Opendoor founder Eric Wu launches NavigateAI, a hands-free AI copilot for construction crews

Eric Wu, who founded and ran the real estate startup Opendoor before stepping away in 2022, is returning to company building with a venture aimed at construction's labor crunch. According to TechCrunch, the startup, called NavigateAI, emerged from stealth in late May with $25 million in seed funding at a $225 million post-money valuation, led by investor Elad Gil.

The cap table maps closely onto Wu's history. Khosla Ventures, where Opendoor co-founder Keith Rabois is a partner, joined the round alongside Fifth Wall, homebuilder Lennar, real estate firm Tishman Speyer and electrical contractor Helix Electric. Angel investors include DoorDash's Tony Xu, Instacart's Apoorva Mehta and Coinbase CEO Brian Armstrong. TechCrunch notes that Gil had previously backed Opendoor.

Wu told TechCrunch he was drawn back because he views AI as the defining technology platform of his lifetime: "I knew if I looked back in 10 years and didn't do something related to it, I'd probably regret that."

A hands-free coach for field workers

NavigateAI is building AI copilots for construction workers and other field laborers. The product runs on smartphones and, in hands-free mode, through Meta's AI glasses. A worker points a camera at the work in front of them and asks, in plain language, whether something is installed correctly, whether the torque is right or whether it meets code. TechCrunch reports the system can surface building specifications, manufacturer manuals and company policy in real time.

Hands-free use matters, Wu argues, since nobody wants to consult a phone mid-job, and NavigateAI is reportedly working with Meta to get the glasses certified as safety eyewear for protected environments. The company is also partnering with AIM, a Meta-backed trade school for fiber installation that guarantees graduates job placement, creating a channel to reach workers before they ever arrive on a site.

Adoption splits along generational lines, according to Wu: younger workers embrace the product, while 30-year journeymen trust their own instincts and are not eager to wear a computer on their face.

A shortage the AI boom made worse

The market case rests on stark numbers. The trade group Associated Builders and Contractors has said roughly 349,000 additional workers are needed this year in the United States just to keep pace with demand, a gap expected to widen as the workforce ages, immigration enforcement removes workers from the country, and megaprojects multiply.

Data centers are a big part of that pressure. TechCrunch reports that a large data center campus once needed about 750 workers at peak, while the biggest projects now require 4,000 to 5,000. Meta's Hyperion campus in Richland Parish, Louisiana reportedly needs around 5,000 construction workers, and OpenAI's Stargate site in Abilene, Texas has involved 6,400. Staffing firm Kelly has said 90% of data center operators now cite staffing shortages as a critical constraint on building or expanding.

Pricing tied to outcomes

NavigateAI started with token-plus-margin pricing, similar to usage-based SaaS, but has shifted newer contracts to a share of the value it creates. If the software helps cut the all-in cost of a home from $300,000 to $280,000, for instance, the company captures roughly 20% of the $20,000 saved. Wu points to Lennar, which he says spends about $9 billion a year on labor, installation and construction, meaning a 5% to 10% improvement would represent hundreds of millions of dollars in value.

Known risks on the road ahead

The model has obvious friction points. Wu openly acknowledges an attribution dilemma: proving a home was built cheaper because of NavigateAI, rather than good weather, a skilled crew or available materials, requires A/B testing across divisions and remains approximate. Savings disputes could get complicated even when the client is also an investor.

Safety liability is unresolved. If the software clears a connection that later fails, it is not yet clear who bears responsibility in a notoriously litigious corner of construction law. And competition looms: the most common alternative today, per Wu, is a worker searching Google or asking ChatGPT, but the large model vendors have the underlying capabilities, and Meta also controls hardware distribution.

NavigateAI's defense, as TechCrunch describes it, rests on workflow integrations and proprietary data that take years to accumulate. The longer-term bet is that data itself: every job completed with the product generates labeled egocentric video of workers building things correctly and incorrectly, a dataset Wu believes could eventually be worth as much to robotics companies as the software is to construction clients.

For now, Wu is keeping governance minimal. The company has no board, and he told TechCrunch he will "try to go as long as I can without one," a stance shaped by his years as a public-company CEO after Opendoor's late-2020 SPAC listing.

Why it matters

NavigateAI is a high-profile test of whether AI's economic value can move beyond desk work into skilled physical trades, where the labor shortfall is most acute. Ironically, the AI industry's own data center buildout is deepening the very staffing shortage the startup targets. The outcome-share pricing model and the plan to bank labeled field-work video as a future robotics asset are both experiments other founders will watch closely — as will the question of whether a seed-stage company can hold off the platform giants long enough to make the data moat real.