← 返回

2026.10.08 | TB | Cloud | Macquarie Technology: Cloud, Government & AI

原标题：Macquarie Technology Group – Cloud & Government Business Growth Outlook & AI Workload Commercialisation

#cloud #datacenter

Third Bridge · 2026-10-08

Analyst-led • 21 Sep 2026

Companies: MAQ AU, AMZN US, AVGO US

Robert Hanimyan (RH) Director, Digital at Brennan IT Pty Ltd

Joanna Chan (JC) — Third Bridge Sector Analyst

Agenda

Key Insights

Introduction

JC [00:00:03]: Welcome to Third Bridge's Interview entitled Macquarie Technology Group – Cloud & Government Business Growth Outlook & AI Workload Commercialisation. I'm Joanna Chan, and I'll be facilitating today's Interview with Mr Robert Hanimyan, former COO at Macquarie Technology Group Ltd.

Robert, before we start today's Interview, please state I agree or I disagree to the following statement: You understand the definition of material non-public information and agree not to disclose any such information, or any other information which is confidential, during this Interview.

RH [00:00:39]: I agree.

JC [00:00:40]: Thank you, Robert. If you would, could you begin with a brief introduction to your background please?

RH [00:00:48]: I'm Robert Hanimyan. I've been in the tech industry in Australia for the past 28 years. The relevant role for this conversation, I was the Chief Operating Officer for Macquarie Technology Group. I was the Chief Operating Officer for the Government division. I was there for just shy of two years.

Questions

00:00:50] In its FY26 results, Macquarie Technology Group's CS&G [Cloud Services and Government] segment delivered quite steady expansion over the past two years, with revenue and EBITDA growing about 11% and 5.5% YoY, respectively. How do you expect the segment's revenue and EBITDA growth to evolve over the next 2-3 years? What will be the primary drivers?

RH [00:01:40]: I think the thing to understand with Macquarie Technology Group is there are four distinct divisions that generate revenue in very different ways, but at the core of every one of those divisions is the investment that Macquarie has in data centres. From a revenue perspective, there is a constant base coming through their data centres. As you've seen in the news, they are building the first AI data centre, or at least that's what it was when they launched it, and expanding their data centre footprint. They've also, I think, just announced the launch of a new data centre. There'll be elements of that. What they're also going to look to do, I expect, is invest into new technologies to diversify the revenue that they are generating through their government customers. One of the core revenue line items for the Government division, which they do talk about, is their Secure Internet Gateway, or SIG. From a government perspective, SIGs are no longer being requested. The Government division has just launched their Azure storage platform, which, in conjunction with Microsoft, is a service that Macquarie builds and sells.

I expect they will continue to build out their SASE play. They've partnered with a number of SASE providers. The third part, which is obviously bread and butter, is their SOC and their security offerings. I'm expecting the next 2-3 years, they are going to invest into developing new technologies. The other thing that will drive Macquarie is the engagement of AI into the way they run their business. One of the really attractive things about Macquarie, and one of the things that made it a great company was the investment that they made into their young talent. They have a graduate programme, which is probably one of the best I've seen in the industry, where they develop people from the day they leave uni through to then, I've seen people move into senior leadership roles, people moving into senior CTO or technology leadership roles. What that means is, from an EBITDA position, their cost base is a lot lower because they are developing and training people who then stay with the company. They've got a fairly low attrition rate. Their retention rate is high, which means that their staff base and their staff costs are well managed.

JC [00:04:43]: How do you think the CS&G segment's revenue and EBITDA growth will evolve over the next 2-3 years?

RH [00:04:59]: Cloud Services will continue to run their show. They are quite heavily involved with Azure, but one of the benefits is the industry trend at the moment is everyone is looking to move away from pure cloud tech because of the costs that they didn't anticipate. Where Macquarie stands out is they are able to actually manage the Azure environments within their own data centres. It's a unique value proposition that they built out. What that means, from a customer lens, is customers can manage their costs based on the infrastructure that they're leveraging to run their Azure environments. I expect the Cloud Services team will continue to grow through their Azure services. Macquarie Government will predominantly be growing through their new Azure service that was launched as well as their security offerings that they're now developing and as well as the existing ones. I expect to see a decline, I think, in their Secure Internet Gateway revenue, but that'll be offset by the new products they're launching.

