Bentley Systems (BSY) Q3 2026 Earnings Call: ARR Growth Hits 13%, Outlook Raised

Bentley Systems (NASDAQ: BSY) reported accelerating recurring revenue growth in fiscal Q3 2026, supported by resources, public works and utilities, grid infrastructure, and large Asset Analytics expansions. Management also raised its full-year ARR growth outlook on a like-for-like basis, including the expected contribution from the Elis acquisition.

Key Takeaways

• Q3 revenue rose 12.2% year over year to $422 million, or 12.7% in constant currency. Subscription revenue, representing 92% of total revenue, increased 12.4%.
• Constant-currency ARR growth accelerated to 13%, while sequential constant-currency ARR growth reached 3.1% and was entirely organic.
• Net revenue retention increased to a record 110%, while account retention remained at 99%. New customers contributed about 300 basis points to ARR growth.
• Adjusted operating income less operating stock-based compensation was $135 million, with margin expanding to 32.1% from 28.1% a year earlier.
• Bentley raised its 2026 ARR growth outlook excluding the Cohesive divestiture to 12.5%-13.5%. Including the divestiture, reported ARR growth is expected to be approximately 100 basis points lower.
• Management said Asset Analytics is already monetizing AI in infrastructure operations, but AI-driven design monetization is not expected to become a material growth driver in 2027.

Core Financial Data

Year-to-date revenue increased 13.2%, or 12.3% in constant currency. Subscription revenue rose 13.5%, or 12.7% in constant currency. Last-12-month recurring revenue reached $1.528 billion, representing 93% of total revenue.

Business and Operating Performance

Resources remained Bentley’s fastest-growing sector. Seequent benefited from sustained mining demand as customers prioritized brownfield optimization, expansion, and critical-mineral development. Bentley said the Elis acquisition adds seismic interpretation to Seequent’s subsurface portfolio, with applications across oil and gas, geothermal energy, carbon capture, mining, and civil infrastructure.

Public works and utilities also delivered strong growth. PLS performance remained notable as utilities invested in transmission-grid expansion and modernization. Asset Analytics secured enterprise expansions in telecommunications tower inspection, road monitoring, and transportation infrastructure AI.

Asset Analytics accounted for approximately half of the improvement in sequential ARR growth compared with the prior-year quarter. The business uses AI to assess the physical condition of infrastructure assets and connects those findings with Bentley’s modeling and simulation capabilities. Its pricing is based on assets per year rather than users per day.

Bentley’s SMB business added more than 600 new customers through Virtuoso. Overall, new logos contributed about 300 basis points to ARR growth, with management describing SMB project backlogs as healthy.

By geography, Asia Pacific was the fastest-growing region. India benefited from domestic water, grid, and offshore energy investment as well as its role as an engineering delivery hub. Australia and New Zealand were supported by resources, while Southeast Asia saw demand from civil projects and power networks.

Growth accelerated in the Americas, led by North America and Latin America. Management cited mining, water infrastructure, public funding, private capital, and AI data-center investment as drivers. EMEA delivered solid performance, with strength in the Middle East and ongoing infrastructure programs in the U.K. China represented less than 2% of ARR and continued to face geopolitical and macroeconomic pressure.

Cohesive was divested on October 13 and merged into Navion, with Bentley becoming a significant minority investor. Bentley expects the transaction to increase subscriptions to approximately 95% of revenue from 92%, making its revenue mix more recurring. Management views Navion as a longer-term channel for integrating Bentley Infrastructure Cloud and iTwin technology with IBM Maximo environments.

Management Guidance

Management said reaching the high end of the ARR outlook requires continued strength in resources, public works and utilities, grid products, Q4 E365 renewals, and additional large Asset Analytics contracts.

Risks and Watch Items

• Bentley expects services revenue to decline year over year through 2027 following the Cohesive divestiture and to represent less than 3% of total revenue, versus approximately 5% year to date.
• Asset Analytics can be more volatile because major enterprise procurement cycles are less predictable. The business also faces difficult comparisons after a particularly strong Q4 2025.
• AI use in infrastructure design remains at an early stage. Adoption depends partly on customers moving from experimentation to production and upgrading to applications with the required APIs and MCP servers.
• Management expects AI design monetization to begin in 2027 but does not expect it to be a material growth driver that year.
• China continues to face geopolitical and macroeconomic headwinds that management expects to persist.
• If end-of-October foreign exchange rates persist through Q4, revenue would face an incremental $5 million negative impact relative to the rates assumed in the outlook.
• Fourth-quarter free cash flow will depend on the timing of year-end collections around the January 1 renewal cycle.

Analyst Q&A Highlights

AI monetization: Management said monetization is already occurring through Asset Analytics, including telecommunications tower inspection and road monitoring. For design applications, Bentley is prioritizing adoption and workflow validation before broader monetization.

Commercial model changes: Bentley expects machine-speed use of engineering applications to require pricing beyond traditional user- and time-based models. Management said engineering firms are also moving toward fixed-fee and outcome-based contracts as AI increases productivity.

Elis acquisition: Management described Elis as accretive to ARR growth and roughly aligned with Bentley’s margins. The most immediate opportunities are in energy, followed by deeper mineral exploration and civil infrastructure applications.

Bentley Infrastructure Cloud: Management emphasized the platform’s openness across Bentley and non-Bentley applications and its policy that customer data remains the user’s data. Asset Analytics is intended to connect existing project data with operating-phase digital twins.

Infrastructure and data centers: Management said increased scrutiny of data centers’ energy, water, and surrounding infrastructure requirements could create additional demand for site planning and infrastructure digital twins.

Go-to-market strategy: Bentley does not expect a major strategic change under its new Chief Revenue Officer. The company plans to reinforce technical selling and strengthen feedback between customers, solution engineers, and product teams.

Full Earnings Call Transcript

Complete Earnings Call Transcript

Management Remarks

Eric Boyer

Good morning, and thank you for joining Bentley Systems' Q3 2026 Results. I'm Eric Boyer, Bentley's Investor Relations Officer. On the webcast today, we have Bentley Systems' Executive Chair, Greg Bentley; Chief Executive Officer, Nicholas Cummins; and Chief Financial Officer, Werner Andre. This webcast includes forward-looking statements made as of November 5, 2026, regarding the future results of operations and financial position, business strategy and plans and objectives for future operations of Betley Systems Inc. All such statements made in or contained during this webcast other than statements of historical fact are forward-looking statements. This webcast will be available for replay on Bentley Systems' Investor Relations website at investors.bentley.comon on November 5, 2026. After our presentation, we will conclude with Q&A. And with that, let me introduce the Executive Chair of Ely Systems, Greg Bentley.

Gregory Bentley

Welcome, and thanks to each of you for your interest and attention as always. We're excited to report '26 Q3 operating performance in which many factors contributed so impressively that are a key metric year-over-year ARR growth after already notching up during each quarter of 2026, and at 13% surpassed our annual outlook range. Our NRR, net revenue retention for the quarter also reached a historical high of 110% and thanks to Nicholas and our teams for stalwart global execution. Being purposeful about programmatic acquisitions, we define our business performance ARR growth to include the generally rather minor ARR reported with these, the last of which had occurred back in '25 Q4.

