Tesla’s Master Plan Part IV: Don’t Say ‘Sustainable Energy’, but say ‘Sustainable Abundance’

Tesla has a new master plan out, Master Plan Part IV, which moves from sustainable energy to sustainable abundance. The document recasts Tesla as an AI-first manufacturer that wants to lower the cost of energy, mobility, and labor in one sweep. You read it well, the next phase is centered around robots, autonomy, and grid storage.

The pillar that will have the biggest impact in the future is labor. Tesla puts its humanoid robot Optimus on the shop floor to take over repetitive, ergonomically tough tasks. The second is mobility. Autonomy remains the growth engine, with Full Self-Driving feeding a future robotaxi network designed for high utilization and low cost per kilometer. The third is energy. Powerwall-based virtual power plants and Megapack deployments turn distributed batteries into paid grid resources, replacing fossil peakers and smoothing demand.

Tesla being at the forefront of sustainable innovation means this is a document every sustainable specialist should have read.

What does sustainable abundance actually mean?

As explained in the intro, Tesla shifts from “sustainable energy” to “sustainable abundance”. With this move it is aiming to drive down the cost of energy, mobility, and labor with one shared AI and manufacturing stack. The company centers three products: Optimus on factory floors, Full Self-Driving feeding a future robotaxi service, and distributed batteries aggregated into virtual power plants (VPPs).

The idea behind it is to scale economics. With every new deployment it cycles data back into cheaper, faster deployments.

Pillar 1 – Labor: Optimus, the factory’s first strategy

Tesla will train a general-purpose humanoid, which start on repetitive, ergonomically harsh plant tasks such as kitting, bin-picking, intralogistics, etc.. Once that is executed, and only then, it will expand outward.

To measure the progress, Tesla will have to check the completion of named workflows without human shadowing. And to check the mean time between failures per task. That should result in unit cost trending down while throughput trends up.

In short, you get ROI when a robot repeats a defined task at a rated cycle time with stable quality. You lose it when environments change weekly. They should realize faster wins in parts kitting and end-of-line logistics than in delicate assembly.

What do rivals do?

• Hyundai | Boston Dynamics focuses on proven, task-specific platforms (Spot, Stretch) and exoskeletons for logistics and inspection. It shows a narrow scope, with real deployments.
• Toyota | Woven builds the software and test city (Arene, Woven City). It opted for teleoperation and safety-led robotics to trump humanoid generality for now.

Pillar 2 – Mobility: Autonomy, from assisted driving to robotaxi

Tesla goes for scaling Full Self-Driving to a robotaxi network with high utilization and low cost per kilometer, coordinated through a dedicated app.

To measure progress, they will have to augment driver-out kilometers per city, and not just beta miles. Collision and disengagement rates should be reported under the same definitions. The end-user economics are the cost per km and fleet utilization.

For this the company will ignore sign-ups and instead track permits, insurance frameworks, and hours of driver-out operation during peak demand windows.

What do rivals do?

• Waymo runs paid, driverless rides in multiple U.S. metros. The model is geofenced but commercial today.
• Mercedes sells SAE Level 3 DRIVE PILOT in select regions (eyes-off under conditions). That is regulatory certification Tesla does not claim for FSD.
• BMW scales hands-off highway assistants under UN rules; with a conservative scope, and a broad availability.
• Chinese AV stack (Baidu, Pony.ai, XPeng) expands pilots and city permits, pairing autonomy with native mapping and superapp integration.

Pillar 3 – Energy: Virtual Power Plants, from homes to the grid

Tesla will aggregate Powerwalls and deploy Megapacks so software can shave peaks, stabilize grids, and pay owners for dispatch.

To measure progress Tesla will have to check the contracted megawatts under VPP programs, the owner payouts per season or per kWh dispatched, and check the peaker avoidance metrics and grid-services revenue booked.

What do rivals do?

• GM Energy | Utilities test EV-to-home and VPP pilots with Ultium hardware.
• Renault Mobilize runs vehicle-to-grid car-sharing in Europe, selling power back during peaks.
• BYD plays at global scale in utility batteries, pressing prices and delivery timelines.

Tesla’s cheaper charging network keeps its customers hooked

Tesla already has a mature, reliable Supercharger network and NACS adoption across brands. It’s rivals are not asleep though, they set up IONNA, a joint venture of legacy automakers. IONNA is now rolling out 30,000+ high-power stalls in North America, supporting NACS and CCS. Stellantis’ Atlante from its expands fast/ultrafast coverage in Southern Europe. The gap clearly narrows as third-party uptime improves and roaming gets simpler.

What about the price between Tesla Superchargers and IONNA and Atlante?

You can have a big network of chargers, but the customers usually look at the price first. So let’s see what the different networks offer.

