Bentley’s YII 2026 Finalists Reveal the Next Stage of Digital Infrastructure

Bentley’s YII 2026 Finalists Reveal the Next Stage of Digital Infrastructure

A shortlist is rarely read as a market indicator, yet the field Bentley Systems has just named for its 2026 Year in Infrastructure Awards works better as economic evidence than as a ceremony programme. Of the 36 finalist projects, 24 reported using artificial intelligence in some form and 22 showed strong or supporting evidence of a connected digital foundation.

Drawn from more than 300 submissions across 53 countries and 12 categories, that is no longer a story about early adopters experimenting at the margins. It is a demand signal from the projects that owners, contractors and engineering firms consider their most significant, and it lands at a moment when the software vendor behind them is turning exactly that behaviour into durable, recurring revenue.

The commercial thesis worth holding onto is straightforward. AI and connected data have crossed from pilot to baseline expectation on serious infrastructure programmes, and the value in the sector is migrating from the one-off design deliverable towards the persistent data asset that survives into operations.

Bentley has spent the past two years restructuring its product line, its acquisitions and its pricing around that migration, while competitors including Autodesk, Trimble, Nemetschek and Hexagon manoeuvre for the same ground. For infrastructure owners and delivery organisations, the finalist field is a useful mirror. It shows what a competitive digital operating model now looks like, and it hints at where procurement, investment and asset strategy are heading next.

Briefing

• Two-thirds of Bentley’s 36 YII 2026 finalists used AI, a marked jump from 2025, when roughly half of finalists incorporated it, signalling that AI on flagship projects has shifted from novelty to expectation.
• Bentley closed 2025 with total revenues of about 1.5 billion US dollars and annualised recurring revenues of 1.46 billion dollars, with subscriptions now around 92 per cent of revenue and net revenue retention of 109 per cent.
• The company is monetising the shift through next-generation AI applications such as OpenSite+, the Bentley Copilot and AI-powered search in ProjectWise, and through an Asset Analytics push extended by the December 2025 acquisitions of Talon Aerolytics and Pointivo.
• Bentley estimates its infrastructure engineering software market at roughly 29.8 billion dollars, and positions itself in horizontal infrastructure, where MicroStation remains a Department of Transportation standard, against Autodesk’s building strength and Trimble’s field and hardware reach.
• Finalists will present at the Fairmont Singapore on 6 and 7 October 2026, in a programme that has now hosted more than 5,500 infrastructure projects across two decades.

AI Crosses From Pilot To Production

The single most telling number in the finalist field is the rate of change. At the 2025 event, then still branded the Going Digital Awards, Bentley reported that nearly a third of submissions and almost half of finalists had folded AI into their projects. A year on, two-thirds of the finalist projects report using it. That escalation matters more than any individual application, because it marks the point at which a technology stops being a differentiator that wins an award and becomes a competence that firms are expected to have. On the projects that engineering leaders put forward as their best work, AI is now closer to a default than an experiment.

Bentley has built its recent product roadmap to capture that transition rather than merely celebrate it. Its next-generation applications lead with OpenSite+ for civil site design, which the company describes as its first engineering tool built on generative AI and which it claims can deliver projects up to ten times faster without loss of accuracy. Alongside it sit the Bentley Copilot, a context-aware assistant that can query and edit models through natural language, a drawing-annotation AI agent that reached general availability in November 2025, and AI-powered search inside the ProjectWise project delivery platform.

A revealing engineering detail sits under the bonnet. The Copilot is built on commercial large language models and is deliberately model-agnostic, so Bentley can switch between providers such as those from OpenAI or Anthropic as performance shifts, a design choice that keeps the company’s productivity gains insulated from any single AI supplier.

The deeper commercial driver is not novelty but scarcity. Chief executive Nicholas Cumins has repeatedly framed engineering capacity as the sector’s binding constraint, arguing that there are simply not enough engineers in the world to deliver the work that needs doing, and casting AI as a way to empower infrastructure engineers, not replace them. That framing changes how the software should be read on a balance sheet.

When a productivity tool addresses a structural labour shortage rather than a discretionary efficiency wish, it moves from a nice-to-have line item towards a necessary one, which is precisely the kind of demand that underwrites subscription renewals and pricing power. The finalist field, dense with AI-enabled delivery, is what that thesis looks like in the wild.

The Connected Data Foundation Becomes The Real Asset

If AI is the headline, the connected data foundation is the more consequential story for asset owners. Twenty-two of the 36 finalists demonstrated a connected digital base underneath their projects, and that plumbing is what turns clever automation into lasting value. Bentley’s strategic argument, sharpened across its 2025 and 2026 events, is that infrastructure is moving away from static project handovers and disconnected files towards digital twins that persist and evolve alongside the physical asset.