I am very enthusiastic about the substantive rationale for our just announced '26 Q4 acquisition of Ellis, as Nicolas will further explain by expanding both the scope and the reach of sequence market-leading comprehensive subsurface modeling it is also accretive to our ongoing growth rate.

Werner will quantify the business performance ARR growth impact of the acquisition and also of pro forma, excluding our copies business, the divestment of which we've also just concluded. Although cohesive will no longer be financially consolidated, I will make the case that its strategic significance for us in support of our generational opportunity in infrastructure operations and maintenance can become greater. Last quarter, I showed how we are uniquely well positioned for this operations life cycle opportunity. In fact, Among our enterprise accounts, those spending over $100,000 per year with us, more at each enterprise tier and spending more with BSY, our owner operators than our project delivery firms, their supply chain, who make up the other half of our business. As I also showed last quarter, 346 of our enterprise accounts are among the 457 ex China, Bentley Infrastructure 500 top owners in the world ranked on their net infrastructure asset value.

To underscore the compelling magnitude of this O&M life cycle opportunity, these existing enterprise accounts, collective infrastructure asset value is $15.5 trillion. In our annual BI 500 reporting, we classify and analyze the top owners and their asset values informatively, A relevant question this quarter is whether the recent surge in government bond yields globally could materially impact infrastructure investment by encroaching on sovereign government budgets. The evidence dispels this potential concern. As you see, perhaps surprisingly, and again, ex China, the proportion of the top owners collective infrastructure asset value that is owned privately is half again as large as the proportion owned publicly.

Drilling down to the public private distribution of BSI's revenue from the enterprise accounts, which represents over 3/4 of our total revenue, as shown here, underscores this resilience. While the largest infrastructure sector for BSY is public work/utilities, this includes the large investor-owned utilities. And half of our revenue in public works/utilities is from the private project delivery supply chain doing the capital project work. All told, less than 1/4 of our enterprise revenue is directly from any level of government and within governments, more of our enterprise revenue is from the combination of state/provincial and local governments than from federal government. Our revenue proportion directly from the U.S. federal government is less than 1%. Among public works at all levels of government, the bulk of infrastructure spending is financed through utility rates, user fees, motor fuel and vehicle taxes and other dedicated sources rather than full faith in credit.

The overriding objective observed everywhere is to avoid subjecting infrastructure investment, given its fundamental primacy to support all other economic, environmental and social priorities to the tightening constraints of public budgets and/or credit factors. Greater stress and global sovereign credit quality and spending capacity accelerates private financing of infrastructure investment both by necessity and choice and increasingly through varied blended finance alternatives in addition to traditional P3 public-private partnerships. Hospiciously for BSY, private infrastructure financing is most often effectuated through design, build, operate concessions, incubating life cycle engineering digital twins as the perfect prescription.

In magnitude by sector, our own digital twin opportunity in infrastructure operations corresponds closely to the BI 500 top owners distribution of asset value by their infrastructure sector classification. For owner operators within the mainstream infrastructure sectors of public work/utilities, industrial and resources, the compelling economic potential of engineering-grade digital twins during the operating life cycle is to enable better maintenance decisions, performing only what and when and how would be determined at the engineering level to be necessary and effective planned in 3D while continuously monitoring in 4D and simulating throughput, resilience and safety. The ideal way to effectuate such improved maintenance be in conjunction with an owner's enterprise asset management, EAM system, responsible for maintenance work orders and their history.

In keeping with our long-standing determination to integrate with rather than to replicate the functionality of enterprise administrative systems, and so as not to handicap our sales force by contending against them, we undertook to partner with IBM Maximo by far, the EAM environment most prevalent among our owner-operator accounts and prospects. To explore and to learn how to introduce operating digital twins, just before our IPO 6 years ago, we acquired the original cohesive, a leading Maximo systems integration firm and then combine several subsequent Maximo services acquisitions into our eponymous captive global digital integrator. The enlarged cohesive would go on to encounter sustained headwinds as Maximo customers responded slowly to IBM's and forced migration from sunsetted on-premise Maximo to managed services.

Cohesive utilization has since then improved substantially, but after the all-consuming distraction of the reluctant migrations, digital twin integration with Maximo remains nascent. During this fray, very capable PE-funded entrepreneurs recognized a compelling opportunity and founded U.K.-based Navion by consolidating most of the world's remaining Maximo business partners. Navion breakthrough was to very successfully institutionalized provisioning comprehensive Maximo hosting plus ancillary cloud services.

We are very pleased to have now joined forces by merging cohesive into Navion in exchange for becoming a significant minority investor in this fast-growing, profitable recurring revenue business now a fully scaled and credibly independent digital twin integrator. Beyond financial investment returns, we will gain as Navion adds Bentley Infrastructure Cloud and our iTwin technology to its Maximo additive repertoire with its services for us, now including our former cohesive colleagues, forward deployed throughout the Maximo footprint. The distribution by sector here of NAVIOM's business post Cohesive reinforces this neutral synergy while serving also as a representative proxy for the stronghold business mix of Maximo itself. Only about 10% is outside of infrastructure. Just as for BSY, the majority of Navion all Maximo centered business is in public works/utilities followed by the resources sector than Industrial.

Among our owner-operator accounts, we remain on the cusp of advancing engineering software and AI throughout the full infrastructure life cycle. A case in point are the 346 ex China, BI 500 top owners, who as BSY accounts already spend on average almost $1 million annually with us each. But so far, corresponding to where we started and our direction of progress as a company, that's mainly for CapEx project delivery, not OpEx asset performance, leaving a big opportunity cost for them and an emerging opportunity for us. On average, these 346 accounts subscribed to 6 brands on average, 3 of our top 10 brands across our portfolio, thus generally including both modeling and simulation functionality within our portfolio of Bentley open applications. As I mentioned last quarter, almost half already leveraged Bentley Infrastructure Cloud as a system of record for at least the portion of their infrastructure engineering performed with our applications.

Most of this usage is of project-wise. A small minority to date do use asset-wise for asset information management, thus sharing the same Bentley Infrastructure cloud schema for OpEx as CapEx and but so far, only rarely throughout the full infrastructure life cycle. But now enabled and empowered by AI asset analytics can be considered the keystone of our strategy to finally close the loop between our CapEx point of departure and our OpEx opportunity. For ultimately every infrastructure asset, instant on digital twins are becoming feasible by leveraging AI through our reality capture software and continuous surveying, including autonomously to inspect as operated conditions of the infrastructure asset updating and 4D monitoring its engineering-grade digital twin and Bentley Infrastructure Cloud.

The differentiating payoff is that asset analytics then closes the life cycle loop by agentically invoking our pertinent modeling and simulation capabilities to intrinsically calibrate and simulate ongoing capacity, resilience and safety, supporting asset performance management algorithms to optimize maintenance strategies and scheduling, uptime and throughput. Our exemplary use cases so far are for integrated grid structural resilience ranging from cell towers to distribution poles. Our imperative is to extend our asset analytics platform across our application portfolio and across all accounts with AI appropriate to create, curate and leverage operating digital twins for every infrastructure asset.