I'm not expecting Macquarie Government to have significant growth in the next 2-3 years. I think they will maintain their revenue position, and then probably in the next 3-5 years they will see growth. The good thing about Macquarie is that they've always had a very capable product development team. The way they go to market and the way they differentiate themselves is that they actually build a service rather than resell a product. They're not a standard system integrator. Because they own the data centre, they can control the entire customer journey. I think from a growth perspective, that will start to play a larger part in how they go to market and how they keep winning business.

JC [00:07:14]: When you say that you expect Macquarie's Cloud Services revenue to grow in 2-3 years, what percentage ranges are you looking at? Could you also provide a range for the Government side?

RH [00:07:26]: I would be guessing on the Cloud Services side, if I gave you a number, I don't know what the percentage growth would be. I'm not expecting it to be off the charts. I expect it to be organic growth. They will look for high levels of retention of their existing customers. Then they will incrementally grow YoY. If they were to hit a 10% growth position, I think that would be a good outcome.

JC [00:08:01]: Do you think that is realistic? Are you expecting high-single digits, for example? Do you have a range for that?

RH [00:08:10]: I think that's realistic for them, yes. They have a great service history. They offer a quality technology platform. They've got loyalty within their existing customer base, which they then leverage when they're selling. I do think that sort of growth is realistic.

JC [00:08:34]: What about for the Government segment?

RH [00:08:38]: I think the Government segment, I think they've been in government now for a very long time, so it's a well-established brand. It is a well-known brand. It's a trusted provider. Again, from a NPS perspective, from a customer longevity perspective, they will have the ability to position themselves as a trusted government provider. It's really going to come down to how well does the market respond to the new products that they are bringing out. When we look at the SASE models, they've got a good partnership with a number of SASE providers, but a couple in particular where they can then take that to market. I think that's going to be a big play for them.

00:09:28] What percentage of revenue growth do you expect to see in Macquarie's Government segment? You mentioned that you're not expecting much growth in the next 2-3 years and higher growth in the next 3-5 years. Could you elaborate on that?

RH [00:09:43]: I think from a revenue perspective, I don't expect the Government division to grow substantially at all. In fact, if they were to maintain their current revenue position, I think that would be a good result, but that's limited insight as to what else they've got up their sleeves. That is not publicly announced yet. If they were to be working on something which I wouldn't put it past them, that was going to take them to market, then I think that would change my answer.

00:10:31] Given that Macquarie is launching new products and moving from an SIG [secure internet gateway] to a SASE [secure access service edge] model and developing more technologies, how do you think the company's EBITDA margin will evolve?

RH [00:10:47]: Macquarie has always been very good at managing their costs. As I said, there's not a high attrition rate, but one of the things that Macquarie has done well is to ensure that when somebody leaves, can they either absorb the role and manage their overall cost base or do they reinvest into a new area? They're very, very good at actually understanding the value of people and how they deliver. With regards to EBITDA, I think they will manage their cost base really well. I think they will also look at all of their suppliers in terms of where their actual costs are coming from, and that will be a big play as well. For Macquarie Government specifically, with the decline in federal government's reliance on the SIG infrastructure, the costs required to maintain and manage that SIG infrastructure will be managed as well. If revenue does drop, for argument's sake, then I think they will do a really good job at managing the related costs associated with that revenue drop, so there won't be any revenue leakage or EBITDA cost leakage, if that makes sense.

00:12:27] How does Macquarie's EBITDA margin for the SIG segment compare to the company's new technologies and SASE model it's developing?

RH [00:12:39]: I expect that the SIG revenue will be significantly higher because it's been a staple product for them for a very long time. That's what Macquarie Government's business was built on. It was built on the SIG technologies. What I'm anticipating is the new technology will… I think the other thing is, the revenue per customer on the SIG platform is significantly higher than what I anticipate the new SASE model is going to look like. What that means is, from a business perspective, they are going to have to sell more licences and services for the SASE model than if they were to continue generating revenue via the SIG in business. They're going to have to work a little bit harder from a sales perspective to close the gap, but there is a gap, I believe.

