Dubai does not build one building at a time the way most cities do. Mega-projects here routinely begin with a brief, move through concept design, and expect a client-ready set of options within weeks, not quarters. In February 2026, Aldar and Dubai Holding expanded their joint venture with AED 38 billion in new developments, anchored by Palm Jebel Ali and a new growth corridor along the city’s eastern edge — a single site there covering four million square metres. Around the same period, BIG–Bjarke Ingels Group released a five-square-kilometre masterplan around the historic Jebel Ali Racecourse, organised into eight walkable neighbourhoods with a public park, equestrian venues, mixed-use blocks, housing, and cultural buildings. These are not isolated towers. They are small cities being designed in a matter of months, for a market that does not wait.
That pace creates a real design problem that traditional massing workflows were never built to solve. A single building’s massing study is already iterative — a few options, some daylight checks, a client review. Multiply that by eight neighbourhoods, dozens of plots, and a compressed design schedule, and the traditional one-option-at-a-time approach starts to break down. This is where generative massing tools, and specifically Autodesk Forma, earn their place in a masterplan workflow rather than being a novelty for a single building.
“Generative design” gets used loosely, so it is worth being precise about what Forma actually does at the masterplan level. Forma’s Site Design tools let a team build a 3D context model of a large site — terrain, surrounding buildings, infrastructure — and then test building massing options against real environmental data rather than guesswork. Autodesk has also introduced Building Layout Explorer, which generates and compares apartment-layout options against a target unit mix, and a Site Design workflow where an embedded assistant can answer natural-language questions about the project site itself. According to Autodesk’s own presentations at Autodesk University 2025, the generative AI layer is meant to automatically produce design alternatives based on environmental analyses such as daylight, wind, noise, and solar radiation, rather than requiring a designer to manually test each variant.
At the scale of a single villa, this is a convenience. At the scale of an eight-neighbourhood masterplan with hundreds of individual plots, it is closer to a necessity. One architecture firm, Haenglim Architects, is cited by Autodesk as having used Forma’s generative AI to produce multiple initial design options, which reportedly cut down review time significantly — though that figure comes from a vendor case study and should be treated as directional rather than an independently audited statistic.
Consider what a traditional massing study actually involves for one plot: block out a few volume options, run a basic sun study, maybe a rough wind assessment if the project justifies the consultant fee, present two or three options to the client, iterate once or twice, move on. That process, done well, might take a team a week or two per plot.
Now scale that to a masterplan with 40, 80, or 150 plots across multiple neighbourhoods, each with its own orientation, adjacent-building shadow conditions, and microclimate exposure — which is exactly the condition a site like the Jebel Ali Racecourse masterplan or the eastern growth corridor site creates. Running that process manually, plot by plot, either consumes enormous design-team hours or — more realistically — gets compressed into guesswork that skips the environmental analysis most plots would benefit from. Generative massing tools do not eliminate the design judgment involved; they change what a design team can realistically test within a real project schedule.
Dubai’s climate makes this more than an efficiency argument. Solar exposure, wind corridors, and outdoor comfort are genuinely different problems at district scale than at single-building scale, because buildings shade and channel wind for each other. A tower placed in the wrong position within a neighbourhood does not just affect its own performance — it can create a wind tunnel at street level for the plot next to it, or block solar access that a neighbouring building’s passive design was counting on. This is precisely the kind of interaction that is extremely difficult to catch with isolated, building-by-building massing studies, and exactly what a 3D context model with integrated environmental analysis is designed to surface early, before the relationships between buildings are locked in by detailed design.