By right, AI-enabled asset analytics for infrastructure operations although more volatile along the way by virtue of its less predictable sales cycles for major procurements should become increasingly material eventually within our overall business proportions, but first, within our incremental growth. And as it happens, Asset Analytics was among the converging factors behind months surpassed business performance ARR growth in '26 Q3. For the fuller picture of all those factors, now over to Nicolas. Thank you.

Nicholas Cumins

Thank you, Greg. Last month in Singapore, we hosted our 2026 year infrastructure event, now focused on our global awards competition. I encourage everyone to explore website to learn more about this year's finalists, winners and founders [indiscernible] and to watch the events recordings, including my own reflections on the practical progress made with AI. Shortly before and after Singapore, 2 important industry publications were released. The American Society of Civil Engineers published its first AI road map for civil engineering. And advisers issued their comprehensive state of the AC industry report. Together, these developments capture what the infrastructure industry stands today with AI and where it is heading. .

The ESC road map points out that civil engineers are not only users of AI, they are the enablers of the physical infrastructure AI requires to exist. Every AI inference ultimately depends on physical fiber backbones, power generation, high-voltage transmission lines, substation interconnections and water cooling networks that enabling infrastructure must be planned, designed, built and maintained by engineers. At the very same time, Global demand for new and bed infrastructure is vastly outpacing available engineering capacity. The only way to bridge discount between supply and demand is by empowering engineers with AI. The AEC advisers report confirms that engineering firms understand this clearly. Faced with severe talent shortages, 70% of AAC firms expect workflow automation to help them most over the next 18 to 24 months, automating repetitive tasks such as routine drafting and documentation to free up engineering capacity.

The report also shows that the larger the firm, the more likely they will use AI to generate and evaluate design alternatives faster, what is called optioneering. We saw this exact transition play out in our year Infrastructure awards where more than 25% of all submissions and close to 70% of all panalists use AI in their projects. The greatest economic leverage remains in design, where early engineering decisions have an outsized impact on an assessed total life cycle cost and long-term performance. Take AESA the winner in our bridges and tunnels category for the Porto metro expansion in Portugal. Traditionally, geotechnical modeling is so labor-intensive that engineering teams can only evaluate a handful of viable scenarios before project deadlines forced them to stop.

AI changes that dynamic entirely. AESA deployed AI agents to evaluate massive simulation outputs from our geotechnical solver plaxes, and run more than 100 simulations instead of just a handful. The result was not just a 30% reduction in design costs, but a significantly superior design that saved 4,500 cubic meters of concrete. What is remarkable about our Yi submissions is that because award entries reflect completed project milestones -- they were based on technology that existed 6, 12 or 18 months ago. When you consider the accelerating pace of AI development, that is a very long time. Back then, teams like IESA had to write extensive custom code to connect AI to ore solvers a capacity out of reach for most firms.

At YII, I use my keynote to show what is possible here and now when you use AI with or engineering applications at machine speed via MCB servers. Yet with all that generative power comes to create responsibility. In Infrastructure engineering, decisions carry real physical consequences. Precision is mandatory and the margin for error is 0. There is no room for AI sloppiness in our space. That is why we have to establish a clear division of labor. Optimization is the domain of the machine, navigating the trade-offs and holding professional judgment remains the domain of the engineer. This principle is fully in line with the AI road map for civil engineering in their AI boundary metric, ASC drew strict red lines Accountability is never delegated, and AI outputs in sealed work must always be verified by a licensed professional.

Our deterministic solvers are precisely the computational engine that engineers and AI agents rely upon through that optioneering iterating across thousands of physics-based simulations to converge on the better design. Finally, as our technology becomes increasingly operated at machine speed, business models must evolve to match with so much productivity being unlocked, engine firms recognize that their own commercial models have to adapt. The AEC advisers report shows that 63% of the largest engineering firms expect a significant move away from time and material billing over the next 3 to 5 years towards fixed fee and value-based pricing to capture the economic dividends of AI hard optimization.

And in our discussion with accounts, they recognize that as their own business models evolve, our commercial model will evolve as well. For now, our focus remains on driving adoption and validating these new agentic workflows. Monetization will come next. Seeing the immense potential firsthand, the sheer power of using our applications at machine speed and the profound value that will unlock across the entire infrastructure life cycle, you can count Bently firmly in the AI optimist camp. We are convinced that artificial intelligence will usher in a new era for infrastructure engineering. One where the capacity and capability of the profession will expand significantly to deliver safer, more resilient and higher quality infrastructure more efficiently.

Now turning to our business highlights for the third quarter. Our year-over-year ARR growth further accelerated to 13%, driven by broad-based demand across our core sectors and disciplined execution across all regions. Our net revenue retention rate reached 110%, underscoring steady expansion across existing accounts. Our renewal performance with E365 accounts also remained strong. even though Q3 is a small quarter for renewals, driven by solid flow uplifts as accounts commit to expanding their engineering capacity with Bentley. New logos contributed about 300 basis points of AR growth once again, primarily through our SME business. Through Virtuoso, our flagship commercial program for SMB accounts, we added over 600 new logos. The underlying sentiment across SMB accounts remains resilient with healthy project backlogs. SMB accounts are also showing increased interest in AI and want to engage with us on it. Turning to our tone of business by infrastructure sector. Resources was once again our fastest-growing sector overall. Sequent delivered another stand quarter driven by sustained strength in mining. Mining companies continue to prioritize brownfield optimization and expansion to meet the urgent global demand for critical minerals, supporting electrification, supply chain security and the rapid buildup of AI data centers.

We recently announced an acquisition to expand the sequence portfolio in resources, and we'll come back to that in a few minutes. Our largest sector, public works and utilities delivered another quarter of strong growth, fueled by sustained global infrastructure spend and grade modernization. The growth of PLS continues to stand out, driven by the global imperative to strengthen and expand the transmission grid. Our asset analytics business also showed strong performance securing strategic enterprise expansions in telecommunications tower inspection and transportation infrastructure AI. Growth in the industrial sector remains solid, in particular in power generation, while commercial facilities remained relatively flat, consistent with recent quarters. Turning to our total business by geographic region. In the Americas, our largest region, growth further accelerated across both North America and Latin America. In the United States, the macro environment remains very healthy.

Our engineered accounts continue to operate at capacity with robust backlogs sustained by continued federal and state public funding alongside substantial private capital deployments. The apartments of transport in to be a growth engine as they advance their digital delivery initiatives. The scale of private investment in AI data centers continues to generate tailwinds across our portfolio. Latin America delivered an exceptional quarter, led by mining and momentum across water infrastructure. EMEA delivered solid performance, supported by sustained infrastructure commitments and strong renewal execution across major accounts. Growth was once again led by the Middle East, despite the ongoing conflict with strong execution in the UAE and Saudi Arabia as major airport and transportation programs continue to move forward.

