What Shepherd’s $42M Bet Tells Us About Where ConTech Is Actually Going
Shepherd, the AI-native commercial insurer for construction and infrastructure, raised a $42 million Series B in March 2026, bringing total funding to $67 million. The story goes well beyond an insurtech deal. Shepherd’s thesis – that automation speeds up existing workflows while autonomy rebuilds them entirely – is the clearest articulation in construction’s adjacent industries of where genuinely ambitious ConTech needs to go. For construction executives watching this space, there’s a direct read-across to what’s being funded, why it’s working, and what it signals about where the real value in the next wave of ConTech will sit.
Construction insurance isn’t a sector most people in ConTech spend much time thinking about. It’s a back-office consideration – something to sort out before work starts, handed off to a broker, and largely forgotten until a claim surfaces. The assumption has always been that whoever does it cheapest and fastest enough wins. Not exactly fertile ground for a $67 million venture-backed platform.
And yet Shepherd has grown revenue more than 7x in 24 months, now insures over $400 billion in project value across 1,500+ policies, and counts the physical infrastructure behind the world’s leading AI labs – data centers, chip fabrication facilities, hyperscaler campuses – among its client base. Intact Private Capital, the venture arm of one of the world’s largest commercial insurance companies, led the Series B. That’s not a speculative bet on an early-stage concept. That’s a strategic institution putting serious money behind a model it believes will become the new standard.
The reason it’s worth paying attention to – particularly if you’re a construction executive or a ConTech founder – is that Shepherd’s core argument isn’t really about insurance. It’s about what separates technology that makes existing processes marginally better from technology that makes them structurally different. And that distinction maps almost perfectly onto the challenge facing most of construction’s digital tools right now.
Shepherd’s $42M raise is a signal about where AI investment in construction-adjacent industries is headed – toward autonomy (rebuilding workflows from scratch) rather than automation (speeding up what already exists). For construction executives evaluating tech investments, and for founders building ConTech products, the framework Shepherd has built is directly applicable to almost every category of construction software in play today.
Automation vs. Autonomy: Why the Distinction Matters
The framework Shepherd built – and what it means for construction
Shepherd’s CTO and co-founder Mo Mahallawy spelled out the company’s philosophy plainly at the time of the Series B announcement. Automation, in their framing, is about making the existing system faster. You take a process that involves people manually copying data between systems, and you make it happen in seconds instead of hours. Autonomy is something else: it’s about asking whether the system itself needs to exist in the same form at all, and rebuilding the workflow around what AI can actually do rather than around what humans used to have to do.
In traditional commercial underwriting, a large construction project submission arrives via email with 15 or more attached files – spreadsheets, PDFs, drawings, schedules. An army of underwriting assistants manually parses that data, keys it into internal systems, and routes it for review. A typical carrier takes two to three weeks to respond with a quote. Shepherd’s AI ingests, structures, and analyzes that same submission in seconds. Underwriters at Shepherd currently handle 25-30 accounts per month – compared to the industry standard of 5-10. The target is 200. Not because they’ve hired more people, but because the AI now does the intake, data enrichment, risk analysis, and preliminary pricing, and the underwriter steps in only for strategic decisions and exceptions.
“Today, an underwriter manually processes about 20 accounts a month and wastes significant time on submissions outside their appetite. In our model, an underwriter will oversee 200.”Mo Mahallawy, CTO and Co-Founder, Shepherd
The self-driving analogy Mahallawy uses isn’t marketing. It’s a practical engineering framework. Waymo didn’t build a car with better cruise control. It redefined what the driving task is and what a human needs to be involved in. Shepherd is doing the same thing to underwriting – not incrementally, but structurally. And that’s precisely the kind of thinking that’s missing from the majority of ConTech product development happening today.
Why Construction Insurance Was the Right Place to Start
A hard market, a data advantage, and a direct line to project outcomes
Shepherd didn’t pick construction insurance arbitrarily. Commercial construction insurance has been in one of its hardest markets in decades, with premiums rising steeply from the early 2020s onward as legacy carriers deployed broad rate increases rather than account-level pricing adjustments. That market dynamic created a genuine opening for a platform that could price risk more accurately, faster, and based on how projects are actually being run rather than on static historical averages.
