Gasgoo Munich- Robotaxi has yet to produce a standalone profit-and-loss statement, but Elon Musk is already discussing how much money it will make next year.
On a third-quarter earnings call held October 22, Beijing time, Musk forecast that the Robotaxi business would start contributing meaningful revenue in the second quarter of 2027. By the third and fourth quarters, it stands to become a significant source of both revenue and profit for the company.
That optimism stands in sharp contrast to Tesla's current financial picture.
In the third quarter, revenue hovered around $28.05 billion, roughly flat year-over-year. Net income attributable to shareholders fell to $787 million, a drop of roughly 43%. The company is juggling operations for its automotive and energy storage businesses while pouring capital into AI, robotics, and autonomous driving.
Robotaxi is indeed gaining traction. Unsupervised driving miles neared 2 million in the third quarter—eight times the total at the end of the previous quarter. The Cybercab, designed specifically for driverless operation, has already begun ferrying passengers in Austin, Texas.
Image source: Tesla
Yet until Tesla discloses standalone operating data, the outside world remains in the dark about how much money the service is actually making.
Elsewhere, Pony.ai has validated the unit economics of Robotaxi in Shenzhen, and Waymo completes hundreds of thousands of paid trips weekly. Still, neither has proven that a Robotaxi business can easily become a profit pillar for a company.
Musk wants to clear that hurdle in about a year.
What advantages does Tesla actually possess to justify such optimism for 2027?
Profits are falling, but spending won't stop
Tesla currently isn't short of funds to invest in Robotaxi.
In the third quarter, operating cash flow reached about $6.26 billion, while free cash flow stood at $1.53 billion. Cash and short-term investments topped $45 billion.
But the income statement is under strain.
Operating income for the quarter slumped 56% year-over-year to $718 million, dragging the operating margin down to just 2.6%.
Automotive gross margins, excluding regulatory credits, rebounded to 18.1% quarter-over-quarter, suggesting existing operations aren't deteriorating across the board. The more visible pressure is coming from expenses.
Operating expenses swelled 39% year-over-year to roughly $4.78 billion in the third quarter, with R&D spending jumping 69% to about $2.76 billion.
AI compute power, chip development, and production readiness for new products are all driving costs higher. CFO Vaibhav Taneja expects operating expenses to keep rising until these new products are commercialized.
The capital outlay isn't over yet.
Capital expenditure came in at roughly $4.73 billion in the third quarter. Although Tesla cut its full-year 2026 capex guidance from over $25 billion to around $20 billion, management still plans to expand spending on the Robotaxi fleet, Optimus production capacity, AI compute, and chip manufacturing in the coming years.
Tesla is adjusting the pace of capital allocation, not ending the investment cycle.
For Robotaxi, this involves a critical financial distinction: buying vehicles and building operational infrastructure burns cash upfront, while those assets subsequently hit profits through depreciation. Even if the service achieves operating profitability, it doesn't mean the initial outlay has been recovered.
Currently, Robotaxi revenue is booked under "services and other," and Tesla has yet to release a standalone P&L for the unit.
Consequently, Musk's 2027 forecast carries two layers of expectation: Robotaxi must not only start generating revenue but also deliver meaningful operating returns as Tesla continues to pour money into new ventures.
To gauge the difficulty of that target, it's worth looking at the companies already leading in commercialization.
How hard is it to make a single Robotaxi profitable?
This March, Pony.ai announced that its seventh-generation Robotaxi achieved monthly unit economics (UE) break-even in Shenzhen.
Image source: Pony.ai
According to the company's disclosures, the calculation includes costs for vehicle and autonomous suite depreciation, charging, maintenance, remote assistance, insurance, and ground operations.
The corresponding operational metrics were: 23 orders per vehicle per day, with daily net revenue of 338 yuan.
The greatest significance of this data set is that it moves Robotaxi commercialization beyond the theoretical argument of "saving on driver wages."
Vehicles must secure a steady stream of orders to cover depreciation and ongoing operational costs.
However, Pony.ai achieved unit economics break-even for a specific model in a specific city—not overall corporate profitability.
WeRide's financial data highlights the distance remaining.
WeRide accelerates overseas Robotaxi commercialization; Image source: WeRide
In the second quarter, its Robotaxi fleet in China averaged over 21 daily orders per vehicle, up roughly 24% quarter-over-quarter. Yet for the first half, the company reported revenue of about 346 million yuan and a loss of roughly 790 million yuan.
These are aggregate financial figures and can't be attributed solely to Robotaxi. Still, the coexistence of improving orders and corporate losses reflects the need for sustained investment in R&D, management, and business expansion.
Making the unit economics work for Robotaxi in one city doesn't guarantee replication in another.
New markets require rebuilding service coverage, order networks, maintenance facilities, and operational teams. Vehicle utilization rates in mature regions may not be immediately replicable elsewhere.
Even Waymo, with its clear lead in operational scale, hasn't publicly proven that the entire business is profitable.
Image source: Waymo
By the end of March, Waymo's cumulative driverless miles exceeded 220 million. By September, it covered 15 U.S. cities with a fleet of roughly 4,000 vehicles, completing weekly paid trips in the hundreds of thousands.
This comparison isn't meant to show how many vehicles Tesla is lagging behind.
What truly matters is that Waymo's years of accumulated operational networks, safety records, and regulatory experience are difficult to replicate in the short term simply by adding production lines.
Robotaxi technology can iterate, and vehicle production can scale, but operational efficiency in a city must be built order by order.
Tesla, however, intends to drastically shorten that process.
Can cheaper cars make Robotaxi profitable faster?