In the U.K., growth remained solid driven by major ongoing national programs, including [indiscernible] national highways and HS2. In Continental Europe, despite a traditionally slow quarter, we are encouraged by early signs of infrastructure funding deployment in Germany, where major rail and highway tenders are beginning to move forward. We also saw major European consultancies shift more engineering work to their delivery centers in India, moderating our consumption growth in Europe but helping fuel our acceleration in Asia Pacific. And in fact, Asia Pacific was our fastest-growing region this quarter.

India delivered very strong growth, driven by both massive domestic investments in water, create an offshore energy infrastructure and by its role as the global engineering hub for multinational firms. Australia and New Zealand delivered strong growth, underpinned by the broad resources sector and notable expansions from key mining accounts. In Southeast Asia, demand across regional civil projects and power networks drove strong growth as well. China, which represents less than 2% of our ARR continues to suffer from geopolitical and macroeconomic headwinds that we expect to persist in the foreseeable future. The strength across the remainder of Asia Pacific more than offset these headwinds. Finally, today, I want to briefly highlight our recent acquisition of Elis, now poucequen, or subsurface company. Sequel is dedicated to helping organizations understand the underground.

And to understand the subsurface, geoscientists relied on a variety of data collection methods. From trading boreholes and analyzing physical rock cores to measuring acoustic, magnetic and gravity signals from the surface. With EDS, we are adding another important signal to our portfolio, seismic. Seismic interpretation provides the highest resolution and the deepest geophysical picture of the subsurface. Historically, seismic methods have been at the forefront of oil and gas exploration, where Elis established its leadership. Yet as seismic data acquisition costs decline, and resolution increases, this method is proving increasingly essential across the broader resources sector. It is already opening new frontiers in energy such as identifying reservoirs for low-temperature geothermal energy and carbon capture and helping mining companies detect deeper critical mineral deposits, seismic interpretation also has direct applicability in civil infrastructure, where high resolution imaging derisk large transport corridors, complex tunneling projects offshore wind farm foundations and other critical assets.

By bringing seismic data directly into leapfrogs 3D implicit modeling, we are giving geoscientists and engineers a far more complete connected picture of the subsurface. This acquisition advances our mission to reshape how infrastructure and critical resources are delivered and optimized globally. And with that, over to you, Werner.

Werner Andre

Thank you, Nicolas. We are pleased with our performance so far this year. Our third quarter results extend the momentum of the first half with ARR growth accelerating once again and subscription revenues and profitability ahead of our expectations, both in the quarter and year-to-date. Total revenues for the third quarter were $422 million, growing 12.2% year-over-year and 12.7% in constant currency. Year-to-date, total revenues increased 13.2% or 12.3% in constant currency.

Our performance continues to be led by our mainstay subscription revenues which represented 92% of total revenues during the quarter. Subscription revenues increased 12.4% year-over-year or 12.9% in constant currency, reflecting continued strength across both our E365 and SMB initiatives. Year-to-date, subscription revenues increased 13.5% and was 12.7% in constant currency. In our small and less predictable revenue streams, perpetual license revenues for the quarter were approximately $13 million, up roughly EUR 2 million year-over-year and approximately $31 million year-to-date, down about $0.5 million and representing approximately 2% of total revenues. Services revenues increased 8.4% or 8.8% in constant currency driven by maximum related services within cohesive.

Year-to-date, services revenues increased 15.2% or 13.8% in constant currency. As Greg described, we completed the divestiture of cohesive on October 13. Year-to-date, cohesive represented approximately 50% of our services revenues and approximately 3.5% and of total revenues. Beginning in the fourth quarter, with the Aforexpect services revenues to decline year-over-year through 2027. And and to represent less than 3% of total revenues compared with approximately 5% year-to-date. While cohesives departure reduces our revenues, what remains is of higher quality. Subscriptions will represent approximately 95% of our total revenues, up from 92%, making our revenues even more recurring and predictable.

As of September 30, our last 12 months recurring revenues were $1.528 billion. an increase of 13.2% year-over-year or 12% in constant currency and represented 93% of total revenues. Our last 12 months constant currency account retention rate remained consistent at 99%, and our constant currency net revenue retention rate increased to 110%. The combination of our high retention rates and new business momentum gives us confidence in the continued durability of our recurring revenue growth. Now turning to ARR, we ended the third quarter with ARR of $1.583 billion at quarter end spot rates. On a constant currency basis, our year-over-year ARR growth rate accelerated to 13%, ahead of our expectations and above the high end of our annual outlook range. marking the second consecutive quarter of acceleration this year.

Year-over-year growth included approximately 40 basis points from our talent acquisition, unchanged from the prior 2 quarters. Sequential constant currency growth in the quarter was 3.1%, all organic compared with 2.2% in the third quarter of last year. and has exceeded the prior year quarter in each quarter of 2026. Asset Analytics contributed meaningfully to both our year-over-year and sequential growth with large contract expansions from existing accounts accounting for approximately half of the improvement in sequential growth over last year. Our GAAP operating income was $123 million for the third quarter and $337 million year-to-date. As I have discussed previously, our GAAP results can be impacted by deferred compensation plan revaluations and other acquisition-related items.

Moving to adjusted operating income less operating stock-based compensation, which was $135 million for the quarter, with a margin of 32.1% and compared to 28.1% a year ago and $392 million year-to-date with a margin of 31.2% compared to 30.7% a year ago. As planned, we invested earlier in the year than in 2025. So our first half margin trailed the prior year and our second half margin improvement is correspondingly higher. Beyond this planned pattern, our results were ahead of our expectations, reflecting a mix shift towards higher quality subscription revenues compared to our original plan, along with with disciplined cost control.

Our year-to-date G&A also reflects investments in our new enterprise wide finance and quote-to-cash platforms, which went live during the second quarter, driving a step-up in costs around the go-live that was contemplated in our outlook. G&A growth moderated to 14% in the third quarter from 27% in the first half. Notably, we did not adjust these costs out of our profitability, we absorbed them within our margin commitment, underscoring the quality of our year-to-date performance while positioning us for greater efficiency and operating leverage. We remain well positioned to deliver on our AOI less operating SBC outlook including our annual constant currency margin improvement. Our free cash flow for the quarter was $101 million, reflecting working capital timing. With the year weighted towards a stronger fourth quarter on the last 12 months basis, free cash flow was $489 million, up 6% year-over-year. Year-to-date, free cash flow was $353 million, down 8%, reflecting 2 factors we have discussed, exceptionally strong collections at the end of 2025, which created a tougher comparison and our plan to invest earlier in the year than in 2025.

We continue to execute a disciplined and balanced approach to capital allocation. Year-to-date, we reduced net debt by $88 million, while returning capital to shareholders through $168 million of share repurchases, up meaningfully from the prior year and $64 million of dividends. Our balance sheet supports significant strategic flexibility. We ended the quarter with cash of approximately $150 million and $1.26 billion of availability in our revolving credit facility, and our net debt leverage was 1.7x adjusted EBITDA down from 1.9x in prior quarter. With our cash interest exposure mitigated by the low fixed coupon on our convertible notes and our $200 million interest rate swap expiring in 2030. This provides ample capacity to support our strategic priorities, including programmatic acquisition such as Elise addressing our mid-2027 convertible notes at maturity, ongoing share repurchases and dividends.