The data edge Shepherd has built is worth understanding. Because the platform quotes less than 40% of what it receives – meaning it sees a huge volume of submissions it doesn’t ultimately write – it has accumulated four years of structured, industry-specific loss data across every submission, regardless of outcome. No traditional carrier has that. They only see data on the policies they write. Shepherd sees the full picture of what comes through the door, and that asymmetry in data quality is becoming a compounding advantage as the model matures.
- Full-funnel data capture. Shepherd structures every submission it receives, not just the ones it writes. After four years, that’s a dataset no incumbent can replicate from policy history alone.
- Real-time field integration. The platform pulls live data from Procore, Autodesk, OpenSpace, DroneDeploy, and Samsara directly into underwriting. Legacy carriers price on static applications. Shepherd prices on how work is actually being delivered.
- Behavior-based pricing. Contractors who use safety technology and demonstrate good site discipline get that reflected directly in their premiums via the Shepherd Savings Program. The model rewards actual performance, not a historical average of the class.
Justin Levine, Shepherd’s CEO, described the direct practical benefit to contractors on the Bricks & Bytes podcast: contractors building insurance cost assumptions into bids years before they actually buy the policy are routinely caught out by rate movements they didn’t see coming. A market that has been running hard for six years means the gap between bid-stage assumptions and actual premium is significant and painful. Shepherd’s rate certainty program – offering predictable renewal rates over multiple years to contractors who maintain consistent technology usage – is a direct attack on that problem. It’s not just a nice feature. For a contractor managing razor-thin margins across multi-year projects, predictable insurance cost is operationally significant.
The AI Infrastructure Play and What It Tells Us About ConTech Investment Signals
Following the physical layer of the AI economy
Shepherd’s client base now includes the firms building and operating the physical infrastructure behind the AI economy – data centers, semiconductor fabrication facilities, energy generation assets. That’s not incidental. It’s the most capital-intensive, schedule-pressured, technically complex category of construction work happening in the US right now, and it’s moving faster than the legacy insurance market can serve it.
Traditional commercial carriers take weeks to quote complex projects. Data center construction moves on timelines where a two-week insurance delay has real schedule and cost consequences. Shepherd’s ability to turn around a quote in under two days – moving toward same-day in some cases – isn’t a minor improvement. For project teams working on $2 billion data center campuses, it’s the difference between a viable construction schedule and a bottleneck.
The strategic investor in the round, Intact Private Capital (the venture arm of one of Canada’s largest insurance companies), is also now a long-term capacity provider for Shepherd’s policies. That structure matters. Shepherd operates as a Managing General Underwriter (MGU), partnering with large carriers to provide coverage while retaining the underwriting authority and data advantage in-house. Intact backing the round isn’t just financial – it’s a signal that the incumbent insurance market is preparing for the autonomous underwriting model to become the competitive standard, and they’d rather be positioned inside it than disrupted by it from the outside.
The Read-Across for Construction Executives
What the autonomy thesis means for your technology decisions
For a construction executive evaluating technology investments, the automation vs. autonomy framework is one of the most practically useful lenses available right now. Most of the ConTech market – by volume and by marketing spend – is selling automation. Better dashboards. Faster reporting. Digitized paperwork. Those tools have value, but they operate within the existing workflow structure. They make the current process marginally less painful. They don’t change what the process is.
Genuine autonomy in construction technology looks different. It asks: if we rebuilt this workflow from scratch around AI, what would the human’s job actually be? For underwriting, Shepherd’s answer is that the human becomes a portfolio strategist and exception handler rather than a data processor. The same logic applies to estimating – where the question isn’t “how do we make quantity takeoff faster?” but “how do we give a senior estimator 200 projects’ worth of market pricing data to make judgment calls on, rather than having them manually build up unit rates?” It applies to site supervision, procurement, scheduling, and most of the other functions where construction firms are currently spending technology budgets on automation-layer tools.
