Are we in a Proptech AI bubble - or just 1997 again?
10 lessons from the dot-com era (and Amazon's internet bubble playbook) that every real estate founder, investor, and exec should apply before this cycle pops.
In August 2024, EliseAI raised $250 million at a $2.2 billion valuation – more than double its worth from just 18 months earlier. The company, which automates property management communications, now generates over $100 million in annual recurring revenue.
Meanwhile, 90% of proptech companies claiming to use AI are essentially rebranding basic automation as “machine learning.” The SEC has started investigating false AI claims across the sector. And private market valuations sit at 15.6x revenue – nearly double the multiples that preceded the dot-com crash.
The global proptech AI sector raised $15 billion in 2024, with $3.2 billion specifically targeting AI-enabled solutions. That’s a recovery from the 83% funding collapse in 2023, when capital fled to $0.68 billion.
I speak to real estate founders, investors, developers and operators every day - and everyone asks the same thing:
Are we watching the birth of a transformative technology cycle, or the inflation of another speculative bubble that’s about to pop?
The answer isn’t what you’d expect. We’re definitely in a bubble – but not the kind that ends in flames.
Today’s Brief:
Why proptech AI is experiencing narrative capture without dot-com style speculation
The Bezos playbook: What Amazon’s 94% crash teaches us about surviving the correction
Which proptech AI models will survive the shakeout (and which are already dead)
The regulatory hammer that nobody’s modeling
What happens next: The 2026-2028 consolidation timeline
The Numbers Don’t Lie (But They Don’t Tell the Full Story Either)
Let’s start with the funding trajectory that’s making everyone nervous:
The proptech funding rollercoaster from 2021 to 2025 tells a story of violent swings. After hitting $6.97 billion at peak hype in 2021, the sector crashed to $0.68 billion in 2023 – an 83% collapse that had VCs declaring proptech dead. Then came the whiplash: $15 billion in 2024, with projections suggesting we could hit $30 billion in 2025.
The concentration of capital tells an even more interesting story. Three companies – EliseAI ($250 million), Bilt ($250 million), and Entrata ($200 million from Blackstone) – captured $650 million of funding between them.
Private market valuations have compressed from 16.3x revenue to 15.6x in the past year. That’s still elevated – public SaaS multiples hover around 8-10x – but the trajectory is toward normalisation, not further inflation.
We’ve Seen This Movie Before (But The Ending Might Be Different)
The dot-com comparison is inevitable, so let’s be precise about it.
In March 2000, the Nasdaq peaked at 5,048. By October 2002, it had crashed 78%. Amazon’s stock fell from $5.33 to $0.30 – a 94% collapse. Of the 457 IPOs in 1999, roughly 75% were dead or effectively worthless within three years.
But here’s what most people forget: Amazon was actually getting stronger during the crash. Sales grew 68% during the worst of it. Average customer spend increased 19%. The company achieved its first profitable quarter in Q4 2001, right as everyone was declaring e-commerce dead.
Jeff Bezos’ response to the crash was instructive. In his 2000 shareholder letter, with the stock down 80%, he wrote: “Amazon is in a stronger position than ever.” He wasn’t delusional. He was looking at customer growth, not stock price.
The parallel to today’s proptech AI sector is striking but imperfect. Like Amazon, the strongest proptech AI companies are seeing real adoption despite market volatility. EliseAI’s $100 million ARR isn’t hypothetical value – it’s actual revenue from property managers who’ve embedded the product into their workflows.
But unlike 2000, we’re not seeing widespread IPOs of profitless companies. The speculation is contained to private markets, where institutional investors can afford to be wrong. That’s a crucial difference.
The Bezos Playbook: Five Lessons From the Last Tech Crash
Studying Amazon’s navigation through the dot-com crash reveals a playbook that today’s proptech AI companies would be wise to follow.
First: “Internet, Shminternet. What matters is great customer service.”
Bezos said this in 1999, at peak hype. Technology wasn’t the strategy – it was the enabler. The same applies to AI in proptech. Companies succeeding today aren’t the ones with the most sophisticated models. They’re the ones solving actual workflow problems. EliseAI doesn’t win because its AI is revolutionary. It wins because property managers save 12 hours per week on tenant communications.
Second: Focus on Fundamentals When Everyone Else is Watching Valuations
During the crash, Amazon cut 15% of its workforce while doubling down on fulfilment centres. Cost discipline during the boom created the runway to survive the bust. Today’s proptech AI leaders are showing similar discipline – lower burn multiples under 2x despite having access to growth capital.
Third: Infrastructure is a Moat, Not a Cost
Amazon’s fulfilment network seemed like dead weight during the crash. It became their insurmountable advantage. In proptech AI, the parallel is data infrastructure. Zillow’s 110 million property records, accumulated over decades, can’t be replicated with venture funding. That’s a moat that survives any correction.
Fourth: Ignore Stock Prices, Measure Customer Value
Amazon’s customer base grew throughout the crash because the value proposition was real. Proptech AI companies that survive will be those with >90% retention at six months, not those with the best PR.
Fifth: The Crash is When You Build Your Lead
Amazon gained more market share during the bust than the boom. Weak competitors died, talent became available, and customer acquisition costs plummeted. The proptech AI correction will create similar opportunities for prepared operators.
Why This Bubble is Different (And Why That Might Not Matter)
We definitely can smell “bubble” in AI startups today. But in my opinion, the fundamentals are different from 2000.
The Narrative Bubble
Every proptech conference in 2024 featured the same buzzwords: “autonomous,” “agentic,” “AI-first.” Aranca Research notes that “merely telling an AI story no longer differentiates.” But we didn’t need a study to understand this. My startup, Buildable, attracts real estate teams in Dubai not because of “AI’, but because we actually solve real problems they deal with every day.