The company’s chief marketing officer, Cate Lochead, put the direction plainly, describing a shift from disconnected files, manual workflows, and periodic inspections to connected information, intelligent automation, and continuous insight, and concluding that the future of infrastructure is not just more digital, it is more intelligent.

Several finalists show why owners care about that distinction in hard commercial terms. Evides, the Dutch utility, is recognised for operational digital twins spanning a 14,000-kilometre water distribution network, the kind of always-on model that changes maintenance economics rather than merely documenting design intent. SUDOP PRAHA’s digital twin of Prague’s main railway station and Pinnacle’s twin of its Madurai campus point the same way, treating the model as an operating tool rather than a deliverable to be archived once construction ends.

The distinction is where lifecycle cost is won or lost. A connected foundation lets an owner interrogate current conditions, test interventions and plan capital against real asset performance, which is a materially different proposition from inheriting an as-built file that begins ageing the day it is handed over.

Bentley has wired this thinking into its platform through the iTwin digital twin technology, the Bentley Infrastructure Cloud and a connected data layer the company calls Bentley Cloud Connect, which allows project information to be crawled for insight from design through to construction sequencing. That architecture is also what makes the AI useful, because an agent is only as good as the data it can reach.

The commercial implication for delivery organisations is that the competitive unit is shifting. The advantage no longer lies solely in producing a good model, but in maintaining a connected, queryable data estate that compounds in value across the asset’s life.

Where Bentley Is Placing Its Commercial Bets

The finalist field flatters Bentley, but the more instructive picture is in its accounts and its capital allocation. The company closed 2025 with total revenues of roughly 1.5 billion dollars, up 11 per cent, and annualised recurring revenues of 1.46 billion dollars, growing about 11.5 per cent in constant currency. Subscriptions now account for around 92 per cent of revenue, net revenue retention sits at 109 per cent, and free cash flow reached 520 million dollars.

Momentum has carried into 2026, with first-quarter revenue of about 424 million dollars and second-quarter annualised recurring revenue near 1.54 billion. These are not the numbers of a company chasing growth through discounting. They describe a mission-critical vendor with high retention and predictable cash generation, which is why the AI and data narrative matters so much to how the equity is valued.

Capital is following the operations phase of the asset lifecycle, where Bentley sees the next tranche of value. Its Asset Analytics portfolio, which applies digital twins and AI to help owner-operators track and improve asset performance, has been reinforced by the December 2025 acquisitions of Talon Aerolytics and the technology of Pointivo, extending the company’s reach into telecommunications and electric utilities at a moment when 5G rollout and grid modernisation are driving asset investment.

That portfolio already includes Blyncsy, which uses crowdsourced dashcam imagery and computer vision to detect roadway defects and inventory assets, and OpenTower iQ for telecom towers. A 2024 partnership with Google adds Street View imagery and Vertex AI to the roadway analytics stack, and the Seequent geosciences business is opening a further front in mining and critical resources. The pattern is deliberate. Bentley is buying and building its way into the recurring work of operating infrastructure, not just the episodic work of designing it, because operations is where budgets are large, continuous and comparatively insulated from project cycles.

There is a credible read of risk here that leaders should weigh. Analysts have trimmed some price targets on softer organic growth expectations, public-sector procurement can be slow and politically exposed, and digital-twin adoption at owner-operators may prove steadier than spectacular. None of that undermines the strategic logic, but it does temper the pace at which the operations opportunity converts to revenue.

For buyers, the more immediate consequence is that Bentley’s roadmap increasingly assumes customers will commit to its cloud and data layer, which raises the stakes on both the value delivered and the terms of that commitment.

The Competitive Battle For Horizontal Infrastructure

Bentley does not own this shift, and the competitive context is where the finalist celebration meets commercial reality. The company’s defensible position is horizontal infrastructure, the roads, rail, bridges, water and energy networks that connect the built environment, where MicroStation remains a standard in transportation departments worldwide and where products such as OpenRoads, OpenRail and OpenFlows carry deep domain credibility.

Autodesk, its broadest rival, is strongest in vertical building through Revit and Civil 3D and is pressing into infrastructure and cloud delivery. Trimble combines software with positioning hardware and field workflows, and its AECO software business alone was generating around 1.1 billion dollars of recurring revenue by 2024. Nemetschek holds significant European share with an open-BIM philosophy, while Hexagon and AVEVA compete hardest in reality capture and industrial asset operations.

