I don’t read gig-economy press releases for the mission statements. I read them for labor cost and who is still holding the wheel. Uber, Lyft, and DoorDash all sell “flexibility.” Underneath that, each quarter is the same fight: keep the app full of workers without becoming an employer under old labor law, and buy enough automation that drivers stop being the whole P&L.

The California rules that still matter

Prop 22 kept app-based drivers in contractor status with a different benefits package than employees. That wasn’t the end of the story. AB 1340 gave covered rideshare drivers a path to organize without flipping them to W-2. PERB certified the California Gig Workers Union on September 9, 2026. Certification gives the union a bargaining role; it is not yet a contract. SB 371 lowered one uninsured-motorist coverage requirement, not the entire rideshare insurance bill.

The legal fight is no longer “employee vs contractor” as a single switch. It is now also about what a statewide bargaining representative can win and whether the terms survive platform changes.

Earnings week is the real product review

Q3 2025 showed strong demand and three different cost questions. Uber grew bookings faster than operating income; its filing doesn’t pin the gap on AV or EV spending. DoorDash earned $244 million and said it expected to invest several hundred million dollars more in 2026 than in 2025. Lyft crossed $1 billion in trailing twelve-month free cash flow while insurance costs remained worth watching.

Same industry, three investor stories: growth with compressed margins, profit that isn’t “enough,” and cash generation as a survival metric. Uber’s Q4 2025 print keeps that pattern alive — beat the easy numbers, still answer for where the money goes.

How they try to shrink driver cost without saying “fire drivers”

Uber’s EV grants look green on the surface. I treat them as unit-cost management: cheaper fuel for drivers, better utilization, data on who will take subsidies. Same company is testing AI data-labeling gigs so a driver between trips can label frames instead of sitting dead. That’s not a side quest. That’s inventing work to keep the labor pool attached to the platform.

DoorDash already won the US delivery share fight that started when they ate Grubhub’s lunch and ended with Wonder buying the carcass cheap. Now the spend is autonomous delivery and international. Labor is still the variable expense; robots are the option on that expense.

Lyft is the smaller survivor. Board drama, rental experiments that died, a long string of “we’re still here” earnings. Cash flow matters more for them because they can’t outspend Uber on every AV partnership.

Robotaxis are the labor hedge with better branding

I keep a separate map of what robotaxi product actually ships. Short version for the gig reader: Waymo sells driverless rides across multiple cities; Tesla reports unsupervised rides outside California but still operates there under a testing-with-driver permit; Uber and Lyft buy access to other companies’ autonomous fleets. Those products do not have the same labor cost or regulatory status.

When Waymo pays Dashers to close doors, the punchline is uncomfortable for both narratives. Autonomy didn’t remove labor. It sliced labor into an $11 task. Gig platforms become the glue between robots and the last dumb physical problem.

What I watch next

  • Does the certified union reach an agreement that changes driver pay or working conditions?
  • Do Uber/Lyft AV miles show up as a meaningful share of trips, or just PR density?
  • Does DoorDash’s “spend again” strategy look smart in 2027 or like 2021 redux?

I write the dated posts when those numbers move. This page is the spine those posts hang on — not a complete history of the gig economy, just the through-line that keeps making me open 10-Q filings.