Skip to content
‹ Back to IndustriesTransportation

How SFO’s digital twin helps carry construction data into operations

San Francisco International Airport rehearsed changes in its $11 billion capital program in a digital twin before building them in the live terminal, helping airlines keep operating. SFO teams also developed an automated workflow that brings mechanical, electrical and plumbing BIM models into the airport’s enterprise GIS and digital twin so design and construction data carries forward into operations.

This story was produced through MarketScale. See how Transportation teams put it to work with Partner & Channel Enablement.

By MarketScale Newsroom · · San Francisco International AirportSfoEsriDigital Twin
Share
How SFO’s digital twin helps carry construction data into operations

Key takeaways

01

SFO developed an automated workflow that brings BIM—including mechanical, electrical and plumbing models—into its enterprise GIS and digital twin so design and construction data can carry forward into operations and asset management.

02

Shared visibility did much of the work during SFO's $11 billion program: construction, operations, security and airline partners read one location-based model, so conflicts showed up as design questions instead of shutdowns.

03

Digital twin programs can now refer to the BSI Kitemark for Digital Twins as a certification scheme that grades governance, security, ownership, and measurable value.

Free workspace

Turn your Transportation expertise into content.

Record interviews, organize footage, and write with AI on a free trial of the MarketScale platform for qualifying companies. No demo required, no credit card.

Try it Free

Every phase of San Francisco International Airport's $11 billion capital improvement program touched something passengers rely on: a security checkpoint, a utility line, a baggage system, the route to the gate. SFO set out to do all of it while staying fully operational. Its method was simple to describe. Teams tried each change in a digital twin of the airport first, then built it for real.

According to a Forbes article on the project, SFO teams recently developed an automated workflow that integrates building information models into the airport's enterprise GIS and digital twin. The workflow covers mechanical, electrical, plumbing and other building systems as well as architecture. For anyone running a large campus, that step could be what keeps a capital program's data useful after the project closes.

Rehearsing construction in a live terminal

The airport's size explains why it rehearses changes first. Kewley calls it a hard version of modernization: no downtime, no spare capacity, and millions of people passing through every month. When one phase moves a checkpoint, the change can also hit utilities, baggage handling and passenger flow, and any of those can set off a chain of disruptions.

The airport the twin has to mirror

54 million
passengers a year moving through while construction continued
18 million+ sq ft
of indoor space modeled, from rooftop solar to underground utilities
100+ buildings
spread over eight square miles, all tied to exact locations
$11 billion
in capital improvements tested in the twin before construction

Forbes (Kathleen Kewley, Esri)

In Kewley's account, construction, operations, security and airline partners all worked from the same view. Issues that might have forced emergency shutdowns came up early, while they were still design questions. Airlines kept operating and passengers kept moving during the program.

This suggests most of the value came from shared visibility rather than raw simulation power. Moving a checkpoint is a security question, a utilities question and an airline question all at once. When all four groups read the same model, a conflict appears on screen before it appears in a concourse. Operators planning phased work inside a facility that never closes, such as a hospital, a distribution hub or a transit station, would face the same coordination problem.

Daily use is less dramatic. Josephine Young, SFO's director of infrastructure information management, calls the twin a "window of entry" for business units across the airport. Users can look inside a building layer by layer. Managers use it to plan a terminal remodel or a new link between baggage claim and public transit. It also handles preventive maintenance, down to when and where to change lightbulbs or upgrade pipes.

Why every data point carries an address

Kewley argues that without that geographic anchor, the model couldn't show how a passenger bottleneck in one area backs up a security checkpoint elsewhere, or how slow cargo unloading affects logistics inside the airport and beyond it.

Esri's architecture team made a related point the same day in an ArcGIS Blog post by Rob Elkins, Julia Colbert, Jay Fowler and Dylan Molnar. They cite the Digital Twin Consortium definition, which turns on a model kept in sync with the real world at a specified frequency and fidelity. They add two conditions people often skip: the twin needs a defined outcome, and people have to adopt it in real workflows. Their transportation example lists the feeds a twin like this can use, including vehicle location, vessel and aircraft tracking, traffic and weather. They also advise choosing data by the decisions the system has to support rather than by how many feeds exist.

That filter is practical for a facilities director scoping a twin. SFO's stated uses are concrete: a remodel, a transit connection, a lightbulb schedule. Each needs only a short list of data layers. Start with the decisions, and the data list comes out much shorter than a list of every available sensor.