Tesla has reason to be confident.
Unlike autonomous driving companies reliant on external automakers for production, Tesla possesses a complete vehicle manufacturing system.
The Cybercab was designed for driverless operation from the outset, lacking a traditional steering wheel and pedals. Management revealed that its parts count is roughly 50% lower than previous models, and the manufacturing team believes the vehicle is well-positioned for mass production.
Image source: Tesla
At the same time, the Cybercab shares the same autonomous driving software as the Model Y, allowing it to leverage the technology and data accumulated from existing passenger vehicles.
These advantages can lower certain R&D and manufacturing costs, potentially helping Tesla scale its fleet faster.
The problem is, the hardest costs to control in Robotaxi may not come from manufacturing.
On the earnings call, when asked about cost per paid mile, Taneja cited a target of 30 cents. He admitted this hasn't been achieved yet but believes a path forward is emerging as operational scale expands.
Taneja did not detail the full cost structure behind this target.
The real operational challenge for Robotaxi lies in the fact that not every mile driven is a paid mile.
Driving to pickup points, deadhead repositioning, and returning to charging facilities all cost money. Even if the vehicle is cheap to build, insufficient orders and high empty-mile rates will erode profitability.
And removing the driver doesn't mean eliminating manual operations entirely.
Vehicle cleaning, charging maintenance, remote assistance, incident handling, and ground dispatching still require ongoing investment.
These expenses share a characteristic: economies of scale are closely tied to the concentration of operational areas.
In Austin, vehicles can share maintenance facilities, dispatch systems, and ground service resources. If orders rise, the cost allocated per vehicle has room to fall.
But if Tesla expands into multiple cities simultaneously, new road environments, weather conditions, and regulatory requirements will generate new operational burdens.
This means there is no simple correlation between fleet size growth and unit cost reduction.
Tesla can build a Cybercab more cheaply, but it cannot secure more orders per vehicle per day simply by lowering the build cost.
More importantly, Tesla has chosen a relatively asset-heavy business model.
Musk has made it clear that the vast majority of Cybercabs will be owned by Tesla and operated primarily as a ride-hailing service.
Building, owning, and operating the vehicles itself allows Tesla to retain a larger share of revenue, but it also means bearing vehicle depreciation, operational infrastructure, and daily service costs.
A different approach is emerging in the Chinese market.
DiDi is partnering with GAC Aion to develop Robotaxis, leveraging existing platforms for orders. SAIC Mobility is combining with SAIC's manufacturing system to advance custom models. XPENG began expanding internal Robotaxi testing in October, attempting to drive operations through partnerships. Pony.ai and WeRide are also collaborating with automakers, ride-hailing platforms, and local operators.
Partnership models require sharing revenue, but they allow for the sharing of vehicle assets and market expansion costs.
The potential advantage of Tesla's proprietary model becomes clearer when order density is high enough: vehicles run continuously, the operational system matures, and profits that might otherwise be shared with partners can be retained in-house.
The problem is, these conditions take time to establish.
If vehicles roll off the line faster than Tesla can secure operating permits, build city service networks, and grow orders, new Cybercabs will initially add to asset investment rather than profit.
Therefore, the biggest commercial variable for Tesla in 2027 is not whether Cybercab can be mass-produced quickly, but whether new capacity can be absorbed fast enough.
What numbers are missing from Musk's 2027 outlook?
Returning to the earnings report.
Tesla hopes Robotaxi will become a significant contributor to revenue and profit in the second half of 2027, but it provided no specific figures and made no explicit promise that the business would achieve overall net profitability by then.
For a company with quarterly net income of about $787 million, a "significant contribution" must be measured by actual operational results, not just the speed of business growth.
Moreover, that profit baseline itself is shifting.
If profits from existing businesses like automotive continue to decline, Robotaxi could command a higher share of total profits even with a limited absolute contribution.
But a change in share is no substitute for the business's inherent profitability.
What really matters is whether Robotaxi can sustainably increase operating profit—and whether those profits are sufficient to support subsequent expansion.
For now, however, the outside world is missing the most critical data points.
Tesla has released driverless mileage figures but hasn't fully disclosed daily paid orders per vehicle. It set a 30-cent cost target but hasn't published current unit costs or their composition. It plans to expand the fleet but has yet to provide a standalone Robotaxi P&L.
This makes Musk's 2027 forecast difficult to verify financially.
From a manufacturing standpoint, Tesla is positioned to lower vehicle costs. Technologically, software reuse offers a chance to reduce incremental R&D spending. Cash reserves and existing businesses also provide a foundation for fleet expansion.
What remains truly unproven is whether these advantages can translate into stable orders and sustained profits across more cities.
Chinese Robotaxi companies have started answering the question of whether a single vehicle can break even, while Waymo has demonstrated the long-term accumulation required to build a scaled service network.
Tesla has chosen a path that emphasizes the integration of manufacturing, technology, and operations—and has set a tighter timeline for itself.
In our previous "Tesla's Moment of Truth" series, we discussed the challenges of moving Optimus from manufacturing to actual work in "What's Holding Back Tesla's 10,000 Humanoid Robots?," as well as the commercialization test FSD faces entering the Chinese market in "Tesla FSD: How Much Longer Must Chinese Drivers Wait?."
With Robotaxi, the test is even more pragmatic.
Musk wants it to become a profit source for Tesla as soon as possible, but building a driverless car and successfully running a driverless taxi service are, ultimately, two different things.
In 2027, what Tesla truly needs to prove is not how many Cybercabs it can put on the road, but whether it can keep those cars busy with orders and turn that revenue into real profit.