| Factor | Traditional Single-Option Massing | Generative Multi-Option Massing (Forma-style) |
|---|---|---|
| Number of massing options tested per plot | Typically 2–3, limited by team hours | Many, generated and screened faster against set criteria |
| Cross-plot shading/wind interaction | Often checked late or informally | Modelled as part of the same context, earlier |
| Environmental data integration | Separate consultant study, often sequential | Embedded in the same massing workflow |
| Client comparison speed | Slower, fewer live alternatives | Faster side-by-side comparison of generated options |
| Where it hands off | Straight to detailed design per building | Still hands off to Revit for detailed design, informed by site-wide analysis |
It would be dishonest to present this as a solved problem. User discussion on Autodesk’s own community forums has noted that Forma’s generative features are, as of recent releases, more limited than what the earlier Spacemaker tool (which Forma grew out of after Autodesk’s 2020 acquisition) offered for some generative workflows. Autodesk staff have responded that the team is currently focused on “Line Building” functionality, described as the planned foundation for broader generative features going forward — which is a polite way of saying the full generative vision is still being built out, not fully delivered today. Separately, academic commentary on AI-assisted architectural design has flagged open questions about whether deep-learning-based generative tools are genuinely ready for everyday production workflows, as opposed to compelling demos.
The honest summary: Forma’s environmental analysis and context-modelling capabilities at site and district scale are real and useful today. Its fully automated generative design capability — the part that gets the most marketing attention — is still maturing. A masterplan design team should plan to use Forma’s analysis strength heavily, and treat its generative automation as an accelerant on top of design judgment, not a replacement for it.
In practice, the tools slot into a sequence rather than replacing each other. Forma (or a comparable site-design tool) is used at the earliest stage to build the context model, test massing and environmental performance across the whole site or district, and narrow down viable options. Once a direction is set for a given plot or building, the detailed design moves into Revit, where the architectural, structural, and MEP teams develop the building properly. The generative massing stage is not a replacement for that detailed design work — it is what determines, with far more confidence than a few manual studies could provide, which massing options are even worth taking into detailed design in the first place.
For a masterplan project specifically, this sequencing matters because the cost of a wrong decision compounds. A massing error caught at the conceptual stage costs a redesign of a 3D model. The same error caught after several plots have gone into detailed Revit design — with structural grids, MEP routing, and facade systems already committed — costs far more in rework, and on a tight masterplan schedule, there often isn’t time to absorb that cost gracefully.
For architects and urban designers working in Dubai’s development market, familiarity with generative massing and environmental-analysis tools is becoming a genuinely marketable skill, not a niche specialty. Large developers and masterplanning consultancies working on district-scale projects are the ones most likely to need this capability, precisely because the scale problem described above is their everyday reality. Being the person on a team who can actually run and interpret a Forma-based site analysis, rather than waiting for a specialist consultant, is a real differentiator on a resume aimed at masterplan-heavy firms.
If you want to build this skill properly rather than piecing it together from scattered tutorials, Orbit Training Centre runs a practical, hands-on Autodesk Forma course in Dubai that covers site context modelling, environmental analysis, and massing workflows aimed at exactly this kind of real project use case — not just the software interface, but how it fits into an actual Dubai-scale design process.
It’s worth walking through a concrete example rather than staying abstract. BIG-Bjarke Ingels Group’s five-square-kilometre masterplan around the historic Jebel Ali Racecourse is organised into eight distinct walkable neighbourhoods, each with its own character, plus a public park, equestrian venues, mixed-use blocks, and cultural buildings woven between them. A project with that structure creates a specific kind of design problem: every neighbourhood needs its own coherent massing logic, but the neighbourhoods also need to work together as a single legible district — shared sightlines, consistent building heights at transition zones, pedestrian routes that actually connect rather than dead-ending at a plot boundary.
Running that kind of cross-neighbourhood coherence check manually, plot by plot, with each neighbourhood’s design team working somewhat independently, is exactly the scenario where things go wrong on a real masterplan — two adjacent neighbourhoods each individually “pass” their own massing review, but the seam between them doesn’t actually work: heights jump awkwardly, a sightline corridor one team was counting on gets blocked by a building the neighbouring team placed without seeing the full context. A shared 3D context model, of the kind Forma is built to maintain, is what lets a masterplan-level design lead actually catch that seam problem before it becomes a construction-stage conflict, because every neighbourhood team is testing its massing against the same live site model rather than against its own isolated file.
