Einride announced a commitment for 500 Tesla Semis on August 18 — the largest public Semi order anyone has made, overtaking WattEV's 370 units from May. It is a genuinely important deal, and the arithmetic around it is more interesting than the headline.
Tesla's dedicated Semi plant next to Gigafactory Nevada has a designed capacity of 50,000 units a year. That is about 137 trucks a day. The biggest order in the programme's history is under four days of nameplate output.
Both halves of that sentence matter, and most coverage picks one.
The Einride deal
| Term | Detail |
|---|---|
| Units | 500 Tesla Semis |
| Start | September 2026, phased over 24 months |
| Geography | California, Texas, New Jersey, Illinois, Georgia |
| End customer | Amazon, plus other Einride shippers |
| Operating layer | Einride's Saga AI platform — routing, charging and energy-cost optimisation |
| Financing | Fully third-party financed |
| Fleet effect | Roughly triples Einride's deployed electric fleet, ~250 → ~750 |
| Commercial frame | Supports converting about $800M of potential long-term annual recurring revenue into active freight |
Two details do the most work. The end customer is Amazon, which means this is not a technology pilot — it is a shipper with hard service levels putting a battery-electric tractor on real lanes in five states. And it is fully third-party financed, which removes Einride's balance sheet as the constraint and replaces it with the asset-finance market's willingness to underwrite Semis. That is a better trade for Tesla and a different risk: fleet finance reprices faster than freight demand does.
The whole public order book
Tesla has never published a Semi order book, so this is assembled from what customers have announced. Treat it as a floor on real demand, not a measurement of it:
| Customer / operator | Units | Timeline | Status |
|---|---|---|---|
| Einride | 500 | Starts Sep 2026, phased over 24 months | Newest and largest; third-party financed |
| WattEV | 370 | First ~50 in 2026, balance by end 2027 | Port of Oakland focus |
| Walmart | ~130 | Various reports | Partial / testing |
| UPS | 125 | 2017 reservation | No public update |
| PepsiCo | 100 | Longest-running customer | ~50 delivered and in service |
| King Fio / NICA / QX Logistix | ~60 | Mostly 2027 starts | California ports |
| Sysco | 50 | 2017 deposit | Soft — no hard obligation |
| Big F Transport | 40 | Early 2027, via Forum Mobility | Port drayage |
| Anheuser-Busch | 40 | 2017 era | On order |
| Others — DHL, ArcBest, Hight Logistics, small fleets, Tesla internal | ~100 | Mixed | Pilots and small fleets |
| Total announced | ~1,515 |
The first chart splits that book three ways, and the split is the point:
- ~50 delivered — PepsiCo, the only customer with meaningful trucks actually working.
- ~1,080 recent commitments — 2025–2026 announcements with named start dates. This is the real book.
- ~385 legacy reservations — 2017–2018 deposits from UPS, Walmart, Sysco and Anheuser-Busch, taken against a truck that arrived eight and a half years later. Some will convert. None of them oblige anyone to do anything.
The number that reframes the story
Set the whole book against the factory:
| Units | As months of output | |
|---|---|---|
| Entire announced public book | ~1,515 | — |
| 2026 production, our base case | ~4,500 | book = ~4 months |
| 2027 production, our base case | ~15,000 | book = ~5 weeks |
| Nevada nameplate | 50,000 | book = ~11 days |
Every publicly announced Tesla Semi order, added together and including the soft 2017 deposits, is about four months of this year's likely build and eleven days of the factory's design rate. That is the second chart, and it is not a small discrepancy.
This is where the popular framing — "demand is way ahead of production" — needs care. It is true for 2026, and it is the correct read of the near term: Einride's phased deliveries, WattEV's first ~50, PepsiCo's remaining units, the port fleets and Tesla's internal use can absorb everything Nevada builds this year, and charging readiness will gate deliveries before demand does. It is not yet true for the factory Tesla actually built. A 50,000-unit line needs an order book roughly thirty times the one that has been announced.
Three honest qualifiers, because the absence of announcements is not the absence of orders:
- Semi orders are announced by customers, not by Tesla. A fleet that does not put out a press release does not appear here at all.
- Tesla does not break out Semi in production or deliveries, so nobody outside the company can size the book or the backlog from filings.
- The programme is one quarter into volume production. Fleets buy trucks after they have seen trucks work, and the largest single order arrived four months after the line started — which is the pattern you would want.
The fair conclusion is narrower than either camp's: demand is comfortably ahead of 2026 supply and nowhere near 2027–2028 capacity, and the gap between those two statements is what the next four quarters are actually about.
What gates the ramp
- 4680 cell supply, shared with other programmes — the constraint Tesla itself names most often.
- Megachargers. A Semi with no corridor charging is a yard tractor. Einride's five-state footprint and WattEV's port focus are both bets on charging arriving on schedule; this, not truck output, is the most likely source of delivery slippage.
- Customer operational readiness — depots, power upgrades, driver training. This is why phased 24-month rollouts exist.
- Autonomy. FSD for Semi is expected to start working late 2026 into 2027, timed with the scale-up. It changes the economics of the purchase, not the delivery date.
- Europe entry in 2027, which is where the addressable book could step up rather than compound.
Production and what it is worth
| Low | Base | High | |
|---|---|---|---|
| 2026 production | 3,000 | 4,500 | 8,000 |
| 2027 production | 10,000 | 15,000 | 25,000 |
| 2027 revenue at ~$180k a truck | $1.8B | $2.7B | $4.5B |
| 2027 revenue as % of TTM revenue | 1.7% | 2.6% | 4.3% |
The price per truck is our assumption, not a Tesla figure — the company has not published current Semi pricing, and configuration spreads are wide. Trailing-twelve-month revenue is $103,619M.
Two things follow. First, Semi is the first of Tesla's new programmes that gets to a visible revenue line quickly: our 2027 base case is worth more than a full point of revenue growth, which is more than the Cybercab does in the same window — we sized that one at six built for every one carrying a passenger by year-end. Second, at nameplate the arithmetic is genuinely large: 50,000 trucks at $180k is $9B a year, about 8.7% of current revenue, from a factory that already exists.
Which is the same shape we described in the factory piece this morning: the building is finished and at stage 5, and everything left is the ramp. Semi is now the cleanest public scoreboard for whether Tesla's ramp skill has caught up with its construction skill — the plant took roughly 28 months to build, and the truck took eight and a half years from unveiling to volume.
What to watch
- Einride's September phase-one units actually moving freight — the first checkable milestone, five weeks out.
- Megacharger site count on the five Einride corridors. Trucks without charging do not deliver.
- Whether Tesla breaks Semi out in any quarterly disclosure. It has not, and until it does the delivery number is unknowable.
- A second order at Einride's scale. One 500-truck deal is validation; two make it a book.