Moving to our outlook. As discussed, year-to-date, we have been exceeding our expectation in constant currency ARR growth, subscription revenues growth and profitability. Our original outlook did not contemplate the divestiture of Cohesive. In order to provide a like-for-like comparison, we are presenting our updated 2026 ARR growth outlook excluding the impact of the divestiture of cohesive. On this basis, given our year-to-date performance, expected continued momentum, like the programmatic acquisition of Elis, which we expect to contribute approximately 100 basis points, we are raising our outlook to 12.5% to 13.5%. And above the high end of our original range of 10.5% to 12.5%. Including the impact from the divestiture, we expect reported ARR growth to be approximately 100 basis points lower.

We are also updating our revenue outlook for the divestiture of cohesive, which we expect to reduce fourth quarter revenues by approximately $12 million, about 3/4 of it in services. Our updated outlook for total revenues at the exchange rates assumed in our 2026 outlook, is $1.673 billion to $1.73 billion. With this, we now expect our constant currency subscription revenues growth towards the high end of our 11% to 13% outlook range and constant currency perpetual licenses and services revenues to remain approximately flat, reflecting the divestiture. We are maintaining our AOI less operating SBC outlook range of $495 million to $510 million. and we continue to expect free cash flow within our outlook range of $500 million to $570 million, including the impact of higher cash taxes related to the cohesive transaction. As in prior years, our fourth quarter free cash flow will depend on the timing of year-end collections relative to our January 1 renewal cycle.

For modeling purposes, we now expect cash taxes of approximately $55 million to $60 million and weighted average diluted shares of approximately $319, million reflecting our higher share repurchases. And also for model purposes, year-to-date through September 30, cohesive contributed approximately $35 million of service revenues, and approximately $10 million of subscription revenues. With regards to foreign exchange rates, for the third quarter, the U.S. dollar strengthened relative to the exchange rates assumed in our 2026 annual financial outlook, resulting in approximately $8 million less revenues from currency. If end of October exchange rates were to prevail throughout the fourth quarter, our revenues will be negatively impacted by an incremental $5 million, relative to the exchange rates assumed in our 2026 outlook. And with that, over to Eric for Q&A. Thank you.

Eric Boyer

[Operator Instructions] The first question comes from Jo Vruwink from Robert Baird.

Question-and-Answer Session

Joseph Vruwink

Great. Thanks very much for the time today. I wanted to ask about some of the acquisition activity in the [indiscernible] space, of course, you talked about your recent activity, but I think there's been 3 acquisitions now since 2025 of some of the more platform companies. And I'm curious what you make of this, obviously, you kind of saw this coming because CPM was way back in 2021, but in terms of the recent M&A activity, what is that maybe a signal about a desire of operators have our software exposure or what AI it might ultimately end up meeting to this specific product portfolio.

Gregory Bentley

Nicolas, maybe you'll speak more to this, but my observation would be, once again, as in the case of EAM, there's -- there are ERP systems for mining in effect that involve for administrative activities, and we don't cover it being in that space. It's the 3D engineering where we have most to add in the geosciences aspect. And otherwise, the competitors are formidable there. But the 3D modeling is unique in mining in that every scoop of earth or substrate gives you new information that you can incorporate right away. And that is our contributing differentiator in mining.

Nicholas Cumins

Yes, I will add that no surprise, there is so much activity going on because mining is very strong. Now depending on which acquisitions you're referring to -- the way we're expanding our portfolio in mining with Sequans as you could tell since the time of the acquisition more than 5 years ago now, is by adding more capacities that help better understand what's in the ground. So we've added capacities to process signals like magnetic signals, for example, or electrical. And with the acquisition of IDS now, it's a seismic -- seismic interpretation, right? So we have a pretty broad and deep portfolio of capabilities to help better understand what's going on in the ground, which is of critical importance for mining for sure. But it is a whole platform for us to grow beyond mining and resources. Idis, as I mentioned in my prepared remarks, is very strong in oil and gas.

And those caparities of EDS seismic interpretation are necessary for us to grow beyond high-temperature geothermal energy production into low temperature. But those capabilities are also very relevant for Civil, and that's what makes us quite distinct as well, which is -- we take all of the sequence type of capabilities to understand the underground and make them available in civil workflows right now in order to derisk project delivery and ensure long-term resilience of infrastructure assets.

Eric Boyer

Next question comes from Matt Hedberg from RBC Capital Markets.

Matthew Hedberg

First of all, congrats on the results, guys, really great to see I guess I'm curious, there was a lot of commentary. I think most of it was from Nicolas on AI. It really does seem like you guys have continued to lean into that from a functionality perspective and it's resonating with clients, and I guess from a monetization perspective, I think, Nicolas, you said it's going to eventually come. It's going to take time. But I'm just kind of curious with the good results here, it sounds like acid analysis is the real start of the quarter. Are you seeing any indication that AI is positively becoming a tailwind to growth today? And just as we start to think about the next few years, how should we think about AI monetization kind of feathering into growth?

Nicholas Cumins

Yes. So AI -- in terms of monetization, this is already happening indeed with asset and ethics, as you pointed out. Our Astec offering is leveraging AI to help owner operators better understand the physical commendations of their infrastructure assets. And Q3 was a great quarter for Asset Analytics. We did a number of deals with existing accounts. both for telecommunication tower inspection and for road monitoring, so modernization for AI is already happening there. That really reflects by the way, where we are with AI overall in the infrastructure ecosystem. The first few cases started many years back in operations and maintenance, and that became the angle that we took with Asset & Analytics.

Now when it comes to AI in design, this is still very early stage, and we are really focusing right now, we need adoption, exploration, validation. We are making sure that from a technical standpoint, we will be able to start monetization in 2027. However, as we said before, we don't expect monetization of AI in design to be a material growth driver in 2027. It will be the year where we start -- the decision to start will be really based on whether accounts are moving now from exploration to production yes, number one. And number 2 is how fast are they going to adopt or new applications that have these MCP servers in order to be able to use these applications and machine speed with AI, yes? So that's where we are.

Our focus again is all about adoption. The feedback that we're getting from the accounts as you said, is very positive. When they understand, they realize the potential here, which is unlocked, they're truly excited and so are we.

Eric Boyer

Next question comes from Jason Celino from KeyBanc

Jason Celino

So I wanted to ask about the acquisition, Elis. No secret has been growing, I think, 2x your corporate average. Maybe can you share some financial profile metrics of the acquisition, if there's any customer overlap with [indiscernible] or yourselves? So any color there would be helpful.

Nicholas Cumins

So the core business of Elis is oil and gas. So there is overlap actually with traditional accounts of Bentley more than of [indiscernible]. From a financial standpoint, it is accretive to our AR growth. And it's roughly in line with us from a margin standpoint. So it really is a very high-quality asset that we were able to acquire here.