Dustin DeVan, who built and sold BuildingConnected and now runs Ediphi, made the pricing version of this argument on the same Bricks & Bytes episode: the construction industry’s budget overrun problem isn’t primarily a collaboration problem or a design coordination problem. It’s a pricing certainty problem. And the technology worth investing in is the kind that attacks that problem directly – not the kind that wraps a better interface around the existing estimation workflow.
Stay ahead of the deals and ideas reshaping construction
2,500+ construction executives and ConTech founders read Bricks & Bytes for the signal, not the noise.
Join 2500+ ReadersWhat Shepherd’s Raise Signals for ConTech Founders
The funding thesis that should inform how you build and pitch
For ConTech founders, the Shepherd raise carries a few pointed messages. First, adjacency to construction – in this case, insurance – is increasingly a viable ConTech investment category in its own right, provided the product is built around construction-specific data and serves construction-specific operational needs. Shepherd isn’t trying to disrupt commercial insurance broadly. It started in construction, built proprietary loss data in construction, and is expanding into sectors that look structurally similar (renewable energy, AI infrastructure). That focused, data-first approach is what produced 7x revenue growth without compromising underwriting discipline, as Costanoa’s Mark Selcow noted in his comments on the round.
Second, strategic investors from adjacent industries – insurance carriers, equipment manufacturers, materials companies – are becoming an increasingly important funding source for construction tech. Intact Private Capital’s role here isn’t just capital. It’s long-term capacity provision, data partnership, and market access. Founders who can identify the adjacent-industry incumbent that benefits most from their platform succeeding – and structure the relationship accordingly – are building a different kind of moat than pure software revenue.
Third, and most directly: investors at Series B and beyond are looking for evidence that a platform is rebuilding its domain rather than optimizing it. Seven times revenue growth in 24 months is the output of a genuinely differentiated model. That growth didn’t come from being marginally faster at the same thing. It came from doing something structurally different – no underwriting assistants, AI-native intake, behavior-based pricing, real-time field data integration – in a market where the incumbent approach hadn’t changed meaningfully in decades.
| Dimension | Traditional Insurance Carrier | Shepherd (Autonomous Model) | ConTech Automation Parallel | ConTech Autonomy Parallel | Implication for Construction Execs |
|---|---|---|---|---|---|
| Workflow approach | Manual data entry, siloed systems | AI ingests and structures submissions in seconds | Digital forms replacing paper | AI-native estimating from live market data | Ask whether the tool rebuilds the workflow or digitizes it |
| Quote turnaround | 2-3 weeks | Under 2 days, moving toward same-day | Faster RFI response tracking | Real-time subcontractor pricing and procurement | Speed advantage only matters if accuracy holds – verify both |
| Capacity per specialist | 5-10 accounts/month per underwriter | 25-30 today, targeting 200 | PM managing more projects with dashboards | PM as portfolio orchestrator across AI-monitored sites | Capacity gains require the human role to be redefined, not just supported |
| Pricing basis | Broad rate classes, historical averages | Behavior-based, live field data from Procore/Autodesk | Cost tracking against baseline budget | Dynamic cost forecasting based on live productivity data | Data-led pricing requires your field data to be clean and structured |
| Rate certainty | Annual renewal, market-driven volatility | Multi-year rate lock tied to technology usage | Fixed-price contract management tools | Long-range cost certainty through predictive market data | Predictability on cost inputs is worth more than marginal software savings on one workflow |
| Data advantage | Own policy history only | Full submission data, including declined risks | Project-level analytics in isolation | Cross-portfolio benchmarking and market intelligence | Platform value compounds with data breadth – evaluate vendor data strategy, not just features |
Shepherd is an AI-native commercial insurance platform founded in 2021, focused on construction, renewable energy, and AI infrastructure. The company raised a $42 million Series B in March 2026, led by Intact Private Capital with participation from Spark Capital and Costanoa Ventures, bringing total funding to $67 million. The capital is earmarked for team growth, platform development, and expansion into renewable energy and power verticals. Shepherd operates as a Managing General Underwriter (MGU), retaining underwriting authority and using proprietary AI to price risk up to 10x faster than traditional carriers. (Source: Shepherd Press Release)
