Overview
Teleo turned existing dozers, loaders, and haul trucks into remotely operated and supervised-autonomous machines. Founded in 2019 by former Google and Lyft autonomy leaders Vinay Shet and Rom Clément, the company pursued a pragmatic middle ground: keep a human in the loop, let that operator oversee several machines, and retrofit the mixed fleet a contractor already owned.[1]
That incremental strategy produced named customers, dealer distribution, $29.8 million in reported funding, and a strategic buyer. Havoc acquired Teleo in March 2026 to add ground vehicles to an autonomy system spanning sea, air, and land.[3][15] The undisclosed price prevents a financial verdict. The operating story is clearer: Teleo made autonomy deployable before it was complete, but scaling safety-critical hardware, field service, and fleet software favored a larger platform.
Founding Story
Shet and Clément brought unusually relevant backgrounds to the problem. Shet had led product work for Lyft's self-driving program after working on Google Street View, Maps, and reCAPTCHA. Clément had worked on self-driving systems at Lyft and on Street View at Google.[1] Public sources do not explain exactly how the pair met. They do show a shared conclusion from consumer autonomous vehicles: waiting for complete autonomy delayed customer value.
Shet described that conclusion directly in a 2023 interview: “I had a front row seat to see that while there is a lot of advancement in machine learning and AI, there’s a limit to how far it can go in building a fully finished product.”[7] Construction and mining offered a different deployment surface. Worksites were bounded, machines repeated defined tasks, and operators already faced dangerous, remote, or uncomfortable conditions.
The product premise was supervised autonomy rather than an empty cab at any cost. In the company's 2023 expansion announcement, Shet said, “We founded Teleo to help contractors supercharge their operators’ efforts by turning their equipment into semi-autonomous machines.”[5] A remote operator could handle difficult moments while software performed repeatable movements. Teleoperation also put the company into real workflows sooner, where it could collect the operating data needed to improve automation.
That choice shaped the company. Teleo built an OEM-agnostic retrofit, a site network, a command center, and an analytics layer for machines that contractors already owned. The compromise increased integration work, but it matched how contractors bought and maintained equipment.
Timeline
- 2019: Shet and Clément founded Teleo.[1]
- Winter 2020: Teleo joined Y Combinator.[1]
- June 2022: Teleo raised a $12 million Series A led by UP.Partners.[2]
- March–April 2023: The company announced Teichert, Tomahawk Construction, and John Aarts Group as customers and launched a dealer network across North America and Europe.[5][9]
- October 2023: Teleo said eight new customers had ordered 42 machines and Tomahawk would deploy autonomous functions on an active jobsite.[6]
- November 2024: Two Series A extensions added $16.2 million and brought company-reported funding to $29.8 million.[3]
- March 11, 2026: Havoc closed its acquisition of Teleo; terms were undisclosed.[15]
What They Built
Teleo sold an integrated system. A customer first installed a mesh network on the worksite. Teleo then fitted a universal hardware kit to an existing machine and installed a remote command center. Operators used live camera feeds and controls to drive equipment from a desk. The same interface let them hand repetitive tasks to autonomy and intervene when conditions exceeded the software's operating envelope.[4]
The cross-brand retrofit was the commercial wedge. Contractors often owned fleets spanning Caterpillar, John Deere, Komatsu, and other manufacturers. Replacing that installed base to gain autonomy would bind the customer to an expensive equipment cycle. Teleo instead tried to make the existing fleet programmable.
Its second product layer was operator multiplication. One person could supervise multiple machines and switch between sites, reducing idle time when a machine lacked an available driver. Teleo Insite added live and historical operating data through a browser, giving managers visibility that a cab-bound operator could not provide.[4]
The system's boundaries were part of the design. A human handled irregular maneuvers and judgment-heavy work. Autonomy handled defined, repeated operations such as material hauling. This approach traded the elegance of full autonomy for deployability and a safer fallback. It also created hard dependencies: reliable networking, low-latency video, fail-safe machine control, trained operators, dealer support, and site-specific commissioning all had to work together.
Market Position
Target Customers
Teleo began with heavy civil construction and mining contractors that owned mixed fleets and could identify repetitive tasks on controlled sites. The company later reported orders from pulp and paper, logging, port logistics, munitions clearing, and agriculture.[3] These customers shared three conditions: costly machines, constrained operator supply, and work that exposed people to dust, weather, distance, or collision risk.
Market Size
The addressable market was substantial even before mining and logistics. The Census Bureau put U.S. construction spending at a $2.1382 trillion annualized rate in May 2025.[12] BLS counted 539,500 construction-equipment-operator jobs in 2024, with 46,200 openings projected each year through 2034 and a $58,320 median wage.[10] AGC reported in 2025 that 92% of contractors struggled to fill positions.[11]
Teleo's reachable segment was smaller than those industry totals: fleets with repetitive workflows, sufficient site connectivity, measurable downtime, and budgets for retrofits and operational change. Public data does not disclose pricing or deployed fleet utilization, so a credible revenue estimate is unavailable.
Competition
Teleo sat between established OEM autonomy and newer retrofit startups. Caterpillar reported 827 autonomous haul trucks operating in 2025.[13] Komatsu reported more than 900 FrontRunner trucks and over 10 billion tons moved as of October 2025.[14] Those systems brought installed support networks and long operating histories, especially in mining.
Teleo countered with heterogeneous fleets, smaller construction settings, and a human-in-the-loop migration path. Dealers were crucial. F-Prime's investment thesis explicitly identified OEM relationships and dealer networks as the channel that sells, services, supports, and teaches autonomy.[8] The dealer strategy reduced a startup's field-service burden, but incumbents still controlled machine interfaces, warranties, parts, and customer trust.