There is also a straightforward economic argument for consultancies bidding on this kind of work, separate from the design-quality argument above. Design fees on large masterplan commissions are typically fixed or capped relative to project value, which means a design team’s profitability on the commission depends heavily on how efficiently they can produce defensible, client-ready options within that fee. A workflow that can generate and screen many massing options against real environmental and zoning constraints in the time it used to take to manually produce two or three options doesn’t just produce a better design outcome — it changes the economics of taking on masterplan-scale work at all, because the alternative (throwing proportionally more staff hours at manual iteration to hit the same option count) erodes margin on exactly the kind of large commission a firm most wants to protect.
One practical point that often gets missed: generative massing tools deliver the most value when an entire project team understands the same context model and can interpret its outputs consistently, not when one specialist on the team runs it in isolation and hands off static images to everyone else. A masterplan team where only one person can actually operate the site-analysis tool creates a bottleneck exactly at the stage where speed matters most, and it also means design decisions get filtered through one person’s interpretation rather than being something the wider team can interrogate directly. Firms taking on masterplan-scale work in Dubai are increasingly training multiple team members on this workflow for precisely that reason — it’s a team capability, not an individual specialist’s party trick.
The difference shows up clearly at the client review stage, which is ultimately where a masterplan design gets judged. With a traditional manual workflow, a design team typically arrives at a review with two or three fully worked options, having already made the hard trade-off decisions internally before the client ever sees them — which means if the client’s actual priorities don’t match what the team assumed, there’s limited room to pivot without a costly new round of manual work. With a generative, context-model-based workflow, a team can arrive with a wider set of genuinely viable options already screened against real environmental and zoning data, and can meaningfully adjust in the room — testing a client’s specific preference against the live model rather than promising to “come back with that option next time.” For a market where client decision cycles move quickly and design fee schedules are tight, that responsiveness is itself a competitive advantage independent of the final design quality.
None of this should be read as a claim that generative massing tools compress a masterplan’s overall project timeline dramatically on their own — authority approvals, infrastructure coordination, utility planning, and construction sequencing remain the dominant factors in how long a masterplan actually takes to deliver, and no software changes that. The realistic claim is narrower and still valuable: these tools compress the specific concept-and-massing phase, and they improve the quality of the decisions made within that phase by testing more options against real constraints. That is a meaningful advantage for a design team, but it is one piece of a much longer delivery timeline, not a shortcut around the rest of it.
No — it’s useful for single buildings too, especially for early sun, wind, and massing studies. But the case for using it becomes much stronger at masterplan scale, because manually running equivalent studies across dozens of plots is simply not practical within a real project schedule.
Not fully. Forma gives a design team fast, useful early-stage environmental feedback to guide massing decisions. Detailed compliance-grade environmental analysis for a large project, especially one going through formal authority approval, typically still involves specialist consultants running more rigorous studies later in the process.
Once massing and site-level decisions are set in Forma, individual buildings move into Revit for detailed architectural, structural, and MEP design. Forma is not meant to replace that detailed design stage — it’s meant to make sure the decisions going into it are better informed.
Not entirely, based on current user and vendor commentary. Forma’s environmental analysis capability is strong today; its fully generative, automated design-option capability is still being developed, with Autodesk currently focused on expanding its “Line Building” functionality as a foundation for that.
It helps, since you’ll eventually hand your Forma massing decisions into Revit for detailed design, but it’s not a strict prerequisite to start learning Forma’s site-analysis and massing workflow itself.
Information in this article draws on Autodesk’s own Forma product communications and Autodesk University 2025 session materials, Autodesk community forum discussion on Forma’s generative features, and 2026 reporting from Gulf News, Khaleej Times, and development-industry press on the Dubai 2040 Urban Master Plan, the Aldar–Dubai Holding joint venture expansion, and the Jebel Ali Racecourse masterplan.