Automation means using technology to make an existing process faster or less labor-intensive – replacing manual steps with digital ones while keeping the workflow structure intact. Autonomy means rebuilding the workflow entirely around what AI can do, redefining the human role from processor to strategist or exception-handler. Shepherd’s framing, borrowed from the self-driving vehicle industry (Waymo’s Level 1-5 autonomy scale), is that most current AI in construction and insurance is operating at the automation layer. Their platform is attempting to operate at the autonomy layer – no underwriting assistants, AI-native intake and risk analysis, human underwriters handling only the strategic and exception work. (Source: Shepherd CEO Blog)
The Shepherd Savings Program evaluates behavioral signals from construction technology tools a contractor uses in the field – including Procore, Autodesk, OpenSpace, and others – and applies those signals to premium pricing. Contractors who demonstrate consistent, safe operating practices through their technology usage receive credits toward improved rates. Shepherd can also offer multi-year rate certainty to contractors who maintain consistent utilization of these tools, providing predictable renewal pricing at year two and year three and beyond. For contractors embedding insurance cost assumptions into bids years before project start, that predictability has direct commercial value. (Source: Bricks & Bytes Podcast)
Intact Private Capital is the venture arm of Intact Insurance, one of the largest commercial insurance companies in the world and the biggest insurer in Canada. Their participation in Shepherd’s Series B isn’t purely financial – they are also acting as a long-term capacity provider for Shepherd’s policies, meaning they are backing the policies Shepherd writes. That structure signals that an incumbent insurance major believes the autonomous underwriting model will become the competitive standard, and has chosen to be positioned inside the disruption rather than on the outside of it. For construction executives, it’s worth noting that the same dynamic – incumbent players co-investing in the platforms that may eventually replace their current model – is playing out across ConTech as well. (Source: Shepherd Press Release)
Shepherd covers a broad range of commercial construction – K-12 schools, healthcare facilities, higher education, utility-scale renewable energy projects, street and road infrastructure, and multi-billion-dollar data center campuses. Its products include general liability, workers’ compensation, commercial auto, excess liability, and builder’s risk. The platform’s recent expansion into renewable energy and power follows the expansion into builder’s risk in April 2025. Clients include contractors at the middle-market to enterprise scale, with average client revenue above $100 million annually. Named clients include Skanska, Bernards, and Nibbi Brothers. (Source: Shepherd Press Release)
The most immediate benefit is speed. Shepherd targets quote turnaround of under two days – moving toward same-day in some cases – compared to the two to three week standard at traditional carriers. For contractors managing complex schedules, especially on data center and infrastructure projects where delays have real cost consequences, that speed difference is operationally significant. Beyond speed, the behavior-based pricing model means that investment in safety technology and site management practices has a direct and measurable impact on insurance cost. Contractors who use Procore, Autodesk, or other connected tools and demonstrate consistent, safe performance can earn lower premiums and multi-year rate certainty. That’s a fundamentally different relationship between technology investment and insurance cost than the current market offers. (Source: Bricks & Bytes Podcast)
Shepherd’s 7x revenue growth in 24 months, with no compromise to underwriting discipline according to its investors, is the output of a model that is structurally differentiated rather than incrementally better. The investment pattern – a strategic incumbent (Intact) co-investing alongside established venture firms (Spark, Costanoa), with the incumbent also acting as a capacity provider – reflects a broader trend of adjacent-industry majors positioning inside ConTech disruption rather than outside it. For founders, it signals that construction-adjacent categories with strong data moats and genuine workflow autonomy are fundable at significant scale. For construction executives, it signals that the technology partners worth taking seriously are the ones rebuilding their domain rather than optimizing it. (Source: BB VC Hot Takes)
Related Articles
The Insurance Game Is Changing and You Should Be Aware
A Venture Capitalist’s Hot Takes On Construction Tech