JC [00:13:52]: When you say it's significantly higher, by how much is it higher? Do you have an estimate for the revenue per customer for SIG vs SASE models?

RH [00:14:07]: I don't actually because the SASE model was being developed just as I left. We had one major customer who had started the process to invest into SASE, but the other customers were still being developed. I think from memory, the revenue difference was I think you need two customers in SASE for every one customer in SIG, but there's a lot of nuance to that particular answer. What I mean by that is, with the SIG infrastructure, they would have a number of services attached to provide additional support. If they were to retain those services, then that gap would be much, much smaller, but if they were to lose all of the services associated with the SIG product set, then that gap would be quite significant.

JC [00:15:32]: How quickly do you think the SIG revenue drop would be?

RH [00:15:41]: It won't be quick because all of those customers are locked into contracts. The other key challenge with government customers specifically is, the time required for government agencies to implement change is quite slow. What has happened a number of times, despite wanting to move from the SIG environment, most customers renewed their contracts because they could not keep up with the rate of change. From that perspective, I expect that Macquarie Government is going to be quite stable for a number of years.

JC [00:16:34]: Over what timeframe would that happen?

RH [00:16:38]: I think three years. I think they will continue to generate reasonable revenue, and they won't have a lot of… What's the word I'm looking for? For the life of me, I can't think of the word when customers leave.

JC [00:17:05]: Churn.

RH [00:17:05]: Churn. Thank you. I don't expect they will have a lot of churn.

00:17:09] Could you break down the primary cost lines involved in delivering CS&G solutions? How do these cost structures differ between government, enterprise and mid-market clients?

RH [00:17:31]: Some of that information is sensitive because it's not public knowledge, but I can share some insights. A few of the key cost drivers for the Government division is obviously people. That's one of the biggest drivers of cost. Then, in addition to that, you have all of the infrastructure that is used to deliver all of their various services, so that would be another substantial cost. Associated with it, that infrastructure is all of the support costs, the licensing costs, etc. From a Cloud business, again, they have reasonable costs from a people perspective. I think it's probably one of the highest parts of their cost base. Then they also have a significant amount of infrastructure that they manage on behalf of their customers. While it is a cost of doing business, so it would likely sit in your cost of goods line item, that's probably where a large chunk of their costs come in.

00:18:54] How do the costs to deliver services differ between Macquarie's Government and general enterprise Cloud Services segments? What is the difference there?

RH [00:19:12]: It's significantly different. With a lot of the government customers, the actual build and implementation of the SIG service is bespoke. Each government customer has their own requirements, their own architecture, their own infrastructure. Cables need to be connected through various third parties. There is architecture work, etc, etc. There is a reasonable difference in the actual work that needs to be undertaken to be able to run that Government business. On the cloud side of it, a lot of it is pretty much cookie-cutter, rinse-and-repeat style product offerings. Less bespoke, which means that their cost to serve, cost to implement and time to implement is more predictable. Therefore, they can manage the costs associated with those particular implementations. The Cloud business will deliver new customers far cheaper, but the size and scale of those customers is a lot smaller in a lot of instances, but the government customers are buying a very different product.

The implementation costs are typically a little bit higher or a lot higher depending on the size and scale of that customer. I was going to say, of the regulatory issues that need to be understood when it comes to standing up services, SLAs, impacts, etc. There's a different expectation on support services when it comes to government contracts.

JC [00:21:20]: When you say the Government costs are significantly higher than Cloud Services, could you provide an estimate of the difference?

RH [00:21:33]: Honestly, I can't answer that. It's comparing two very different things, and we never actually looked at what the difference was between those costs.

00:21:50] How long does it take Macquarie to onboard new government and cloud enterprise clients?

RH [00:22:03]: Enterprise clients are a lot faster. They're a lot more agile and dynamic. In some cases, they're also a lot smaller. Enterprise customers can be implemented within weeks. Again, because the service itself is pretty much a standard offering. There aren't a lot of variables associated with it. It's just making sure that connectivity, customer engagement, understanding, that's all in place. With government customers, a large chunk of the time to implement is typically down to customer delays. Federal government agencies don't tend to work at pace. Timelines are mostly irrelevant. They work to make sure that they are compliant. They work to make sure that the service delivers value, but time to implement and turn on the service is not a primary driver. It takes a strong project manager within the government space to push the customer to turn on products. It can take a couple of months to turn on services with some of the slower government customers.

