PropTech means property technology: the software, sensors and data tools that handle buying, financing, leasing and running buildings. It used to be a label for startups. In 2026 it is how most of that work gets done.
The change is easiest to see in one number. In February 2026, 26% of loans bought by Fannie Mae and Freddie Mac closed with no traditional appraisal, replaced by an automated valuation, according to the American Enterprise Institute Housing Center’s tracking of appraisal waivers. On no-cash-out refinances the waiver share was close to half. A step that once meant a person walking the property with a clipboard is now, very often, a model.
This article covers what actually works in real estate technology right now, what quietly stopped working, and what the new rules require of you. Every figure comes with its source, and the last section is a 12-month plan you can start this quarter.
Key Takeaways
- Automated valuation has moved from pilot to default on a large share of US mortgage loans.
- Tokenized assets grew fast in 2026, but property is a small slice of a market led by credit and Treasurys.
- Smart building systems earn their keep on energy and maintenance, not on novelty features.
- Algorithmic pricing and AI tenant screening now carry real legal exposure, with settlements to prove it.
- Proptech funding has stabilised at a much lower level, concentrated in a handful of large rounds.
- Virtual land was the clearest dead end of the last cycle. Treat new frontier claims accordingly.
Why location stopped being the whole story
Address and square footage still set the floor on what a property is worth. What changed is everything layered on top: how fast someone can see the space, how easily they get in, and what the building costs to run once they are in it.
Tenants now expect the same self-service they get from everything else they buy. A contactless tour at 9pm on a Sunday. A move-in handled in an app. A maintenance request that produces a ticket number rather than a voicemail. None of that is exotic technology, and that is the point: it is cheap enough that not offering it reads as neglect.
The commercial market makes this urgent. US office vacancy reached a record 21% in the first quarter of 2026, and 10 of the preceding 13 quarters showed net contraction, according to Moody’s Analytics. When a fifth of the space is empty, a landlord competes on service and running costs. Our guide to the future of the office covers that shift, and office design trends looks at what occupiers now ask for.
Two rules keep this from becoming a shopping list: measure app usage, service tickets and renewal rates rather than feature counts, and retire any platform that has not moved one of those three.
Where the money is actually going in 2026
The funding picture tells you which parts of PropTech investors still believe in. Global proptech venture funding reached $4.53 billion across 231 rounds in the first half of 2026, according to the Center for Real Estate Technology and Innovation (CRETI), almost flat against the same period in 2025 at 0.6% lower.
Flat is the story. The boom years are over and the market has not broadened to replace them. Eleven rounds of $100 million or more took 49.6% of all capital, while 75 rounds of $5 million or less took 2.8%. The money goes to a small number of established companies working on housing supply, construction, energy, building operations and real estate capital markets.
Consolidation follows the same logic. Rocket Companies bought Redfin, then closed a $14.2 billion acquisition of mortgage servicer Mr. Cooper in October 2025, putting search, lending and servicing under one roof. CoStar completed its purchase of 3D scanning company Matterport in February 2025.
What this means for you: fewer independent vendors, more suites. Ask any vendor who owns them, and what happens to your data and pricing if they are acquired.
AI valuation and market intelligence
An automated valuation model, or AVM, estimates what a property is worth from data rather than an in-person inspection. It reads public records, past transactions, listing activity and local market signals, then returns a number in minutes.
AVMs are not new. What changed is that lenders and regulators now accept them for real decisions at scale. Fannie Mae and Freddie Mac also offer a middle option: a waiver combined with property data collected by a trained inspector rather than a licensed appraiser. In February 2026 that covered roughly 3% of purchase loans at each enterprise.
What modern models add, and where they break
Newer models pull in signals an appraiser would not systematically collect: transit plans, permit filings, local employment data, listing-to-sale ratios. On standard suburban housing stock with plenty of recent sales, they are genuinely good.
They are weakest where you would expect: unusual properties, thin markets, and anything recently renovated in a way no record captures. A model cannot see that the kitchen was replaced last month. Treat an AVM as a fast first read that flags which properties need a human look.