Gregory Bentley

I think that it's accretive to our revenue multiple as well, reasonable from that standpoint. And then in terms of overlap, it happens that the oil and gas owner operators as we think of them are now all of the above energy operators, and they are interested in applying sequence modalities beyond seismic to opportunities such as geothermal and carbon capture. So we have now the account access to go beyond pilots into scaled offerings for the oil and gas energy -- diversified energy providers as they diversify into other underground sources. .

Eric Boyer

The next question comes from Daniel Jester from BMO.

Daniel Jester

All right. I wanted to go back to the slide in the presentation that talked about sort of how different companies by size, expect the change in their business models. And it looked like there was a lot of diversity in terms of expectations enterprise, I think, 60% expect a big change in SMBs less so. So I guess a 2-part question. I guess, -- number one, any hypothesis as to why there's such a diversity in terms of expectations for how those business models may change? And then two, as your enterprise customers do change their business models, how may you have to change yours to meet their future needs.

Nicholas Cumins

Good question, Daniel. So on the first one, the way I would interpret that difference is that the larger firm that was another slide -- the larger the firm, the more likely they're going to be starting to use the AI in design to not just automate, but to optimize designs. And my [indiscernible] on this and also talking to larger accounts. is that as they do this, then they really realize the potential, and based on this, they also realize that a business model which is primarily based on time and material is probably not right for them to capture their fair share of the value that's going to be created, that the conversation has to shift to value and outcomes. .

And so there is -- that the incentives are perfectly aligned with what ultimately is good for the owner operators, which is optimal designs. So that's that. Then on the second one, as a counsel is saying, yes, they're going to have to shift to something which is maybe fixed price or indeed outcome based, et cetera. There's not necessarily an implication that our business model will have to change to a committed for that. I think these 2 things are actually quite compatible. The relation, however, it's a little bit the same for us, which is if our commercial models are somewhat user-centric and somewhat based on how much time users are spending with our applications -- then going forward, this might not be the right way for us to capture also a fair share of the value that's going to be created. And we have to shift to something else.

So it's actually very interesting in this conversation with the accounts because there's a lot of reciprocity, right? They realize that they have to evolve their business model. And based on this also understand that we also have to evolve our business model. that especially with E365 as a program, application days, it is a proxy for how much time in a way. Engineers are spending with our software, and if they start to use the application on a machine speed, then on top of that, there should be another price.

Eric Boyer

Next question comes from Sitikantha Panigrahi from Mizuho.

Sitikantha Panigrahi

Great. Congrats on the quarter. If I asked the ARR growth that reached almost at least 13% and about half of that better sequential growth came from large expansion in existing accounts with asset analytics contributing that. So for this year, the question is how much of that momentum is already locked into Q4 given your outlook now implied 11.5% to 12.5% this? And for 2027, how do you rank the drivers like uplift in renewals, new logo and asset analytics and now [indiscernible] you talked about AI-related consumption -- so which of them do you expect to add growth versus just sustain it heading into 2027?

Nicholas Cumins

So Siti, first of all, we definitely have a lot of momentum. And our results in Q3, 13%. This was really based on strengthening resources, strength in public work utilities and large enterprise deals in asset analytics. And that was all organic, no acquisitions. And we have 13%. So for the remainder of the year, in the updated outlook that we gave -- this does factor in already then the acquisition of IdeS. And for us now to be able to get to the top end of that range, we will need to continue to do very well in resources and fabric work and utilities. We'll have to do very well with our renewals. And you know that Q4 is our biggest quarter for 365 renewals, and we'll have to bring again important tier enterprise deals for asset analytics. .

So obviously, this is not the time for us to talk about our outlook for 2027. This will be the focus of our call when we review our Q4 results. But we don't see, at this point, given the strength of the end markets and the strength of us as a company or competitive strength, we don't see why we show goes away from our long-term financial framework and that we've used in all the past few years.

Eric Boyer

The next question comes from Jay Vleeschhouwer from Griffin Securities.

Jay Vleeschhouwer

Thank you. In the prepared remarks, you made some reference to Bentley infrastructure cloud adoption, I believe, among some of your largest customers. could you speak more broadly about infrastructure cloud adoption across all or most of your cohorts and the adoption dynamics in other words, is it the case that Bentley Infrastructure Cloud helps you induce more applications adoption? Is it really more the reverse where customers buy enough applications and they adopt infrastructure cloud. .

And then secondly, you referred to the Navion relationship is representing not just cohesive but some of your apps. Is that going to be a significant enough channel partner relationship that you're percentage of revenue from the channel could slightly start to increase more meaningful lag.

Nicholas Cumins

On the first question, maybe Greg will address the second one. So on the first question, Bently Infrastructure Cloud. The reason why we grow with Ben Infrastructure Cloud is, number one, our openness. The fact that we actually work great with Bentley and Non-Bently applications. We work even better with Bede applications, yes. But we work great with nonmetapplications. And so Bently infrastructure cloud is just a fantastic platform for collaboration, and data exchange for entire infrastructure value chains across engineering disciplines, across engineering organization because of that -- the strength of that openness. And the other reason why we are winning more and more is because of the strong stance that we took when it comes to data where we basically said that the data that is embedding fast cloud is our users' data always.

So we don't use it to train our own AI models. We don't use it to train AI unless there is explicit direction from the accounts, and that makes us quite distinct from other platforms that are there, as we discussed in previous quarters.

Gregory Bentley

And I'll say that my reason to bring up Bently Infrastructure Cloud, in particular, as to the footprint of our owner-operator accounts, that half of our business. It's mostly project-wise in those accounts. And so they have a Bentley infrastructure cloud schema that they can interrogate and query, including autonomously, to utilize and maintain their assets during operations and maintenance, but until now they don't. But they're just 1 step away from it. And my point is that asset analytics provides the Keystone that 1 step where they can have an as operated physical condition of their asset to feed back into resimulating and remodeling based on the design information they have already.

So it's become possible now to have instant on digital twins for these owner operators with AI and asset analytics. In the case of Navion and Maximo, of course, as we've learned as anything but instant on when you're adapting an enterprise system to expand from maintenance information to engineering information to inform the maintenance Navion is a really impressive outfit run by entrepreneurs who are looking to make more of Maximo. And it's early in the relationship. Obviously, the divestment just closed, and we're going to be on their board and encouraging their brief to include digital twin, as I say, a digital twin integrator as well.

There really is an opportunity for that, but it's more of a long-term opportunity, I acknowledge.

Eric Boyer

Next question comes from Andrew DeGasperi from BNP.

Andrew DeGasperi

So I wanted to touch on a bit in terms of your strength in the Americas, and I think you mentioned that data centers and directly benefited that region. Obviously, there's been a change since Tuesday in terms of lease to high-level headlines in terms of -- particularly in certain geographies of the U.S. like Texas, and there's some concern about the slowdowns or project delays with regards to data centers. So just curious to know if what you make of that in terms of what is your exposure there? Like do you think this -- there could be in terms of at least where your products are deployed, there could be some impact. And maybe if you can -- given you've been around for 40 years, I'm sure you've seen these type of cycles before. Just curious to know -- what do you -- what could we expect in terms of that in the future?