Also depends on the size and scale of what's being turned on. As I said earlier, a SIG implementation, a SASE implementation, two very different engagements, both have a fair amount of customer involvement in actually turning on the service, managing staff, managing staff devices, etc. It makes it a more complex project, and therefore durations will be longer.

00:23:50] How do you expect the cost line dynamics for cloud and government solutions to shift over the next 2-3 years?

RH [00:24:02]: I think like every business in the world, the cost dynamic is going to shift to what can we leverage AI to manage and support, take away all of the administrative burden as a starting point. I expect the Government division will be a lot more limited in how it can leverage AI, though, because there will be a limitation on where the data is stored for Macquarie Government to be able to leverage it effectively, especially when it comes to any documents that are security rated.

00:24:50] How do you expect the underlying infrastructure, staff, services or licensing cost line dynamics to shift in the next 2-3 years?

RH [00:25:08]: It's a tough one actually, to be honest. If I think about how the industry is changing, as a starting point. There is an expectation that the staff costs will continue to slightly go up, but they will be managed. As I was saying earlier, there is a significant investment into their grad programme. A lot of the roles that will be developed over the next five years, I think will continue to be developed through the grad programme. When people do leave the company, I think they will manage those costs really well. That's how they've always done it. I think from that perspective, staff costs will be managed. I don't expect it to blow out. I expect the people that are remaining or the people that do leave, a chunk of their roles will be replaced through some form of automation with AI. I think there's an element of staff costs will be managed really well. What was the other part of your question?

JC [00:26:38]: Touching on infrastructure costs, licensing costs and so on.

RH [00:26:43]: I think that one's going to be far more interesting. One of the things that happened in the industry a few years ago, and this impacted everyone, was when VMware, bought out by Broadcom, changed their entire licensing model and the financial impact that had on every customer globally was substantial. Macquarie's response to that and how they manage their way through it says that they are on top of their vendor costs and their licensing costs, and they've got some very good vendor relationships. They'll continue to, I think, negotiate really well. They've got some very clever people in there who are very good at negotiating. They will build that out. I don't expect it to significantly increase in the next five years.

JC [00:27:41]: You still think that will increase. Do you think it will be a single-digit increase? Would it be low, mid or high-single digits?

RH [00:27:48]: I expect it to be single digits. It'll be aligned with inflation, if nothing else. Unless there's again some drastic shift. For example, with AI costs coming in, if AI costs go through the roof, that will be offset with some form of cost saving somewhere else, so it'll always be balanced. Macquarie has and always will be very fiscally responsible. That's why they've had continued growth and success for such a long time.

00:28:25] How are Macquarie's cost structure and scale threshold evolving for running dedicated NV1 [Negative Vetting Level 1]-cleared engineering teams for federal agencies vs serving clients through a shared 24/7 SOC [security operations centre] pool?

RH [00:28:54]: When you talk about NV1-cleared staff, you do pay a premium. They are in high demand. It depends on the level of the role. If I was to use an example, a senior project manager that is NV1-cleared in Canberra, so and again, location matters, might end up costing you, say, AUD 160,000 or AUD 180,000, whereas a senior project manager in Sydney that is not NV1-cleared might cost you AUD 140,000. You might end up paying a premium for some of these people.

JC [00:29:44]: If we're comparing the same city and same title, what percentage is the premium for NV1 demand on average?

RH [00:30:01]: I would probably say NV1 demand is, on average, it probably picks up a 15% premium. If I'd had to put a percentage on it, but that's… It's not a factual number that you're going to be able to reverse, calculate or be able to determine impact to the business. I think the other thing that's probably just as important to understand in this process is, with Macquarie, if they didn't have those NV1-cleared engineers, then they wouldn't actually be able to go to market and sell their services. Once you do pay a premium for NV1-cleared staff, and typically, the NV1-cleared staff in Canberra are more expensive because they are in higher demand and there is a lower supply for the group. Anyone who is actually incredible at their jobs, and they are quite niche, typically, they'll start their own consulting business because they can then charge themselves out across multiple agencies for different rates. Canberra, you do pay a premium for location as well as then clearance. For Macquarie Government, if they didn't have that investment, then they wouldn't be able to sell the services that they do. It's a necessary cost.