Everyone in that set is now racing on AI, which makes differentiation harder and interoperability more valuable. Trimble has moved to acquire construction risk and compliance capability through Document Crunch, and Nemetschek has bought AI specialists including Firmus AI and field-digitisation tools such as GoCanvas. Against that backdrop, Bentley’s wager is that AI grounded in genuine infrastructure context, and anchored to lifecycle digital twins, will outperform more general construction intelligence for owners of complex, long-lived assets.

The company has paired that with an open-data message and an explicit commitment that customer data will not be used to train AI models without consent, a stance calculated to ease the trust barrier that often stalls enterprise adoption. For procurement teams, the practical takeaway is that the market is consolidating around a few platform ecosystems, and the decision that matters most is which vendor owns the authoritative model and connected data at each stage of the lifecycle, because that choice shapes cost, lock-in and optionality for years.

What The Finalist Field Signals For Owners And Delivery Teams

Read as a body of work rather than a list, the finalists map where digital methods are creating the most operational value. Resilience engineering is prominent, with PT Wijaya Karya’s Semarang-Demak toll road designed for a coastline under pressure and Arcadis automating the design of flood-protection structures through its AutoFW work. Subsurface intelligence is another cluster, reflecting the reality that ground risk drives cost and programme certainty on major projects. Mott MacDonald’s 3D geological modelling for Changi Airport’s Terminal 5 and Intertechne’s geotechnical modelling of a tailings dam both show teams using data to de-risk what lies beneath before committing capital above it.

Construction rehearsal is a third and commercially significant theme. AtkinsRéalis is advancing 4D construction at scale on Hinkley Point C, the Skanska Costain STRABAG joint venture is using digital methods to rehearse HS2’s London tunnels before crews break ground, and Wills Bros is applying 4D planning to marine works at Dublin Port. The common thread is that expensive, high-consequence sequences are being tested in software first, compressing risk and rework on site.

Elsewhere the energy transition dominates, from APD Global’s AI-enabled delivery of the Clean Energy Link-North transmission project to NV5’s mapping of risk across 17,600 miles of grid, alongside geothermal and pumped-storage schemes in Indonesia and China. The signal for owners and delivery teams is unambiguous. Organisations that can capture reality, model the subsurface, rehearse construction and carry that data into operations are pulling ahead, and those still handing over disconnected files are accumulating a competitive disadvantage that will show up in cost, schedule and asset performance.

The Direction Of Travel For Infrastructure Delivery

Taken together, the YII 2026 finalists describe a sector past the question of whether to digitise and into the harder question of how to operate as a connected, AI-assisted business. The escalation in AI use, the weight given to connected data foundations, and Bentley’s own pivot towards asset operations all point the same way.

Value is concentrating in the persistent data estate and in the productivity that AI unlocks against a genuine engineering-capacity shortage, and that concentration will increasingly separate firms that treat data as an operating asset from those that treat it as project overhead. The Singapore event in October will supply the case studies, but the strategic direction is already legible from the shortlist alone.

For infrastructure leaders, the practical response is less about admiring the finalists and more about auditing their own position against them. The pertinent questions concern how much of the estate sits in a connected, queryable form, where AI could relieve the tightest engineering bottlenecks, and which platform decisions will govern data ownership and cost across the coming decade.

The vendors will keep competing on features and acquisitions, and that competition is healthy for buyers. The organisations that benefit most will be those that decide, deliberately and soon, how they intend to own and exploit their infrastructure data before the market decides for them.

