Why Founders Build IoT
The Internet of Things promised to connect the physical and digital worlds, creating smart cities, automated homes, and intelligent supply chains. This vision attracted 25 startups in our dataset that collectively burned through $3.2 billion before failing. You were drawn to IoT because the addressable market seemed infinite: every object, every process, every interaction could theoretically be enhanced with sensors and connectivity. The hardware renaissance of the 2010s made it feel accessible, with cheap sensors, ubiquitous wireless connectivity, and cloud infrastructure that could handle massive data streams.
What made IoT particularly seductive was its cross-industry applicability. Among the 25 failures, 11 targeted consumers with smart home devices and connected products, while 8 focused on industrial applications like supply chain monitoring and agricultural automation. The remaining startups spread across IT infrastructure, materials tracking, and utility management. This diversity masked a fundamental challenge: IoT startups weren't just building software or hardware, they were building both, plus the connectivity layer, plus the data infrastructure, plus the user experience that made it all worthwhile.
The category peaked in failure rates between 2017 and 2020, with a surprising resurgence in 2025, suggesting that even recent AI-enhanced IoT plays haven't solved the core problems. With an average lifespan of 4.8 years, these startups typically survived long enough to achieve product-market fit on paper, deploy hardware at scale, and then discover that the economics simply didn't work. The biggest casualties like Mobike ($900M), Bird ($776M), and Infarm ($600M) all died from unit economics, not lack of vision or technology. They proved the concept but couldn't prove the business model.
How IoT Startups Die
IoT startups die primarily from two interconnected problems: they get crushed by competition (36%) or discover their unit economics don't work (32%). Together, these account for 68% of all failures in this category. The pattern is brutally consistent: you build compelling hardware, achieve initial traction, then realize that either well-funded competitors can undercut your pricing or that the cost of manufacturing, deploying, maintaining, and supporting physical devices in the real world exceeds what customers will pay. Unlike pure software plays, you can't easily pivot when the hardware is already manufactured and deployed.
The remaining failures split between running out of cash before achieving scale (20%), discovering no real market need (8%), and outright product failures (4%). The relatively low rate of product/tech failures is telling: IoT founders generally succeed at building working technology. The challenge isn't making devices that connect and collect data; it's building a sustainable business around them.
IoT markets attract well-capitalized incumbents and tech giants who can subsidize hardware losses with adjacent revenue streams. You're competing against companies that can afford to lose money on devices to capture data, platform lock-in, or ecosystem control. Once your proof of concept validates the market, larger players can replicate your functionality with superior distribution and deeper pockets.
SEE ANTIPATTERN →Physical devices have brutal economics: manufacturing costs, inventory risk, warranty obligations, field maintenance, connectivity fees, and customer support all scale linearly with users. The top three failures by capital burned (Mobike, Bird, Infarm) all died from unit economics, proving that even massive scale doesn't necessarily fix the math. Your hardware costs stay high while software competitors' marginal costs approach zero.
SEE ANTIPATTERN →IoT startups require multiple expensive rounds to fund hardware development, manufacturing scale-up, and market deployment before revenue materializes. You need capital for inventory, tooling, certifications, and field operations that software startups never face. When the funding environment tightens, you're left with physical assets that can't be quickly monetized and burn rates that can't be easily reduced.
SEE ANTIPATTERN →The low rate of market need failures suggests IoT founders generally identify real problems. However, the few who fail here typically discover that while the problem exists, customers aren't willing to change behavior or pay enough to justify the solution's complexity. The friction of installing, maintaining, and integrating physical devices creates a higher bar for value delivery than software alone.
SEE ANTIPATTERN →Pure technical failure is rare in IoT, indicating that founders generally succeed at the engineering challenge of building connected devices. When products do fail, it's typically due to underestimating real-world deployment challenges: battery life in actual conditions, connectivity in edge cases, or durability under sustained use. The technology works in the lab but fails in the field.
SEE ANTIPATTERN →The Biggest IoT Failures
These are the most well-funded IoT startups that failed. Click any card to read the full autopsy.
What To Build Today
The landscape has shifted dramatically since most of these failures occurred. Edge AI chips now enable sophisticated on-device processing that reduces connectivity costs and latency. Energy harvesting and improved battery technology address power constraints that plagued earlier generations. Most importantly, the infrastructure layer has commoditized: you no longer need to build your own connectivity stack, device management platform, or data pipeline. The pivot themes from failed startups reveal a clear pattern: founders want to rebuild with AI-first architectures, B2B focus, and modular deployment models that reduce upfront capital requirements.
The key insight from the failure data is that consumer IoT with complex hardware deployment models is a trap, while B2B applications with clear ROI and predictable unit economics remain viable. The successful rebuild opportunities share common characteristics: they leverage existing infrastructure rather than building it, they focus on high-value use cases where customers have budget authority and pain is acute, and they design for software-like economics even when hardware is involved. The integration of AI isn't just a feature upgrade; it fundamentally changes the value proposition from data collection to automated decision-making.
You should look for opportunities where IoT enables a business model transformation rather than just operational efficiency. The failed startups proved that customers won't pay much for dashboards and alerts. They will pay for systems that autonomously manage resources, predict failures before they occur, or eliminate entire categories of labor. The rebuild opportunities below reflect this shift from monitoring to autonomous action.
