Mecka AI's Sequoia Talks Value It 7x Higher Than a Round Closed Three Months Ago

Mecka AI, a two-year-old startup that pays gig workers to record themselves doing everyday chores for robot-training datasets, is in talks with Sequoia Capital that would value the company near $500 million, according to TechCrunch. The terms are still being negotiated and the exact check size hasn't been disclosed, but the number itself is the story: it would mark roughly a sevenfold jump from the valuation implied by a $60 million raise the company disclosed in June, itself only three months earlier.

That June disclosure was unusual in its own right. Mecka revealed it had already closed two previously unannounced rounds, a $25 million Series A in November 2025 and a $35 million follow-on, led by Framework Ventures, a crypto-focused fund rather than a robotics specialist, alongside Menlo Ventures, SV Angel, Kindred Ventures, and angel investor Ted Xiao, a former DeepMind researcher now on Jeff Bezos's Project Prometheus team. At the time, Mecka told reporters it expected to end 2026 at a $100 million annualized run rate based on signed contracts, though it declined to name customers or disclose a valuation.

Mecka isn't the only company being repriced this fast. XDOF, a UC Berkeley spinout founded by Philipp Wu and Fred Shentu that emerged from stealth with a $70 million Series A in June, is now in talks for a Series B near $1.2 billion, an even sharper jump that TechCrunch reported investors pursued unsolicited. XDOF says it has around $50 million in annualized revenue and 20 customers, including several frontier AI labs, and it collects data through teleoperation, human operators piloting robot arms, a different bet from Mecka's approach of outfitting gig workers with body sensors and iPhones to capture human motion directly.

The financing pattern echoes what happened in text-based AI data, where Scale AI, Mercor, Surge AI, and Micro1 built businesses selling labeled data to model developers. That market fractured after Meta took a 49% stake in Scale AI for roughly $14 billion in June 2025, prompting OpenAI, Google, and xAI to route work elsewhere; Surge AI is reportedly now in talks near a $25 billion valuation and Mercor near $20 billion off more than $2 billion in annualized revenue. Investors appear to be trying to fund robotics-data equivalents before any single vendor locks up the same position Scale once held in language models.

Sequoia's interest also fits a specific mandate rather than a general enthusiasm for robotics. The firm unveiled a $10 billion commitment in early August aimed at what it calls "reindustrialization," spanning manufacturing, defense, energy, and robotics, and has already backed Physical Intelligence, a robotics foundation-model company. A data-supply company like Mecka would occupy a distinct layer of that thesis from the model and hardware bets Sequoia has already made, supplying the raw motion data those models are trained on rather than the models or the robots themselves.

What's largely absent from the Mecka valuation story is any public accounting of how that motion data gets collected and from whom. MIT Technology Review reported in February and again in April that gig workers across more than 50 countries, including India, Nigeria, and Argentina, have recorded themselves with head-mounted cameras for humanoid-robot training, in some cases for hours of repetitive tasks, and that workers aren't always told the footage will train systems that could eventually replace similar jobs. India's IT Ministry has reportedly raised questions about the practice. None of that reporting names Mecka specifically, and the company's own consent and pay practices for its data workforce remain undocumented. Mecka's August 2025 seed announcement, when it raised $8 million from NEO and SV Angel, named 1X Technologies as an early customer building on egocentric human-video datasets; the September funding coverage doesn't say whether that relationship still holds, or who else is now paying for the data a fivefold valuation increase is being built on.