JC [00:31:42]: How much of a premium do Canberra NV1 staff demand on top?

RH [00:31:49]: I would probably say some of the people in Canberra could generate a 20% increase on costs.

00:31:59] What operational levers does Macquarie use to manage the premium cost it has to pay for NV1 and any wage inflation?

RH [00:32:20]: Like every company, there are performance reviews. There is a fairly rigorous recruitment process, but ultimately, when you hire the right person, their roles deliver value that is commensurate to what they're being paid. It's not like you are overpaying someone to do that role to earn you less money.

00:32:52] How is the salary premium for NV1-cleared staff demand trending? Is it increasing?

RH [00:33:01]: I don't think it's increased in any deviation from standard wage increases. For argument's sake, you're not paying a senior architect AUD 300,000 now that's NV1-cleared vs then a senior architect who's not NV1-cleared, paying him AUD 180,000. There's no substantial increase for NV1-cleared staff vs other staff. The other thing to remember is, people can go through the process to become NV1-cleared. Typically, what I've seen happen with Macquarie, one of the things that we did was we would bring on teams, we would bring on people who we would then put through the NV1 process, and they would be cleared within a maximum of six months, sometimes three, depending on the agency that was doing the vetting. There's ways in which you can manage around the premiums, but the premiums themselves typically come from people who work in Canberra and have federal government experience.

00:34:50] When Macquarie is trying to improve its EBITDA margins or considering optimising its internal service desk operations, how is the company doing that? How sensitive are government agencies to these operational shifts before risking any SLA breaches?

RH [00:35:05]: When Macquarie looks to improve EBITDA margins, there's a few things that we consider or at least that I would consider. The first one is, what is the capacity of the existing teams to take on more work? Second part, then, is what can actually be automated without breaching any form of security policies or security expectations from a government perspective? Then the third one is, do we actually need to spend the money doing whatever we're doing? Is there an alternative? Some of it is introducing new technologies. For example, we were using one technology that, for argument's sake, cost AUD 200,000 per year to run. We found a different provider who gave us a better quality of service, but it only cost AUD 50,000. There was a reduction of AUD 150,000 in operational costs. The other types of projects that Macquarie will continuously undertake to ensure that… When they are going to spend the money, and they're not afraid to spend the money as appropriate. There's a value exchange.

That's typically how Macquarie has gone through and managed their costs. The second part to that is the sensitivity to tier 1 government agencies is incredibly… It has to be done. It's not a choice of if. It's a question of, what do we need to do to make sure that we remain compliant? Otherwise, you risk a breach, a material breach. It's not just on an SLA side of things. The SLAs are slightly different in this scenario. This is to a, "Can the service still actually function?" It's not actually an SLA issue. Is the service still capable of functioning as a protected or an IRAP-approved service?

00:37:33] How can Macquarie deliver cost savings through new technologies and automation? How much cost savings can that deliver? Could you provide an example?

RH [00:38:01]: Let me give you one example, which I know customers would know about. We implemented a secured version of ServiceNow. When I was in my role, we had implemented a new version of ServiceNow that was secured.Interestingly enough, that protected instance should probably cost more. That was down to making sure that we negotiated the right options. We were able to get the right implementation, but the implementation itself we did in-house. That also managed our overall cost, bringing in third parties, etc. Within ServiceNow, we then turned on a number of functions that helped us retire, I think, five different applications that we were using and paying third parties a licensing cost for.

We were able to turn those off. The cost savings increased again. That's the type of work that I think the team thrived on. Now in addition to that, there are some very, very clever people who are building out AI offerings built into the gateway as well as into the SOC. That, again, is something that I look at and I go, the team had done a phenomenal job at reducing the cost to serve without impacting the quality of service or the security position. A lot of the focus is there are very, very clever people who will continue to build automations and then report and track progress towards a goal. That's how Macquarie Government would reduce the OPEX without impacting service.