Handover data that keeps working

The newest piece of SFO's twin is an automated workflow that brings BIM models into the airport's enterprise GIS. It covers mechanical, electrical, plumbing and other building systems as well as architecture. Kewley says the immediate gain is a fuller, continuously updated view of infrastructure. The larger one is that design and construction data carries forward into operations and asset management.

SFO teams recently developed an automated workflow that integrates BIM models into the airport's enterprise GIS and digital twin. It covers mechanical, electrical, plumbing and other building systems as well as architecture, so data from design and construction keeps paying off after the project ends.

For owners running long capital programs, that changes where the twin gets built. If MEP models feed the operating system directly, the twin is only as good as what contractors hand over. Owners would care about model completeness, consistent asset identifiers and accurate locations. MarketScale's read, not a published SFO requirement, is that these expectations belong in design and construction contracts written now, years before the equipment shows up on a maintenance schedule.

AVEVA reports that engineering firm AP Consultoria e Projetos used a cloud-based engineering environment where multidisciplinary teams worked from a single source of truth. The firm cut analysis time for pipe support type and position by 90% and reduced total project delivery time by nearly 50%. The post was first published by the Financial Times Tech for Growth Forum. The mechanism looks similar to SFO's: when every discipline works from one model, problems could surface earlier.

Where the word twin stretches

Not everyone uses "digital twin" the same way. In an Oct. 21 guide, Tributech CEO Thomas Plank cites a 2018 classification by Kritzinger and colleagues that separates three levels by how data flows between the asset and the model. Under that test, a model that receives data from the asset automatically is a digital shadow, and only one that also sends data back automatically counts as a digital twin. Operators could weigh their own planning uses against that line.

In the Kritzinger et al. (2018) classification, a digital model is updated by hand in both directions. A digital shadow receives data from the asset automatically, but changes go back by hand or not at all. A digital twin exchanges data automatically in both directions, so commands from the model can reach the physical asset.

That isn't a knock on SFO's results. It changes what a buyer should specify. For coordinating construction and planning maintenance, two-way control may not be needed, and it carries extra security weight: Plank notes that once a twin can write to an asset, it becomes part of the control path, and ISA/IEC 62443 security rules apply. For many facilities programs, a read-only twin is the safer and cheaper choice.

Two other problems Plank lists apply to any large owner. Models drift from reality as assets change, and ownership is often split among OT, IT and engineering. SFO's automated workflow that brings BIM models into its enterprise GIS and digital twin suggests one way owners could keep a model from drifting.

A yardstick for twin programs

Buyers now have an outside reference point. BSI announced the BSI Kitemark for Digital Twins in an Oct. 7 press release and called it a new certification scheme. Its purpose is to let organizations show that the digital twin capability they run is secure and well governed, and that it delivers measurable business outcomes. Further requirements are added on top for governance, competency, security, risk management and value realization. Assessment covers a specific twin as well as how well the organization can deploy, govern and improve twins over time. The first certification went to Neuron at a ceremony in Hong Kong.

Certification won't suit every program. Owners who are early in the process can still use the maturity model as a checklist. Together with SFO's example, it points to a few questions to settle before scoping a twin or writing it into a capital program:

  1. Which operating decisions will the twin support, and which data layers does each one actually need?
  2. How will design and MEP models enter the GIS at handover, and who keeps them current afterward?
  3. How often does the model sync, and does it only read from assets or also send commands back?
  4. Which team owns the model when OT, IT and engineering each hold part of it?

The next step at SFO is easier to measure. Kewley writes that as automation and AI improve, the twin should get better at flagging discrepancies and spotting patterns behind the scenes. If it starts surfacing problems that no one asked it to look for, it will be doing more than mapping the airport.

Featured companies

Your experts belong here

Every story in MarketScale Transportation starts with a company putting its fleet managers, logistics engineers, and safety leads on the record. Buyers are already reading this topic. The only question is whose experts they find.

Fleet and logistics buyers compare quietly, and your operators become the evidence that settles it.

Book DemoSee how it works15 minutes, straight to a calendar.

About the author

MarketScale Newsroom
MarketScale NewsroomEditorial Team, MarketScale

The MarketScale Newsroom reports on the companies, technologies, and trends shaping 16 B2B industries. It turns primary sources and expert commentary into clear, useful coverage for the people doing the work.