Gregory Bentley

Let me address the election generally. And of course, we're familiar with divided government in the U.S., but it's about to thoroughly take place after this lame-duck session. What we think is that infrastructure investment is where there's likely to be bipartisan concurrence. Of course, the IIJ ended after 5 years at the end of September, the projects in process, of course, will be completed. It's been extended. It's spending has been extended until December 11, most of that spending, and that could go on further. But in the interim, there's been something called the Build America 250 bill, which in its committee in the House, was approved by a 62 to 2 vote. So it's a bipartisan opportunity.

Now we don't think that's going to be acted on in the lame duck session. When it comes to next year in the new Congress, it's likely that, that already adequate spending for surface transportation roads and bridges to that, we'll probably be added now rail and transit spending as well. And in general, the Democratic-led congresses have had have tended to fund infrastructure at greater degree. So that probably lies ahead. What's very interesting is permitting reform, which is important to a lot of what we do. And that has finally come so far that a bipartisan bill again, has been reported out of all the committees, the applicable committees in the Senate, and it's actually scheduled for a vote on next Monday, to come -- to bring it to a debate on the full floor. And it's not out of the question that, that could be enacted because it is a bipartisan priority even during this year still because it's thought by -- although there are some environmental groups in opposition, it's thought to be about the best that's going to be achieved by many who have to make that decision.

In Congress also is the biennial water bill that has bipartisan support and will, for the first time, include funding for digital initiatives, including digital twins. And then finally, back to your question about data centers. It's likely that, as you say, across the country and in various jurisdictions, there'll be more scrutiny of the environmental impacts. These are impacts having to do with energy provision, with water provision, the surrounding infrastructure. That's where our software is utilized. And we can imagine site plans for data centers now and needing to include infrastructure digital twins and we actually see that as an opportunity.

Eric Boyer

Next question comes from Kristen Owen from Oppenheimer.

Kristen Owen

Congratulations on the nice results. And Nicolas appreciate the Spiderman reference that you gave us. You talked about this in some of your prior questions, but I wanted to put a finer point on it. Just help us understand with the lease acquisition, where that creates the most immediate commercial opportunity. Is it selling into sequence existing tools, selling sequence into those existing lease customers, is it bringing that seismic interpretation into your mining customer base? And how do you think about those synergy opportunities contributing to incremental ARR growth?

Gregory Bentley

Kristen, I'm going to jump in to say that what I learned about seismic interpretation is that, of course, you make the earth rumble and listen to what happens, it used to require a tractor trailer and a big rig to generate these waves. Now it's down to something that can be towed behind a golf cart. So it's applicable to civil projects in general. And probably that lower cost of acquisition will continue to be more advantageous. So to me, the big opportunity is applying it where it's been prohibitive in the past. And that -- there's many other such modalities where that's been the case over the life cycle of Sequent. Sorry, Nicolas, so I just had to say I'm imagining creating a seismic wave with a golf cart size machine as something that sticks in my head.

Nicholas Cumins

All good. Kristen, and I will say energy, first of all, because Egis is already strong there, and there's so much more to do. As I said, this company on its own already growing quite fast and will be accretive to our AR growth. So there's more in energy, traditional energy sources, but new sources of energy as well. I will price saining second because seismic allow you to find minerals at a much deeper, deeper level and with a much higher resolution than is possible with traditional methods or other methods and then Civil. .

Eric Boyer

Next question comes from Alexei Gogolev from JPMorgan. Actually, I'm sorry, I think it's Faith Brunner from William Blair.

Faith Brunner

Can you hear me okay? Just wanted to double click on some of the conversations you guys are having with your design customers in regard to AI, there's a lot of interest. There's a lot of excitement about the opportunity, but is there anything that's maybe driving a wedge between the willingness and readiness in the sense of you talk to how business models may have to change around material in time and different optimization. But is there anything else maybe on the customer end from a change managment perspective that we should be aware of that might add some time between initial interactions and a more scaled adoption of these AI solutions.

Nicholas Cumins

I will say 2 things. First 1 is education. And I mean it with a ton of respect but education on what's really possible here and now. This is a big priority for us with our marketing team. It's also a big priority with our solution engineering team. Those are what we consider 4 depot engineers who are working hand in hand with our accounts to show them what's possible right now. And the second thing is what I indeed touched on earlier when we talked about AI monetization and API monetization going forward, we're going to be dependent on how fast they adopt the new applications, the new versions that have these APIs that are instrumented and they have these NCP servers. I think as they realize the potential, they will have a big incentive to move faster to these new versions that have had done in the past. .

Eric Boyer

Next question now is from Alexei Gogolev from JPMorgan.

Alexei Gogolev

So Bret, you described Asset Analytics as the Keystone to close the CapEx to Opex, strategically, do you see asset analytics becoming a stand-alone platform, category inside mentally or remaining capability delayer distributed across each domain portfolio.

Gregory Bentley

Well, we're managing it as an asset analytics group at the moment so that they can move quickly, and as you know, we've grown by acquisition in asset analytics because we don't want to start from scratch where we don't have to, with a particular AI for a particular asset types, like Talend brought us distribution poles and so forth. And I think that will remain the case for a while. It's different in that we charge per asset per year, not per user per day at the enterprise level. It's incremental in that respect. And the competitive set is different. It tends to be hardware-oriented competitors at the present time.

Again, what's our differentiation, which is the reason I referred to it as a keystone is not only do we capture the current physical condition of an asset, but we feed that back in to resimulate or remodel to remodel and resimulate -- does that mean it's still safe? Is it still resilient? And what maintenance action would be sensible based on what's worked elsewhere. So our differentiation is over on the -- what we have thought of as the CapEx side of our business, but the modeling and simulation, which the owners own and paid for and have, they have never used during the operating life cycle. So asset analytics is meant to open that up. But for now, organizational, it's a platform in its own right where of course, the digital twin information has persisted in Bentley Information cloud, but we're competing in major procurements with other data acquisition approaches, and that requires some specialization.

It's a distant sales group now and so forth. Over time, as you know, I think it will -- it can grow -- should grow to be comparable in scale, our revenues per asset should be comparable to our revenues per user over time, but that's -- we'll have to complete the portfolio to get to the other asset types of course, along with the organizational development that's underway.

Eric Boyer

Next question comes from Taylor McGinnis from UBS.

Taylor McGinnis

So the 13% ARR growth was clearly the highlight and really strong. So I believe constant currency net new ARR was actually over $50 million or so higher than it was last year. So could you just help us understand like in that incremental bookings that you guys have -- how much of that is coming from asset analytics, sequence momentum or any other areas? And then second, if we look at the 4Q guidance range, it actually seems that you expect that net new ARR, our booking strength, growth to continue on a year-over-year basis despite some of the tougher sequent comps. So I guess what are you seeing at the start of the quarter that's giving you that comfort

Gregory Bentley

Andre do you want to address it?