JC [00:40:26]: As new technology develops and automation matures over the next 2-3 years, how much cost savings can Macquarie generate from these actions? Could you provide an estimate in percentages?

RH [00:40:47]: The targets every year would be a minimum of 10% reduction every year. That's what I anticipate is the target for the team every year. 10% reduction in OPEX.

JC [00:41:03]: Do you think that is realistically deliverable?

RH [00:41:07]: Yes, I think that is realistic. I think they've got some very, very clever people there who will drive a cost reduction across the business through inventing new ways of working, developing new tech. I think they've got some very clever people there.

00:41:29] Macquarie is shifting its product evolution from capital-intensive private cloud hosting to its lower-CAPEX software platforms such as Managed Fabric and Launch AI. How does this mix shift alter the company's underlying service and delivery cost structure and EBITDA margins?

RH [00:41:51]: As I understood it, the Managed Azure platform has been part of cloud now for a while, but they've just launched Managed Azure into government space. I'm not sure I would call it capital intensive as such vs lower CAPEX because a lot of it still sits within the data centres that Macquarie manages and runs. While establishing the original offerings, yes, there was an investment that Macquarie made to implement the platforms, but each customer… It's like a little LEGO block set. Each customer pays for their add-on. There's costs obviously built into the onboarding and implementation, and then the monthly run rate. A lot of the CAPEX or the expenditure to expand the Managed Azure environment would be done per customer, but the original base that would not be capital intensive. It would be a capital investment to set up the original pilot or the first group.

Then after that, they would continue to build. Where there is a cost to Macquarie, is typically in occupancy. When you're building out the Managed Azure environment, you need to know that you have capacity or space in the data centres to then expand out that offering. It's an opportunity cost is that that space is then reserved for those particular offerings.

00:44:04] What does Macquarie's utilisation rate look like now, given the demand for Managed Azure and new technologies?

RH [00:44:16]: Explain to me what you mean by utilisation rate.

JC [00:44:18]: You mentioned how there is a cost to serve. If the infrastructure and the additional expenditure are built and it is underutilised, what would that look like?

RH [00:44:34]: The occupancy rate affecting data centres. I don't know where it's currently at. What I do know is it's one of those key metrics that Macquarie monitors and has a very strong view on. What I can share is there is dedicated space for Macquarie to deliver federal government customers within their data centres, and those costs are allocated to the Government division. If there was an occupancy issue for the federal government space, that doesn't get solved by selling Cloud Services. Whereas if there was capacity issues with the cloud infrastructure space, that could actually be consumed by the data centre team.

00:45:32] Given that software vendors and hyperscalers have aggressively raised their licensing and compute prices, how much of these input cost increases can Macquarie pass directly on to its end customers under its multi-year contracts? Where do you see margin compression occurring?

RH [00:45:59]: With Macquarie, with the cloud team, those costs can be passed on to the customer directly in a number of instances, especially if it's a managed environment, where the costs are delivered. Where it's part of a service that Macquarie delivers, there is the ability for Macquarie to move those costs to customers. What has happened with the VMware instance was Macquarie brought together a Tiger team, and they found a way in which they could actually do some quick implementations and changes that would enable the customer to either maintain or reduce their contract costs or take on a smaller increase than otherwise. Every time a vendor passes on significant cost increases, it just means customers have a decision to make and you're creating a burning platform for them to make that decision. Macquarie has been very good at reacting to those situations fast and making sure that customers always have a choice as to how their infrastructure is being managed.

JC [00:47:24]: What percentage of these costs can Macquarie directly pass on to its end customers?

RH [00:47:35]: If it's a direct customer-related cost, then all of them. I'm pretty sure with Macquarie Cloud, all of those costs get passed on as appropriate. If it is, for argument's sake, an increase in an application or something that is used to support the customer, but it's a shared platform, then those costs are typically not passed on, but as I said earlier. If ServiceNow pricing increased, unless the customer is specifically paying for a specific ServiceNow licence, then those costs would have to be absorbed by Macquarie, but what typically happens in that scenario is they go and renegotiate a different outcome. They very rarely end up paying for that additional service or those additional costs.