B2B Weekly

The week in Transportation, and sixteen other industries, every Monday.

Ten stories, one-line takes, five minutes. Free.

Transportation: are you visible to AI?

Before they reach out, Transportation buyers ask AI engines which vendors to trust. Explore how your experts, customers, and partners can become useful content for buyers and AI search.

Free Trial

You just read one Transportation expert. Your company is full of them.

This article was produced through MarketScale. The same platform turns your fleet managers, logistics engineers, and safety leads into the articles, video, and social content Transportation buyers are searching for. Start a free trial and see it with your own people. For qualifying companies, no credit card, no demo required.

NPS +73 · 1,000+ creators · 38+ countries

What your free trial includes

Hands-on access to the MarketScale platform
Media requests to your crowd, remote recording, AI writing tools
No demo required. No credit card.
For qualifying companies. Company confirmation required.

More Transportation Insights

ZYT says Prax can pay back in a year on fuel savings

ZYT says Prax can pay back in a year on fuel savings

Zhuoyu Technology (ZYT) launched Prax, a driver-assist system for heavy trucks, with Shacman's Delong X5000 already running Chinese parcel routes. ZYT claims 4.5% to 7.5% fuel savings. That overlaps rival Inceptio's fleet results, so driver usage rates and deliveries across seven truck makers are the better test of its value to fleets.

  • 01When two driver-assist vendors claim fuel savings in the same band, the share of highway kilometres drivers actually run with the system on becomes the number to compare.
  • 02ZYT's one-year payback assumes a truck covering 800 km a day, so fleets on shorter or slower routes should rerun the math with their own mileage and fuel prices.
  • 03Hundreds of Shacman Delong X5000 trucks are already running Prax, while other partner programs are at earlier stages; Gasgoo says 2027 will test ZYT as more models move toward delivery.

Nov 16, 2026

Why IndexBox sees EV wires growing twice as fast as wire harnesses

Why IndexBox sees EV wires growing twice as fast as wire harnesses

IndexBox forecasts EV wires growing 12.4% a year through 2035, more than double the 5.8% for wire harnesses overall. 800V platforms drive the gap. Automaker supplier approval takes years, so today's sourcing decisions set much of the 2030s supply base.

  • 01EV wire demand is forecast to roughly triple by 2035, and the suppliers who fill those orders are likely the ones already inside multi-year OEM approval cycles.
  • 02Wire value per vehicle is rising faster than vehicle counts: EVs carry 20-30% more harness content than combustion cars, and 800V platforms need thicker cable and better insulation.
  • 03Replacement harnesses and charger cables make up about a fifth of EV wire demand and are far less program-locked, so certification and distribution networks matter more there than OEM approval.

Nov 16, 2026

FAZT pegs EV driving at 5.4 US cents a kilometer at its new Guadalajara station

FAZT pegs EV driving at 5.4 US cents a kilometer at its new Guadalajara station

FAZT opened a US$500,000 fast-charging station in Guadalajara with 12 stalls of 120 kW each and a flat US$0.45 per kWh rate. It puts EV driving at 5.4 US cents per kilometer, against 8.4 for gasoline. Heavier fleet vehicles need their own math.

  • 01At US$0.45 per kWh, public fast charging beats a 15 km/L gasoline car only while the EV uses less than about 19 kWh per 100 km, so vehicle efficiency decides the savings.
  • 02A single published US$0.45/kWh rate at every FAZT station gives fleet budgets one public-charging price to plan against.
  • 03Charging density in Mexico City, Monterrey and Guadalajara does not yet extend to secondary highways and industrial parks, which matters most for fleets whose routes leave the metros.

Nov 15, 2026

Explore More Transportation Insights

Read more expert perspectives from across Transportation.

Browse Transportation Hub

About the Expert

MarketScale Newsroom
MarketScale Newsroom

Editorial Team

MarketScale

The MarketScale Newsroom reports on the companies, technologies, and trends shaping 16 B2B industries. It turns primary sources and expert commentary into clear, useful coverage for the people doing the work.

For B2B teams

Your experts could be publishing here

Stories like this one run on content MarketScale captures from real practitioners. See how your team's expertise becomes coverage in Transportation and beyond.

Book a Demo

Or call us. No forms required. We pick up. 214-945-2512