Key Industry Questions

1. Why does the proportion of AI-using finalists matter more than the technology itself? Because the year-on-year jump, from roughly half of finalists in 2025 to two-thirds in 2026, marks a change in expectation rather than capability. When AI appears on most of the projects that firms nominate as their best work, it has moved from a competitive edge to an assumed competence. For owners and contractors, that reframes investment decisions. AI-assisted delivery becomes part of the baseline case for winning and executing complex work, rather than an optional enhancement, which in turn strengthens the commercial rationale for the software subscriptions and data platforms that support it. The trend also gives boards a benchmark for judging whether their own digital maturity is keeping pace with the leading edge of the market.
2. What is a connected digital foundation, and why do asset owners care? A connected digital foundation is a shared, structured data environment in which design, construction and operational information stay linked rather than fragmenting into separate files at each handover. Owners care because it changes lifecycle economics. Instead of inheriting an as-built model that ages immediately, they gain a living digital twin they can interrogate to plan maintenance, test interventions and allocate capital against real asset performance. The Evides operational twins across a 14,000-kilometre water network illustrate the point, turning the model into an operating tool. The commercial payoff shows up over decades in lower whole-life cost, better resilience and more defensible investment decisions, which is why the connected foundation, not any single application, is the strategic prize.
3. How is Bentley turning these trends into revenue? Bentley has restructured its product line, acquisitions and pricing around connected data and AI. Its next-generation applications, including OpenSite+, the Bentley Copilot and AI-powered ProjectWise search, target design and delivery productivity, while its Asset Analytics portfolio targets the operations phase through tools such as Blyncsy and OpenTower iQ. The December 2025 acquisitions of Talon Aerolytics and Pointivo extended that reach into telecommunications and electric utilities. Financially, the model is subscription-heavy, with recurring revenue around 92 per cent of the total and net revenue retention of 109 per cent, meaning existing customers tend to spend more over time. The strategy concentrates on capturing continuous operational work rather than episodic design work, where budgets are larger and steadier.
4. Which competitors are best placed to challenge Bentley? Autodesk is the broadest rival, dominant in vertical building through Revit and Civil 3D and pushing harder into infrastructure and cloud delivery. Trimble competes strongly where field work, positioning hardware and construction execution matter, and its AECO software already generates substantial recurring revenue. Nemetschek holds significant European share with an open-BIM approach, and both it and Trimble are buying AI capability, through deals such as Firmus AI and Document Crunch respectively. Hexagon and AVEVA press hardest in reality capture and industrial asset operations. Bentley’s defensible advantage is horizontal infrastructure depth and lifecycle digital twins, but the competitive intensity around AI and interoperability means differentiation is narrowing and buyers hold more leverage than before.
5. What does the finalist field suggest about where digital value is concentrating? The projects cluster around a few high-value uses. Resilience engineering features in schemes such as the Semarang-Demak coastal toll road and Arcadis’s automated flood-protection design. Subsurface intelligence appears in Mott MacDonald’s geological modelling for Changi Terminal 5 and in tailings-dam geotechnics, reflecting how ground risk drives cost certainty. Construction rehearsal is prominent on Hinkley Point C, HS2’s London tunnels and Dublin Port, where costly sequences are tested in software before work begins. The energy transition runs through transmission, grid-risk mapping, geothermal and pumped-storage projects. Collectively, these point to value concentrating where teams can capture reality, model risk, rehearse construction and carry data into operations, rather than in isolated design tasks.
6. How should procurement teams respond to platform consolidation? The market is coalescing around a small number of platform ecosystems, so the decisions that carry the longest consequences concern data ownership and interoperability rather than individual features. Procurement teams should establish early who owns the authoritative model and connected data at design, construction, handover and operations, because that governs cost, lock-in and future optionality. It is worth scrutinising data portability, the terms under which information can move between tools, and any commitments a vendor makes on using customer data for AI training. Bentley has publicly committed not to train models on customer data without consent, and comparable assurances from rival vendors should be sought. Treating the platform choice as a decade-long data strategy, rather than a software purchase, protects value.
7. Is the engineering-capacity argument for AI credible or marketing? The capacity constraint is real and widely acknowledged across the sector, which lends the argument substance beyond vendor messaging. Infrastructure programmes face persistent shortages of experienced engineers at a time when investment in transport, water, energy and grid modernisation is rising. That imbalance makes automation of repetitive tasks, such as drawing annotation, documentation and option evaluation, valuable in its own right rather than merely aspirational. The important qualification is that current tools augment rather than replace skilled judgement, and their value depends on the quality of the underlying data. Used well, AI relieves specific bottlenecks and frees scarce engineers for higher-value work, which is a more grounded proposition than claims of wholesale transformation.
8. What should infrastructure leaders do with this information now? The most useful response is a candid audit of their own digital position against the leading edge the finalists represent. Leaders should assess how much of their asset estate exists in connected, queryable form, identify where AI could relieve the tightest engineering bottlenecks, and clarify which platform decisions will govern data ownership and cost over the coming decade. Pilots are worth running where they address genuine constraints rather than chasing novelty, and data governance deserves attention before scaling any AI deployment. The organisations likely to benefit most are those that decide deliberately how they intend to own and exploit their infrastructure data, rather than allowing vendor roadmaps and market momentum to make those choices for them.

Strategic Takeaways

1. AI on flagship infrastructure projects has shifted from differentiator to baseline expectation, and firms treating it as optional risk falling behind peers who now assume it in their delivery models.
2. The durable value is migrating from the one-off design deliverable to the persistent, connected data estate, making digital-twin continuity into operations the real competitive prize rather than any single application.
3. Bentley’s capital allocation signals where the sector’s next margin lies, in the continuous work of operating infrastructure through asset analytics, rather than the episodic work of designing it.
4. Platform consolidation means the highest-stakes procurement decision is data ownership and interoperability across the lifecycle, a choice that will shape cost, lock-in and optionality for a decade.
5. With engineering capacity as the binding constraint, productivity software that addresses a genuine labour shortage is becoming a necessity rather than a discretionary purchase, underpinning both vendor pricing power and buyer urgency.