00:48:40] How do customers react to cost increases when Macquarie is trying to negotiate contracts with them?

RH [00:48:51]: I suspect you get a wide range of responses from customers. Some of them would be outraged and disgusted that they were being taken advantage of. Others would completely understand. I think it comes down to the maturity of the customer. It would come down to the capability of the salesperson having that conversation. It would come down to what that percentage increase is. The reactions you would get from customers would vary significantly. When it came to the VMware Broadcom increase, every customer understood because it was public news, and it was happening to everyone, regardless of size and scale and shape. This was a global Broadcom conversation. From that perspective, I go, yes, the reactions would be far more manageable.

00:49:55] Customer onboarding delays directly lag MRR [monthly recurring revenue] recognition. Where do implementation bottlenecks most commonly occur between contract signing and full deployment?

RH [00:50:18]: Let's take the first part. Is there any impact to monthly recurring revenue recognition with delays in customer onboarding? The answer to that is it depends on what is in the contract. If it is a brand-new service that is not co-terminating with any other service, then day one of that particular service or invoicing for day one is on the day the actual service goes live. There's no actual lost revenue as such. It's delayed revenue. Whereas, if a service is, and again, contractually, the revenue turns on from the date of contract signature, but cannot be invoiced until the service is live, then yes, there is lost revenue until the service is turned on. It varies. In government contracts, it was more prevalent that a lot of the add-on services. For example, if you had some form of e-mail security service that would be linked to a SIG contract. Therefore, if you didn't turn on the e-mail service within a set timeframe, then you would lose revenue because it would obviously be contracted to align with the SIG contract.

Whereas if it was a brand-new service, then that would be aligned to whatever the agreed contract timeframe is. In terms of the actual delays that people would experience, a lot of it came down to, in government, specifically, staff availability. A lot of federal government agencies have a small team, every one of which actually does every project that comes in, and federal government agencies like to do multiple projects at once. They can be slow. It could come down to risk reviews, IRAP assessments. There's a whole bunch of things that could delay a service being turned on. It could be something like you need a cable run within the data centres to connect two racks to enable the service, then having that lined up takes time because you need, again, people with the right clearances, third parties who actually have time to do the job. There's a whole bunch of things could line up. It's a service that requires hardware. Any delays to hardware procurement could actually delay revenue. It's quite a complex process to make sure that all of those things, all of those items, line up correctly.

JC [00:53:25]: What is the typical time delay for government to successfully work out the full deployment?

RH [00:53:35]: Government delays were typically down to government staff being made available. That was primarily the biggest delay that we saw.

00:53:47] Regarding Macquarie's sovereign data platforms such as Managed Microsoft Fabric and Launch AI, are Australian enterprise and government clients actively converting AI pilots into recurring multi-year managed cloud contracts, or is adoption still mostly experimental?

RH [00:54:08]: For government agencies, I can't answer you with any level of authority on this just at the moment because I haven't been part of that for the last year. What I can tell you is, when I was working directly with federal government agencies, everyone was experimenting. The biggest question almost every federal government agency has right now is, how do we manage data sovereignty without breaching any of the security pieces? If I give you an example, if you look at Claude as a platform, all of the data is stored in the US for 30 days. That is not something that a federal government agency can actually tolerate. If I then go implementation of Copilot built into the various platforms that government agencies have. That, again, is something that I go unless there's a protected instance delivered by Microsoft, which I'm not sure there is just at the moment, then those government agencies would not be allowed to use that service. That's the biggest gap right now. I think it's the ability for federal government agencies to have an IRAP protected instance of an AI platform that is available for them to use.

JC [00:55:39]: How long do you think it would take for this data sovereignty issue to be resolved and for the sovereign data platforms to mature, successfully deliver and be monetisable?

RH [00:55:58]: I think the monetisation side of it is already clear because it's being done in the consumer world. I don't think it would take long to monetise the platforms once they're up and running. The question is going to be who is capable of getting an IRAP certified platform or IRAP assessed.