Business Model
Teleo sold hardware, software, and operational support into an enterprise workflow. Its published terms describe recurring annual subscription services, while product materials show a required retrofit kit, site network, command center, training, and analytics. The company did not publish prices, revenue, gross margin, renewal rates, or service costs.
The economic pitch rested on utilization and operator productivity. Tomahawk said two trucks had sat idle for 10 months because it lacked operators; a remote command center could put them back to work.[9] If one operator truly supervised several active machines, the customer gained output without a proportional increase in labor. If the machine still demanded frequent intervention or specialist support, the savings narrowed.
Funding underwrote the gap between a working system and repeatable deployment. The $12 million Series A in 2022 and $16.2 million in 2024 extensions supplied capital for hardware, AI development, dealer enablement, and field installations.[2][18][3]
Traction
Teleo disclosed orders and demonstrations rather than recurring revenue. In 2023 it named three early contractors, then said eight new customers had collectively ordered 42 machines.[5][6] By November 2024 it reported 34 more machine orders from nine customers and a dealer network reaching the United States, Canada, Europe, Australia, North Africa, and the Middle East.[3]
Operators in Dallas controlled machines at a site in Finland, more than 5,000 miles away.[3] That demonstration proved distance. Public announcements did not separate orders, completed retrofits, paid deployments, autonomous hours, or renewals, leaving production depth unresolved.
Post-Mortem
Supervision turned an autonomy limit into a product
Teleo's most important choice was accepting that full autonomy would arrive unevenly. The company used human operators as the exception handler and sold the useful subset first. This made jobsites a source of customer value and training data instead of a future promise. It also avoided the binary standard that had consumed far more capital in robotaxis.
The counterargument was strong. A 2020 Hacker News discussion challenged both the labor-shortage premise and the safety of removing an operator from the site.[17] Those objections set the burden of proof: a retrofit had to beat the cost and performance of hiring or training an operator. Teleo answered with idle-machine cases, multi-machine supervision, and deployments across named contractors. It did not publish enough utilization or cost data to settle the ROI question.
Mixed fleets created a wedge and an integration tax
OEM independence matched contractor reality and differentiated Teleo from Caterpillar and Komatsu. The same promise multiplied hardware variants, machine-control interfaces, safety validation, field installation, and support obligations. A software update could not erase the physical work required at each site.
Teleo's attempted remedy was distribution through equipment dealers. Dealers already held customer relationships and service capacity, while Teleo supplied the autonomy layer.[8] That model expanded reach, but it introduced another organization into commissioning and support. Scaling required product maturity and channel coordination at the same time.
The acquisition joined a broader control architecture
Havoc said customer demand for a single system controlling assets across domains drove its purchase of Teleo and drone company Mavrik. Havoc CEO Paul Lwin said customers wanted autonomous systems “to operate together as a coordinated force rather than isolated platforms.”[15] Teleo supplied ground autonomy and demanding industrial deployments; Havoc supplied a broader command architecture and new defense distribution.
Shet framed continuity rather than retreat. He wrote that the team would keep building fleet-based ground autonomy while connecting it to Havoc's sea-air-land system.[16] The product survived inside the buyer. The standalone company did not.
Consolidation around orchestration explains the deal. Individual autonomy products had to build perception, control, operator interfaces, networking, deployment, and support. Havoc's thesis was that customers increasingly wanted those assets coordinated through one operational framework. Teleo became more valuable as a ground component of a multi-domain system while independence kept the full-stack cost on one startup.
The outcome remains financially opaque
Teleo raised $29.8 million and found a strategic acquirer, which is materially different from a shutdown.[3] Yet neither company disclosed the purchase price, cash-versus-stock mix, employee retention, or investor return. The evidence supports one conclusion: Teleo demonstrated enough technology, customers, and strategic fit for Havoc to acquire the company and continue the product.
Key Lessons
-
Ship the useful middle. Teleo made supervised autonomy the product instead of treating it as an embarrassing waypoint. That choice brought the system onto customer sites while full autonomy remained incomplete.
-
Installed assets can be a distribution wedge. Retrofitting mixed fleets avoided forcing customers through an equipment replacement cycle. It also left Teleo responsible for more integration and validation work.
-
Channel partners carry operational knowledge. Equipment dealers brought service capacity and customer trust that a 30-person robotics company could not reproduce quickly. The channel still had to learn a new safety-critical system.
-
Orders are weaker than utilization. Machine orders and named customers established demand. Production hours, intervention rates, uptime, renewals, and measured customer savings would have established repeatability.
-
Strategic fit can end independence without ending the product. Havoc bought Teleo to make ground systems part of a broader control architecture. The company disappeared as a standalone entity while its core capability continued.
Sources
- Y Combinator: Teleo
- Teleo: Announcing the Series A
- Teleo: $16.2 million Series A extensions
- Teleo technology and installation process
- Teleo customer deals and dealer expansion
- Teleo: customer launch and 42 machine orders
- Equipment Journal: Teleo supervised autonomy interview
- F-Prime Capital: The next wave of autonomous vehicles
- Inside Autonomous Vehicles: customer and dealer expansion
- BLS: Construction Equipment Operators
- AGC: 2025 construction workforce shortages
- U.S. Census Bureau: May 2025 construction spending
- Caterpillar: 2025 segment highlights
- Komatsu: FrontRunner Autonomous Haulage System
- Havoc: acquisitions of Mavrik and Teleo
- Vinay Shet: Teleo joins Havoc
- Hacker News: Teleo remote-operation discussion
- Construction Dive: Teleo Series A