Risk, income, and ESG scoring for investment analysis
Investment platforms now layer risk and sustainability scores onto cash-flow models. ESG stands for environmental, social and governance: for a building, it mostly means energy use, emissions and climate exposure. A flood-zone score or an energy-intensity score changes the hold period and the capital expenditure plan, so it belongs in the underwriting rather than in a separate report. Our ESG framework guide explains how to structure those metrics.
Personalised leasing, and the screening rules you cannot ignore
AI can segment prospects, time follow-ups and handle renewals, which frees leasing staff for conversations that need a person. The risk sits in automated screening.
In November 2024, SafeRent Solutions agreed a $2.28 million settlement in Louis v. SafeRent Solutions, a class action alleging its tenant scoring algorithm disadvantaged Black and Hispanic applicants and housing voucher holders, and agreed to change how it scores voucher holders. If you buy a third-party screening score, you are letting someone else’s model make a fair housing decision for you. Ask for the fairness testing, keep records of adverse decisions, and keep a human review path open. Our piece on bias in algorithmic screening covers the testing methods.
Algorithmic pricing after the RealPage settlement
Revenue management software that sets rents is now a regulated activity in practice, even where no new statute names it. On 24 November 2025, RealPage reached a proposed settlement with the US Department of Justice in an algorithmic price-fixing case. Under its terms the company agreed to stop using competitors’ confidential data in real-time pricing, to stop training pricing models on active lease data, to drop features that discouraged price cuts, and to accept court-appointed monitoring. Greystar, the largest US landlord, settled separately.
The line the DOJ drew is about inputs, not algorithms. A pricing tool fed by your own data and public market data is ordinary business. One fed by rivals’ confidential, current numbers looks like coordination. Ask every pricing vendor in writing which data sources train the model. Our guide to dynamic pricing strategies sets out the safer patterns.
Blockchain and tokenized real estate
Tokenization means recording ownership as digital tokens on a blockchain, so a single property can be split among many holders and the fractions traded. The promise is liquidity for an asset class that normally takes months to sell.
The market grew quickly. Tokenized real-world assets reached $26.4 billion of on-chain value on 8 March 2026, close to four times the level a year earlier, according to tracker RWA.xyz (the figure excludes stablecoins). Six categories passed $1 billion: private credit, commodities, US Treasurys, corporate bonds, non-US government debt and institutional funds.
Read that list again. Real estate is not on it. Tokenization found its first real market in instruments that were already standardised and liquid, not in buildings. For property in 2026, the rails work and the volume is thin.
What a pilot should look like
If you want to test it, keep it contained. One property, one share class, a platform that handles identity checks and investor reporting, and a clear answer on where tokens can be resold. Secondary market depth is what disappoints people, so check it before you raise rather than after. Our coverage of blockchain in business gives the wider context.
Smart contracts for title, escrow, and payments
A smart contract is code that releases money or transfers a record automatically when agreed conditions are met. For rent distribution among token holders, or staged escrow releases, it removes reconciliation work.
It does not remove the legal layer. Whether a coded transfer satisfies your jurisdiction’s requirements for conveying title is a local question, and in most places the deed recording system still governs. Validate the template with counsel, write in a dispute resolution route, and keep conventional oversight on anything with a complicated title history. Our guide to smart contracts in business covers the drafting points.
Smart buildings and IoT: energy, maintenance, and tenant experience
IoT means internet of things: equipment with sensors and a network connection, so it reports its own state. In a building that means meters, thermostats, air quality monitors, door controllers and lift telemetry feeding one building management system (BMS).
The value here is unglamorous and reliable. Heating, cooling, ventilation and lighting dominate what a commercial building spends on energy, so controls that match them to actual occupancy cut a bill you already pay.
Integrated control, comfort, and cost
The win comes from connecting systems that used to run blind to each other. Occupancy data from access control tells the heating system which floors are in use, so you stop conditioning empty space at full load. Tenants notice the comfort; finance notices the bill. Honeywell, Johnson Controls, Siemens and Schneider Electric all sell here, and most large portfolios end up with a mix.
Predictive maintenance that lowers downtime
Predictive maintenance uses sensor readings to spot equipment degrading before it fails, so you service a chiller on a Tuesday morning rather than at 11pm on a Friday. Fewer emergency call-outs, longer equipment life, fewer angry tenants. Start with the assets whose failure is most disruptive: lifts, chillers, boilers. Our predictive maintenance guide covers how to set up the data side, and IoT in business explains the sensor layer.
