my watch. I would say that the acquisitions, at least over the last 5-10 years, have been focused on strengthening our category positions and building out more scale. Anixter was actually game-changing because we took two equal-sized Fortune 500 companies and put them together and did that in the beginning of the pandemic and successfully integrated those. That was a big scale move.
So that's still on the table, Dave. Look, any deal that's getting done, we get a call. And many times we get last call. And sometimes we get only call. But anything that's happening, we will look at. Some we're choosing to put into play. So I wouldn't rule that out. But to the extent we lean in that direction, it does take us out of the market for a while to do the other priority of acquisitions, which is services. And I think probably the best example of what we've done organically in the business, and it was organically, was our utility business when I joined 20 years ago was sub-5% EBITDA. Now it's our highest profitable business because we're much higher on the service value proposition, direct end user customer. We're more than products. It's services, supply chain solution management, and includes some fee for service-based business models.
So we added services organically at a relatively slow rate over the last decade and a half plus. Now we want to use our balance sheet to inject services quicker. The most recent example is post-Anixter, the first deal that we did was Rahi Systems two years ago, November 2022. Wasn't that large a company. But that put us in a great place for a more complete data center solution. It added services. And they had a rack and roll program, which is not just products. They would build out the server rack, test it, qualify it, certify it, and support global deployment of direct end user customers. So we climbed the value chain a bit, coupled it with legacy Anixter's data center position. And it's the basis upon which we're growing our data center business that went from teens growth in Q2 to we're north of 40% growth.
Our data center business grew in Q3. It's a $2 billion of our $22 billion plus run rate and accelerating. So we want to tuck in these acquisitions that are services and use that as we cross-sell with our end user customer base and kind of expand the scope. We announced a deal the day of earnings, after earnings, called Ascent, another data center services company. This has no products. It's not a product distributor. It's pure play services. It includes an on-site staffing model helping manage the operations of the data center. And so we're exceptionally well positioned now to kind of participate across the lifecycle of the data center. One thing that's not getting written about quite enough, quite frankly, is when you think about this mega data center growth, this secular growth. Everyone's talking about building these new data centers, right? 800 megawatts, gigawatt mega data centers.
With GenAI, you're going to see a significant amount of investment around upgrading current data centers, and you're talking about an order of magnitude in terms of power density, and so if you're in there working the whole lifecycle of the data center, even better positioned to participate in those upgrade or retrofit projects, so it's services. We're willing to pay a little higher multiple if it's got higher margins and if the growth rate, if the equation works such that we get the payback quickly. That was the case with Rahi, and we expect that to be the case with Ascent. It's been running a 30% CAGR for the last three years, so again, we're willing to pay up a little bit, again, with the eye on the prize being delivering the outsized growth tied with the secular growth trends.
But to your point, Dave, this balancing act is going to be delicate. We want to still use our balance sheet right. It's the biggest balance sheet, strongest balance sheet we ever had. We are seeing the power of scale in our business. I mean, WESCO and Anixter coming together is, I know it's textbook, a case in point of what scale can do. And so there may be some of those opportunities. You can't time those right. You got to be positioned. Our pipeline now, I say, is very robust in terms of a number of different types of deals. And many of them have high services content.
Okay. Any questions from the audience? Let's think about just before we leave the Ascent acquisition. I wanted to ask you if you could. I don't pretend to know exactly how data centers are run and this sort of thing. Could you talk about Rahi and Ascent and how that fits into the data center operations? And importantly, I think there's been some questions that I've gotten from other industry players that have asked, is there any kind of channel conflict potential with these sorts of deals relative to your other potential customers that you'd be selling to in the data center arena? Again, I don't know, but maybe you can explain it to us.
I'd say the data center position started with the legacy Anixter business. They served data centers either directly as an end user customer. This would be whether it's a hyperscaler, one of the big global hyperscale players, or an MTDC, a multi-tenant data center. That's a company that builds a data center and then sells out the capacity to others to use that compute capacity, or enterprise-class data center, which is a customer that has their own captive data center on site dedicated to them. Those are the end user relationships that Anixter had when we acquired them, merged with them, integrated them. They also went to market through specialty integrators. There's a number of specialty integrators that also serve those end users. It's very interesting when you map out the whole channel flow. There's direct end users.
And then there's these specialty integrators that will also do projects for data center customers, not unlike what can be done with an industrial customer with electrical at some times, where you have a specialty contractor or you do a project direct for them. So Anixter's blue chip relationship with the end users was something that we really coveted because Wesco's data com business didn't really have those end users for data centers. So that's the structure of the industry. And Anixter, by and large, was a product-based distributor for the white space of data centers. The white space, that is, under HVAC control. Again, we're not Ingram Micro or Tech Data that merged with SYNNEX. We don't provide the switches or the equipment that goes in the racks.