00:57:28] When federal government agencies and enterprises evaluate their sovereign cloud providers vs global hyperscalers, what specific operational factors drive their decision to sign with Macquarie?

RH [00:57:44]: What is Macquarie's point of differentiation is what you're asking?

JC [00:57:48]: Yes.

RH [00:57:49]: Different answer for federal government vs enterprise customers. One thing they have in common is quality of service. If you look at any of the content that Macquarie produces, including a book that was written about them, The Macquarie Way. Everything that they have built comes down to exceptional customer service. They've got an NPS that is 80% of the business. That is a big selling point. What it means is customers get looked after. Most customers want peace of mind. Whatever they're buying, wherever they're buying, they want peace of mind. The second part is always going to be down to price. There's always cost sensitivities around price, which, again, Macquarie, I think, prices themselves reasonably well. The reason that customers would go to the bigger providers in terms of the hyperscalers would be because of an actual service that they specifically needed, or what also tends to happen is the actual hyperscalers buy the business.

What I mean by that is Amazon will actually go in and do all of the implementation for free, provide benefits to the government financially from a contract perspective to say, we will cover X percentage of the overall costs for the first three years while you're going through your implementation and onboarding and blah, blah, blah. It can actually shift. There's significant investment from Amazon right now into building a defence grade data centre, which I think they're doing in conjunction with defence. There's all those joint investments as well that I think companies like Amazon have the ability to do. Macquarie has done joint investments, but different levels. I think the last part with Macquarie, specifically in the government space, and the reason that they have been so successful is they've had a longevity in government, which very few companies have. They've consistently maintained their quality reputation. I think that makes a big difference in terms of the risk factors associated with government agencies going, do we know who we're working with? Do they have a track record of success? Can we trust them?

01:00:46] How have Macquarie's retention, stickiness and churn trends evolved across government and enterprise or mid-market accounts?

RH [01:01:03]: Let's start with the first part of your question. I think client retention. A lot of clients have a couple of things that they think about when they're looking at either retention or churn. First is, do they see value for money? The second part is, is there a better alternative? If I spend the same amount of money, do I get more for my dollars? The third part is quality of service. Then the fourth part is, where are they at from a growth perspective for their development vs then the partners that they're with. There's so many factors to it. What I have seen from an industry trend recently is we've gone to a point of insourcing again. There's a lot of companies that are insourcing their businesses. That's just a trend that seems to happen every few years. It's a cyclical thing across IT. The stickiness, specifically in Macquarie Government, especially around the SIG services, was there was no way to replicate the SIG easily. It is actually quite a complex environment. It has a significant amount of work and effort to not only run the environment, but to actually build it to a specification that keeps government agencies protected.

When you say we're protecting government agencies, the amount of… There was something to the tune of millions of hacks that were perpetrated every day that the Macquarie Government team would block consistently. I think that is a big factor. When we talk about retention and stickiness, they're the factors that I think are critical. Where government is evolving is… I think we touched on it at the beginning of the call, there is no longer a mandate for a SIG service. A SIG is considered archaic technology. Now while government agencies will remain on the SIG until their contracts expire, it's also an expensive service, by comparison to shifting across to an ISP and putting a SASE model over the top. The cost for money or the value for money exchange is now something that's being reviewed. There is also a significant cost to change. Removing the SIG, redirecting everything to SASE, implementing SASE, building all your rules, etc, that is an investment that government agencies need to make.

When you think about retention and stickiness, they're all the factors that are there. Mid-market, I think less. It's all about what is it costing. Mid-market, you've got small business, medium-sized businesses. They need to ensure that they are running their business effectively. To do that, they've got to manage their costs. Churn in mid-market, from my experience, is typically about price. It's not about much else actually. Macquarie does a good job at managing that cost conversation and ensuring that customers not only get great service, but that's a part of the conversation that means that they become quite sticky. The Macquarie team also has quite intimate knowledge on the customers' environments, which means it's harder to move because that knowledge provides a level of support that customers need. There's all those little factors.

JC [01:05:29]: We will now conclude the Interview. Let me close by saying thank you for all of your input and contributions.

Transcription ends at 01:05:46 of the recorded material