Werner Andre

Yes. Happy to. So Q3 was raising 150 basis points is the combination of, I would say, 2/3 is just better-than-expected underlying performance led by Resources grid and Asset Analytics. Asset Analytics was particularly strong compared to Q3 last year and year-to-date -- compared to year-to-date last year. And then the Ellis acquisition, as I mentioned, contributed about approximately 100 basis points of ARR growth. We will see that being added in Q4. And we had about 40 to 70 basis points in our outlook for programmatic acquisition. So Elis is contributing approximately 1/3 of the overperformance that we see year-to-date. And going into Q4, I think to get to the high end of the range, I think Nicolas hit on that before. it just will require the continuous strength across like resources, public rux utilities and especially grid just was our strongest renewal quarter.

So we need a strong renewal quarter, but we have reasonable visibility into the pipeline. And then it will require larger asset analytics deals. Asset Analytics had a particularly strong Q4 last year. But as we said, like if all that aligns then we -- I think we will see a solid Q4 again.

Eric Boyer

Next question comes from Joshua Tilton from Wolfe.

Joshua Tilton

Awesome. Nothing like Taylor stealing my exact question right off my computer screen before I have to ask one. Maybe to step back then, like more of a high-level question, Nicolas, I really appreciated you addressing kind of the business model transition that has to happen, but if we step back, it sounds like what you're saying is as you continue to embed AI more into your products, you're essentially selling your customers the ability to be more productive and automate more of the processes. But right now, your customers are kind of charging their customers by the outdoor.

So those incentives might not exactly be aligned. I guess what I'm trying to understand is what -- at least from my perspective, that kind of feels like a big shift, so what kind of gives you the confidence that the industry is going to get this pricing shift correct? And like why is this not like a lot more longer-term dynamic than a shorter-term dynamic? Does that make sense?

Nicholas Cumins

It absolutely makes sense, yes. And I'll try to be super brief. So you've characterized it well. We've always been in the business of making engineers more productive since 1984. Obviously, the kind of productivity gains we're talking about here, both in efficiency and effectiveness is considerable. And we do see a clear correlation for those accounts who have already shifted their business model when they can to something like fixed price or outcome-based, they lean on the eye much harder than others.

So when they can because ultimately, it will depend also on the own operators that don't completely dictate on their own as engineering firms. -- we're surprised talking to CEOs of large engineering firms at AC advisers just last week, of how many of them have already made a significant transition to business models that will be -- there will be no incentive for them to become way more productive using AI.

Gregory Bentley

Joshua, I think you could say that the transition to outcomes-based is hard and may take longer. The transition to fixed price is a waste station that's relatively easy. And there's no pushback, less pushback to that.

Eric Boyer

Next question comes from Tomer Zilberman from BofA.

Tomer Zilberman

I wanted to ask maybe a more conceptual question around AI. If you look at the data from the labs, it shows that maybe some of your core customers kind of civil engineers and the like are using AI a lot less than some other knowledge work type roles. I think it's like 30 to 300x less. So the question is 1 part of it is the data is a little bit stateless as of a few months ago and models have improved pretty significantly in terms of engineering capabilities. But the question is when you're talking to your customers and they're talking about interest in AI, is it more so about embedding AI into your existing products? So it's almost as if they're not using AI as the interface? Or are you seeing increasing demand either for stand-alone AI applications from yourselves? Or even the concept of head list where maybe they're interacting with the third-party model, but they're connecting to you guys at the back end?

Nicholas Cumins

So when we look, for example, at the results of the AC advisor survey, we can see that we still had in the early stages of adopting AI when it comes to design. And the use cases, the first use case is that engineering firms went after where use cases that you could argue are really about knowledge workers. It's really about how to help them find information faster about the project, understand better the product requirements and so on and so forth. When it comes to using AI for the actual design, so to automate pieces of the design, or to be able to optimize design, definitely very early stage.

And depending on the sophistication of the accounts, we will hear either count so say, Bently, you please do this. We have our own Bendeka pilot, which is leveraging third-party L&M. So yes, we're going to use Bendeka pilot together with your applications. The more sophisticated ones are saying, no, we actually have our own. We basically picked whatever AI harness, Microsoft Copilot or Claude or whatever. And we have made it at such that it has so much scale that really understands how we work as an engineering firm, and we want that AI harness to be interacting with your applications. So those are the most sophisticated ones. But that's why it was very important for us very quickly to clarify that -- it has been a long tradition for Bentley. We embrace openness that we will have an open architecture when it comes to AI interacting with their applications that you will be able to use that LLMs are provided to us than our own [indiscernible]. You'll be able to use, obviously, AI models that we create ourselves, but we will also be able to use third-party AI models with our applications.

Tomer, one more thing. Actually, if you do get the AEC advisers report, you will see some questions regarding opens, and then we'll reinforce what I just said.

Eric Boyer

Great. The last question comes from Mason Marion from Cantor.

Mason Marion

So in August, you installed a new Chief Revenue Officer, have any changes been made thus far to your go-to-market strategy? Or do you expect any material changes over the next year?

Nicholas Cumins

We don't expect to change, we expect a reinforcement of what already had started with the previous Chief Revenue Officer, who had hired our new Chief Revenue Officer and prepared him as a successor. Andy, our new Chief Revenue Officer, his background is technical sales. And among other things, he has a very rich background, a lot of experience. But at Bently was in charge of solution engineering. That's the team I referred to earlier. -- saying those are the 40 float engineers, right? And through Andy's leadership, we have significantly improved the feedback loop between our accounts and our product teams, best solution engineering, basically being the keystone here in that feedback loop. .

And that is so important right now given the profound technological change going on with AI that we have this very close feedback loop. So you can expect more of that with Andy as our Chief Revenue Officer.

Eric Boyer

Sorry. One last one from [indiscernible]

Unknown Analyst

just on the asset analytics , I just wanted to touch base. Last quarter, you said you were running around $50 million in revenue. I'm not sure are you going to update that annually or sort of what's what kind of metrics we can get around the asset analytics progress? And what are you doing there in terms of go-to-market partnerships. This Obviously, the cell tower business is kind of more concentrated, but road maintenance is like huge. So how are you guys going to market and leveraging the technology with partners?

Nicholas Cumins

Blair, I think the $50 million number, this was our run rate at the time of the [indiscernible] acquisition. So last year, Q4, but it was not all ARR at that time. That's why we said run rate because some of the revenue could not be qualified as ARR, so when it comes to decommission towers, we have 2 sorts of growth here. One is winning new accounts, and we talked about last quarter of a very important telecommunication entire owner who is looking to use our, and will use our solution and is actually now using our solution to do regular inspection of the tariffs. But we also have equipment manufacturers, and in this quarter, we had significant expansion deals with equipment manufacturers that were already accounts of ours, so that's on telecoms tariffs.

We do see a natural expansion from telecommunication towers into the electric grid starting with the distribution tariffs. That was 1 of the main objectives of the acquisition of Talon because there were already supporting use cases where their technology was deployed on distribution towers. So that's a natural expansion. And then road and monitoring our focus remains, but not exclusively, but the primary focus is still with the DOTs and there's still a lot to do with the DOTs and we had some significant expansion with DOTs in Q3.

Eric Boyer

And that concludes our call today. We thank you for your interest in Bentley Systems.