But everything else that's in the data center, the racking, the fiber optic connectivity, the thermal solution, IP cameras that are in the room, how it ties into the infrastructure, we're heavily involved at the design level. So Anixter has a captive engineering team that works in conjunction with its end customers to help design and spec in the connectivity solutions, let's say. So that's how it works. Now, what did Rahi do? Rahi gave some additional services to the Anixter portfolio, but also this rack and roll capability where they would build these racks, fill up, test them, certify them that they work to a spec, and then would help globally deploy them to that hyperscaler or MTDC or enterprise-class customer. So it climbed the value chain a little bit. What does Ascent do? Ascent has no product sales. They're in there with the data centers operating.
They have an interesting software package that helps monitor the KPIs of the data center, can give early warning indicators if there's issues. They help support maintenance and other upgrades. They also have part of that software will manage some of the capital projects and ensure they're getting ROI as upgrades are made to the data center. So they're embedded in the operations of the data center. So now we have more of a full lifecycle view. And again, as I mentioned, I think it's going to be very interesting as we go forward because obviously new data center builds, these big mega data centers are great opportunities. But the retrofit, renovation, and upgrade of data centers represents another great, I'll call it demand stream. Pre-AI, the half-life of a data center was 4.x years. What does that mean?
They would take everything in the rack, change it out because again, Moore's Law has not reached its limit, so okay, there would be good payback to basically increase the capacity and speed and compute power of that data center. With GenAI, it dramatically reduces that half-life, right? And there's massive investment and upgrades required to go, which includes much more exotic cooling solutions, a la liquid cooling for GenAI rise because of the power demand. So it's just a holistic look at that value chain, Dave. We're so much more than a product-based distribution company. The services and in looking at the whole lifecycle, this is why at our investor day recently, we didn't have each of the three businesses present. We picked two major topics: our digital transformation, which we gave a lot more detail than we had previously on that publicly, and data centers.
I'd encourage you to listen to that replay.
In the last three minutes here, as long as we're on the data center thing, which I think is super important and a really interesting part of the business, could you talk about the opportunity? You talked about the white space. Talk about the gray space, given that that's kind of legacy WESCO, it's where you reside.
It's where we reside, but it's interesting because if you look at how that industry grew up, and these giant hyperscalers were at one time small tech companies, and they've been managing their own hyper growth, no pun intended. The electrical switchgear for the gray space of data centers historically has been provided direct from manufacturers to the end user. So the Eatons of the world, Schneider Electrics of the world, our big supplier partner, Siemens of the world, they've been providing that direct. As a structural matter, that's how the industry worked. We are now picking up some of those projects. We have a much more complete solution under one roof, the combined Wesco-Anixter and our direct end user relationships because now they're trying to execute globally to a set of design standards and deploy globally. They all have massive spend budgets.
These are the big companies we're talking about are Magnificent Seven Plus. You see what their annual investments are. They're trying to execute globally with these data center buildouts. There's no one else who has the global execution capability. That's a value-added distribution function. Anixter built it organically over two decades. They were in 53 countries around the world when we bought them. No one else has it today. So they are now looking at what they used to do inside their four walls. This was much more complex. And they're looking for, quite frankly, one outsourced partner to help manage that for them. And we're filling that bill. Yes, they want one throat to choke. But I think it's an opportunity for us to pick up gray space. And then the final point, which will be the ultimate determinant, quite frankly, is power.
The gray space is our EES business. The white space is our CSS business. The power is our UBS business. And at the end of the day, you need to have the power solution or you can't even begin to think about where to site the data center. And that's instrumental to the siting. Here's a stat. And there's a lot of estimates you'll see. Total data center power demand in the U.S. this year is about 25 gigawatts. Estimates of that'll be 80-100 gigawatts in 2030. So you're talking north of 10 gigawatts per year added if it was linear. One gigawatt's the equivalent of one new nuclear power plant. Think about that. So this is why I've said for quite some time that the utility industry, which was a cyclical GDP industry, is fundamentally shifting to secular growth.
You're not going to be able to talk about data centers without talking about power. But we're uniquely positioned for that too. And I do think back to utility, I think we're entering, again, electricity demand has not grown since 2007, but now we're facing this rising power demand curve. So very exciting opportunity.
Great. Well, thanks for the download on data centers. That was great. And so what I'm hearing you say is you expect growth next year. Hopefully with that, we can get some margin expansion, 10 basis points, $0.30 per share. It's got a lot of operational leverage. And in a better economic environment, I think this can be a really interesting story.
Thank you, Dave.
Thank you, John and Scott. Appreciate it.
Thank you.
The guys will be available at the Salon C breakout session.