When everyone claims the same innovation, no one actually has it…
But narrative bubbles can deflate without destroying the underlying sector. The key question is whether the narrative has gotten ahead of reality (it has) and by how much (not as much as you’d think).
The Valuation Bubble
Private valuations at 15.6x revenue are clearly elevated. But unlike 2000, these valuations aren’t being marked to market daily through public trading. Private market corrections happen slowly, through down rounds and quiet shutdowns, not spectacular crashes.
More importantly, the capital raising the valuations is different. Blackstone’s $200 million into Entrata is strategic investment from an operator that manages $500 billion in real estate assets. They’re not betting on bless and whistles or disruption. They’re buying operational efficiency.
The Adoption Reality
Here’s the sobering statistic: Only 700 of 7,000 tracked proptech companies actually use AI in any meaningful way. That’s 10% – hardly the “everything is AI” narrative the market is selling.
This suggests we’re earlier in the cycle than the hype would indicate. Mass adoption hasn’t happened yet, which means both that the crash isn’t imminent and that most current players won’t survive to see real scale.
The Regulatory Hammer Nobody’s Modelling
The biggest risk to proptech AI isn’t a funding crash. It’s regulation.
RealPage is under federal investigation for algorithmic price-fixing in rental markets. The case alleges that their revenue management software enables coordinated rent increases across competing properties. If the government wins, it doesn’t just hurt RealPage. It calls into question every AI-powered pricing tool in real estate.
Then there’s data privacy. GDPR in Europe, CCPA in California, and India’s DPDP are creating a patchwork of compliance requirements that add 20-30% to operating expenses. Most proptech AI companies haven’t built these costs into their unit economics.
The combination of algorithmic accountability and data privacy regulation could trigger the correction that market forces alone won’t create. One high-profile enforcement action – imagine the SEC fining a major proptech company for false AI claims – and suddenly every investor starts asking harder questions.
The Consolidation Timeline: 2026-2028
Based on current trajectories, here’s how I think the proptech AI shakeout will unfold:
2025: Peak Narrative
Funding could approach $30 billion as everyone rushes to claim AI positioning. Valuations stay elevated but stop growing. Switched on investors start getting selective.
2026: The Reality Check
One or two high-profile failures – likely companies that raised at billion-dollar valuations without achieving product-market fit. Regulatory scrutiny intensifies. Private valuations compress from 15.6x to 10-12x. Funding drops 40% but doesn’t collapse.
2027: The Consolidation
Platform players like Yardi and AppFolio acquire the best point solutions. International expansion becomes the growth story as US/UK markets saturate. Only companies with embedded workflows and real moats survive as independents.
2028: The New Normal
Proptech AI becomes just “proptech.” The companies that survived own specific workflows completely. Valuations stabilise at 8-10x revenue. The next innovation cycle begins.
One thing I have left out is the comparison between modes. So, in the internet boom, we had new modes of mobile, cloud computing, and APIs. And with the AI boom, we have new modes of voice AI, and text response, and even visual. I’m not sure whether this either makes the crash and recovery faster and more shallow, or heavier?
What Survives: The Three Categories of Winners
Not every proptech AI company faces the same risk. Based on historical patterns and current market dynamics, survivors will fall into three categories:
The Workflow Embedders
Companies like EliseAI that become so embedded in daily operations that ripping them out would be painful. These businesses are not point solutions or minor features. They actually replace previous manual workflows and have the stripes to prove it. They’re systems of record that own critical workflows. Retention over 90% at year two, with switching costs measured in months of disruption.
The Data Fortresses
Platforms sitting on proprietary, impossible-to-replicate data sets. Think MLS systems with decades of transaction history or inspection platforms with millions of property condition reports. The AI is almost secondary – it’s the data gravity that creates the moat.
The Infrastructure Builders
The overlooked opportunity: companies building the rails everyone else needs. Payment processing for property management. Compliance automation for algorithmic pricing. Identity verification for digital leasing. Boring, critical, defensible.
Point solutions without one of these characteristics are already dead.
The Bottom Line: It’s 1997, Not 1999
YES, we are in a proptech AI bubble. But it’s more 1997 than 1999. The narrative has gotten ahead of reality. Valuations are elevated. AI-washing is rampant. But the fundamental transformation is real.
The difference between 1997 and 1999 was two years and a lot of bad decision-making. The proptech AI sector has time to course-correct before a crash becomes inevitable. Whether it does depends on three factors:
First, capital discipline. If burn multiples stay under 2x and companies focus on profitability over growth, the sector can deflate gradually rather than collapse.
Second, regulatory navigation. Companies that build compliance into their DNA survive. Those that ignore it will struggle massively.
Third, customer value over technology novelty. The winners won’t be the ones with the best AI. They’ll be the ones that make property management 10x better, whether that’s through AI, automation, or just better software.
For real estate operators and investors, the playbook is clear:
Bet on companies with embedded workflows and real retention metrics. Avoid anything that sounds like “Uber for property management.” Expect valuations to compress 30-40% over the next 24 months. Use the correction to acquire strategic capabilities at reasonable prices.
The proptech AI transformation is real. The bubble is real too.
We need to look past the headlines of companies raising the most money or making the boldest AI claims. There is a lot of bull**it out there. The winners will be the companies that solve real problems for real customers and build the operational discipline to survive when the music stops.
That’s the lesson from Amazon, from the dot-com crash, and from every technology cycle before this one. Technology changes. Market dynamics don’t.
The correction is coming. But for prepared operators, it’s not a crash – it’s an opportunity.
P.S. I launched my startup Buildable earlier this year. If you’re analysing real estate deals in Dubai, I’ve built this for you. Book a demo on the site if you want a run-through!
Thanks for reading - see you next week!