AI-Native Industrial Predictive Maintenance
Build edge AI systems that predict equipment failures and autonomously schedule maintenance for high-value industrial assets. Unlike earlier IoT monitoring plays, modern edge processors can run sophisticated models locally, eliminating connectivity costs and enabling real-time intervention. Target industries with clear failure costs (manufacturing downtime, utility outages) where your system's ROI is measurable in days, not months.
Modular Smart Building Systems for Commercial Real Estate
Create plug-and-play IoT modules for HVAC, lighting, and occupancy optimization that building managers can deploy without infrastructure overhaul. The opportunity is in software-driven energy savings with hardware as a loss leader or rental model. Commercial real estate has budget authority, long-term contracts, and direct financial incentives (utility costs) that consumer smart home plays lacked.
Vertical-Specific Supply Chain Intelligence
Focus on single industries (pharmaceuticals, perishable food, high-value electronics) where tracking and condition monitoring have regulatory or insurance implications. Sell to enterprises that already budget for compliance and risk management. Your differentiation is in automated compliance reporting and insurance premium reduction, not just visibility. The unit economics work because customers have existing budget lines you're capturing.
IoT-as-a-Service for Equipment Manufacturers
Partner with industrial equipment makers to embed connectivity and intelligence into their products, enabling them to offer uptime guarantees and usage-based pricing. You handle the entire IoT stack as a white-label service. This model solves the unit economics problem by making hardware manufacturers your customers rather than end users, and it leverages their existing sales channels and customer relationships.
Survival Guide for IoT
Key Takeaways
- Unit economics killed 32% of IoT startups and contributed to many competition failures. Calculate your fully-loaded cost per device including manufacturing, deployment, connectivity, support, and warranty before you raise your Series A. If the math doesn't work at 100,000 units, it won't work at 1 million.
- Consumer IoT is a graveyard unless you have a credible path to being acquired by a platform player. 11 of 25 failures targeted consumers, and the biggest casualties (Mobike, Bird) had massive user bases but couldn't make the economics work. B2B customers with clear ROI and budget authority are far more forgiving of hardware costs.
- Competition crushed 36% of failures because IoT markets attract well-funded incumbents once you prove demand. Your defensibility must come from proprietary data, vertical-specific expertise, or integration lock-in, not just being first to market with connected hardware. Plan for competitors with 10x your resources.
- The average 4.8-year lifespan means you'll likely survive long enough to deploy at scale before discovering fatal flaws. Front-load your unit economics validation with pilot programs that include full-cycle costs: installation, support calls, device failures, and replacement logistics. Don't extrapolate from lab conditions.
- The 2019-2020 failure peak coincided with the end of cheap capital for hardware startups. IoT requires sustained funding through multiple expensive milestones. Structure your roadmap so you can demonstrate improving unit economics at each funding stage, not just user growth or engagement metrics.
- Only 8% failed from no market need, suggesting IoT founders are good at identifying real problems. The challenge is that customers often won't pay enough to justify the solution's complexity. Validate willingness to pay and procurement processes before building hardware, not just problem severity.
- The pivot themes from failed startups all emphasize AI integration and B2B models. Modern edge AI fundamentally changes IoT economics by enabling autonomous action rather than just monitoring. If your value proposition is a dashboard showing sensor data, you're building 2015 technology in 2025.
Red Flags to Watch
- Your unit economics depend on achieving massive scale before profitability. If you need millions of devices deployed to make the math work, you'll run out of cash first like 20% of failures did.
- You're building proprietary connectivity infrastructure or device management platforms. These are now commoditized services. Capital spent on infrastructure is capital not spent on differentiation.
- Your primary value proposition is data visibility rather than automated action. Customers have dashboard fatigue and won't pay premium prices for monitoring alone.
- You're targeting consumers with hardware that requires professional installation or ongoing maintenance. The support costs will destroy your margins and customer acquisition costs will exceed lifetime value.
- Well-funded competitors or platform players could replicate your core functionality in 6-12 months. Without deep moats, you're validating a market for someone else to dominate.
Metrics That Matter
- Fully-loaded cost per device per month including hardware amortization, connectivity, support, and replacement reserves. This must be significantly below your monthly revenue per device with margin for customer acquisition costs.
- Time and cost to deploy per device in real-world conditions, not lab environments. Installation complexity killed many consumer plays and inflated B2B sales cycles beyond what financial models assumed.
- Device failure and replacement rate in the field after 12+ months of operation. Early prototypes work fine; production hardware at scale reveals quality issues that destroy unit economics.
- Customer payback period on their investment in your system. B2B customers need to see ROI within their budget cycle (typically 12-18 months) or they won't renew and your churn will be fatal.
- Percentage of revenue from software and services versus hardware sales. Sustainable IoT businesses shift toward recurring revenue that has software-like margins, using hardware as an enabler rather than the primary revenue source.
Also Study These Categories
All IoT Failures
VIEW ALL 25 ON THE GRAVEYARD →BROWSE ALL DEEP DIVES →