ESG dashboards and certification readiness
Energy meters feeding a carbon accounting tool turn an annual scramble into a standing report. That matters for LEED and WELL submissions, for lender and investor questionnaires, and increasingly for leases with large corporate tenants who carry their own reporting duties. See our guide to carbon accounting software for how those numbers get assembled.
Three upgrades are deployable inside a quarter: smart locks with self-guided access, lighting controls with connected thermostats, and one work-order system so vendor jobs stop living in email. Check network reliability and device security first: a sensor on a flat network is a route into your building systems, and cheap hardware is the usual weak point.
VR, AR, and digital twins
A digital twin is a live 3D model of a real building, kept current with data from the building itself. Virtual reality (VR) puts a viewer inside that model through a headset or browser. Augmented reality (AR) overlays model data onto what someone on site sees through a tablet.
Immersive tours and self-guided leasing
3D tours are table stakes now rather than a differentiator, which is why CoStar bought Matterport. A prospect in another city can walk a floorplate at any hour, which widens your funnel and filters out visits that were never going to convert. Connect tour analytics to your CRM so follow-ups reference what the viewer lingered on. Our CRM trends guide covers that integration.
Construction and facilities
On a development site, AR overlays let a team spot clashes between mechanical, electrical and plumbing runs before they are built, which is where change orders come from. After handover, an as-built model gives operations clean data for work orders and asset registers instead of a box of drawings.
Digital twins then support monitoring, maintenance forecasting and training on equipment it would be unsafe to practise on. Pilot one building, prove the saving, use it as a template. Our guide to digital twin technology goes deeper on the tooling.
Platforms redefining brokerage and property management
The real change in transactions is that the whole sequence now leaves one audit trail. Listing, offer, approval, notarisation and signature used to live in five systems and an inbox. On a modern platform they are timestamped steps in one record, which shortens the deal and turns compliance reporting into a query rather than a project.
End-to-end digital transactions
Compass and Opendoor run versions of this on the residential side with different models: Compass as a brokerage platform, Opendoor buying homes directly and reselling them. As of its August 2026 guidance Opendoor was still targeting profitability by the end of 2026, which tells you how hard the direct-purchase model is to run at scale.
Property management as a service
AppFolio, Buildium and similar platforms automate rent collection, applicant screening and maintenance coordination. In most portfolios the honest benefit is not fewer staff. It is that the same team handles more doors and spends its time on retention rather than chasing payments.
Compare licence cost against the labour you actually remove, not the labour the vendor’s slide removes. A platform that does not talk to your accounting system creates a second data entry job, and standardising field names early keeps owner statements reconciling across a portfolio.
Fintech-enabled financing and investment access
Fintech has changed the speed and the evidence base of lending decisions more than the decisions themselves. Direct bank-data income verification replaced weeks of document collection, which matters most for self-employed and contract earners whose income is real but awkward to document.
Mortgage tech and AI risk assessment
Models can read irregular income patterns that a rules-based underwriter would reject outright. That widens access where it is paired with human review and clear records of why an application was declined. Lending sits squarely inside fair lending law, so the documentation is not optional. Our overview of fintech trends covers the wider market, and insurtech trends the parallel shift in property insurance.
Crowdfunding, REITs, and tokenized rails
Platforms such as Fundrise and Arrived let retail investors buy fractions of property portfolios with small minimums. A REIT, or real estate investment trust, does something similar through public markets, with daily liquidity and mature disclosure rules.
Compare four things before you list on a platform or invest through one: the minimum, the fee stack, how and when you can exit, and the reporting quality. Exit terms are where the surprises live. Our look at decentralised finance in business transactions covers the on-chain versions.
Data-driven city planning and zoning
City-scale modelling lets you test how density, transport and climate scenarios affect a site before you commit to a design. Scenario tools combine transport networks, flood and heat data, and demographic projections.
Be realistic about what this buys. The most famous smart city project of the last decade, Alphabet’s Sidewalk Labs development in Toronto, was cancelled in 2020, and the unit was folded into Google at the end of 2021. The technology was never the hard part. Consent was.