The Tesla Supercharger (US/EU) has a pricing that varies by site and time (peak/off-peak). You can see the live kWh price in the Tesla app, and also counts for non-Tesla drivers who can buy a Supercharging Membership to get “Tesla-owner” rates (US $12.99/mo; Europe commonly advertised around €9.99/£8.99/mo where available). The off-peak pricing is around US $-€0.24/kWh across the board. Important, Tesla also applies congestion (idle) fees if you stay plugged in after the threshold.

IONNA (US/Canada only) has a public pay-as-you-go pricing shown on the charger/apps. Early sites launched at about $0.48/kWh; and they’ve run promos like $0.25/kWh (Labor Day). Many sites are in the mid-$0.30s/kWh though. You can pay with card or partner apps (BMW, Mercedes, Kia, GM, Hyundai, ChargePoint, etc.).

The Atlante pricing is based on either a flat-rate option (€3.99/month and a flat €0.43/kWh on fast and ultrafast chargers) or a standard (pay-as-you-go) formula. In practice, you’ll see typical EU public fast-charging ranges between ~€0.35–€0.60/kWh for 50–150 kW and ~€0.40–€0.90/kWh for 150 kW+.

In short, in Europe and the US, Tesla beats all competition with its off-peak pricing.

Humanoid robots market

The humanoid robots market is much bigger than you would think, and when it comes to pure logistics or reptitive production, no job will be safe from these AI powered robots. And there are regional dynamics.

In China we see a rapid scale in robotics broadly (install base, patents) with cost pressure from local vendors; government programs push domestic service and industrial robots, with humanoids showcased at the 2025 World Robot Conference.

In the United States/EU the demand concentrates in auto, third-party logistics, and big-box retail DCs. Here, regulators scrutinize safety and “driver-out/worker-out” zones. While BMW (US) and Mercedes (DE/HU) illustrate European OEM interest; US retailers test in controlled environments.

The jobs that will get automated first include the following:

• Structured, repeatable tasks in brownfield plants and warehouses: part placement into fixtures, tote induction, pallet moves, pick-to-cart, line-side replenishment. These win because facilities can fence hazards, instrument the flow, and measure cycle time and mean-time-between-failure (MTBF). BMW’s trials explicitly targeted fixture placement; Amazon’s Digit pilots target repetitive flow between zones.
• Customer-facing or outdoor tasks (storefront service, curbside delivery) stay experimental until manipulation, perception, and liability frameworks mature. Even Amazon’s “humanoid delivery” exploration sits in testbed mode.

And there is a huge market open for these machines:

• 2030–2035 window: Goldman Sachs lifts its 2035 humanoid TAM to $38B (≈1.4M units), assuming material-cost deflation and faster iteration. UBS pegs $30–$50B by 2035 with ~2M units in the workforce.
• Long arc to 2050: Morgan Stanley sketches a much larger composite opportunity (hardware, software, service) crossing $5T by 2050, with the majority in industrial and commercial use.

As mentioned, pilot projects are already well underway. Here are three of them in specific sectors:

• Automotive manufacturing: The pilots focus on kitting, bin-picking, machine tending, and end-of-line logistics. BMW has a commercial agreement with Figure and ran trials at Spartanburg for sheet-metal placement; Mercedes invested in Apptronik and is testing humanoids for component moves and quality checks.
• Warehousing and contract logistics: Agility Robotics’ Digit is being trialed at Amazon and GXO for tote handling and case movement. This represents low-variation workflows with clear safety and uptime metrics.
• Retail back-of-house. Sanctuary AI ran paid, small-scope tasks with Canadian Tire stores and DCs (stock moves, packing, basic inventory interactions).

The market for humanoid robots is real but staged with paid pilots in factories and warehouses for now. The broader deployment will come as vendors prove cycle time, uptime, and safe integration.

Sustainable Abundance, but only when deployments prove it

Tesla’s new plan says the company isn’t just about clean cars anymore. It wants to cut the cost of three things at once: work, transport, and electricity. The way to do that, according to Tesla, is one shared AI-and-manufacturing stack that runs across humanoid robots (Optimus), autonomous driving and robotaxi services, and home-to-grid energy storage.

What’s real today sits mostly in energy and driver-assist. Powerwalls and Megapacks already earn money in virtual power plants, and FSD subscriptions keep rolling. Inside Tesla’s own factories, Optimus is learning repetitive tasks, but full, unsupervised workflows are still ahead. Robotaxis remain a goal; the proof will be city-scale, driver-out operations with transparent safety data.

Their bet is clearly on integration. Cars generate data and cash to fund robotics; robots lower factory costs; software-coordinated batteries stabilize grids and create new revenue. If it works, the outcome is lower cost per delivered kilometer, lower cost per repetitive task (and a massive job loss a result of a successful implementation), and lower cost per dispatched kilowatt-hour.

---

---