Using planning data to reduce delays
What works is narrower and genuinely useful: feeding model outputs into your design and pro forma early, so parking ratios, resilience measures and density targets are set before drawings harden. The same datasets let you make a data-backed case to planners and neighbours, which is where most of the calendar risk sits. Our guide to urban tech trends covers the tooling.
“Early visibility into constraints turns unknowns into negotiable tradeoffs.”
Metaverse and virtual land: what actually happened
This is the clearest cautionary tale in recent real estate technology. In 2021 and 2022, brands and speculators bought parcels in Decentraland and The Sandbox on the argument that scarce virtual land would behave like scarce physical land.
It did not. Average floor prices for land in The Sandbox fell roughly 95% from their 2021 peak, and Decentraland land fell roughly 89% from its 2022 peak, measured to June 2024, according to CoinGecko research. The scarcity was real and the demand was not.
The lesson generalises. When a new asset class is sold on an analogy to property, ask who buys it once the speculators leave. If the answer is unclear, cap your exposure at an amount you would write off without flinching.
Regulation, privacy, and cybersecurity
Every system above collects personal data about tenants, buyers or staff, which makes data practice a property management duty rather than an IT one. A breach touches valuation, insurance cover and tenant trust at once.
Data governance, identity, and zero trust
Start with the boring foundation: know what personal data each platform holds, why, for how long, and who can reach it. Zero trust is the principle that no user or device is trusted for sitting inside the network; every access request is verified and scoped to the minimum needed. For building systems spread across many sites, that is the practical model. See our guides to zero trust security and data privacy trends.
The AI rules that affect property businesses
In the European Union, the AI Act treats some uses as high-risk, including assessing creditworthiness and deciding access to essential private services, which reaches into lending and some tenant screening. That deadline moved: the Digital Omnibus, in force since 27 July 2026, pushed the high-risk duties from 2 August 2026 to 2 December 2027. Transparency duties for systems that talk to people directly, such as leasing chatbots, sit on a separate track. Our EU AI Act compliance guide sets out the timetable and RegTech solutions the tooling.
In the United States there is no single federal AI statute for property. Exposure runs through existing law: fair housing, fair lending, state privacy statutes and antitrust, as RealPage and SafeRent both show.
- Check cyber insurance wording against your actual controls, before a claim tests it.
- Run vulnerability scans and incident drills on a schedule someone owns.
- Confirm vendor encryption, logging and breach-notification timelines in the contract.
Your 12-month PropTech playbook
Start from your own numbers, not a vendor’s roadmap.
Assess
List your data sources, largest operating costs and most common tenant complaints. The overlap between expensive and annoying is your shortlist. Three months of work-order data usually tells you more than a consultant will.
Select
Prefer open interfaces and documented APIs over one suite that promises everything, because lock-in costs surface at renewal. Compare vendors on integration depth, security posture and realistic implementation time, and ask for a reference from a customer your size.
Pilot
Run narrow, time-boxed pilots with named owners: 3D tours on your slowest-letting units, access control at one property, an energy dashboard on one meter. Set the success threshold before you start, so the result is not a matter of opinion.
Measure
Track a short scorecard: net operating income, energy cost per square foot, vacancy days, turn times and service tickets. Use the same definitions everywhere so portfolios compare honestly.
- Train: give managers a short playbook and a named person to ask.
- Decide: scale what cleared the threshold, stop what did not, and move the budget.
- Revisit: recheck vendor pricing and ownership at every renewal.
Conclusion
The useful split in 2026 is between technology that reduces a cost you already pay and technology that promises a new market. The first group has a track record: automated valuation at scale, building controls that cut energy use, predictive maintenance, 3D tours, transaction platforms with one audit trail. The second, tokenized property and virtual land most obviously, is still mostly promise.
Spend accordingly. Put most of the budget on proven operational wins, cap the experiments, and keep governance ahead of both. RealPage and SafeRent show that the compliance bill for automated decisions arrives after adoption, whether or not you understood the model you bought.
Over the next 12 months, pick two or three pilots tied to numbers you already report, measure them without sentiment, and scale only what clears the bar.
Found this useful?
Make SmartKeys a preferred source on Google, and our articles will surface more often in your Top Stories, AI Overviews, and AI Mode.
Add as Preferred Source
