Canada Lists Roberts Bank Terminal 2 National: 2.4m TEU, 2035, $5.9bn

The Analysis

Put the map on the table first. Ten years ago the west coast of North America was read as one lane with three doors: Los Angeles and Long Beach, Oakland, and the Pacific Northwest pair of Seattle and Tacoma with Vancouver attached as the Canadian annex. That reading is finished.

On 30 October Canada listed Roberts Bank Terminal 2 at the Port of Vancouver as a project of national interest. The Vancouver Fraser Port Authority project is 5.9 billion Canadian dollars for three container berths, adding 2.4 million TEU of annual capacity, lifting Vancouver container capacity by about half and supporting up to 100 billion Canadian dollars of annual trade. The listing gives upfront certainty on the remaining federal authorisations and pulls commercial operations forward by up to four years, to 2035 from 2039. A final investment decision is due by mid-2027.

Four years pulled forward is the headline. What I want to know is which bone on the map actually moved, and whether this is a pulse or a turn.

Three numbers carry the story: 2.4 million containers of annual capacity, 5.9 billion Canadian dollars, and the year 2035. Everything else is commentary on those three. They get more useful if you push on them. If 2.4 million is about half again as much as Vancouver handles today, the port's existing container capacity sits somewhere near 4.8 million TEU and the finished port would stand near 7.2 million. Divide the money by the capacity and you get roughly 2,458 Canadian dollars of capital for every container of annual throughput, which is a heavy figure for a greenfield terminal and tells you this is a structure built for a forty-year life rather than a quick bolt-on.

Divide the trade figure by the capacity and you get about 41,700 Canadian dollars of trade value per container, on the assumption that the 100 billion is port-wide annual trade and not this terminal's own book. Neither division appears in the announcement. Both are the kind of arithmetic that tells you what sort of asset this is.

Why 2026. Canada has spent two years being reminded that access to its largest trading partner is not a guarantee, and a port is the physical expression of that anxiety. Building west coast capacity that reaches the Canadian market without crossing an international border is not a transport project in the old sense; it is insurance written in concrete. The mechanism is new in spirit as well: a national interest listing that squares away the remaining federal authorisations up front, rather than letting a project chew through approvals for another decade. The timing has a narrower edge too. The final investment decision is due by mid-2027, which means the next eight months decide whether this ever appears on the map at all.

Who feels it, and when. Not the shipper booking a box next spring, because nothing about 2035 changes a 2027 routing decision. Carriers and terminal operators feel it first, in network planning in the early 2030s. The railways feel it next, because 2.4 million containers a year is a rail problem long before it is a berth problem. Canadian importers feel it in the 2030s as optionality. American west coast ports feel it as share, slowly, over a decade. Shippers with Canadian inland destinations feel it last and least. Which is another way of saying this: the news matters to you mainly as information about where not to put your 2027 and 2028 volume.

Here is what the coverage skips. The listing buys permission, not steel, and not track. The four years saved come off the authorisation path. The building still has to happen, and from a mid-2027 investment decision to commercial operations in 2035 is about eight years of construction on a marine site with a long causeway and a rail connection that has to be built or doubled alongside it. So the capacity arrives as a single step in 2035, and the nine years between now and then get nothing at all. If you are planning a network for 2028, 2030 or 2032, this project is not in your plan. That is the gap nobody writes about, and it is eight years wide.

Then there is the second half of the same point. Adding berth capacity at Roberts Bank does not add inland capacity. Containers leaving Vancouver cross a causeway and then climb through the Rockies on a corridor that already runs tight in winter and has pinch points that no quantity of berth length will fix. If that corridor is not widened on the same timetable, 2.4 million containers of berth capacity arrives into 2.4 million containers of rail constraint, and the queue simply moves inland, from the anchorage to the yard. Berths are the visible bone. The corridor is the one that actually carries the load. On this coast I have watched the same film before: a gateway adds quay length, dwell improves for a season, and then the constraint reappears forty kilometres inland where nobody was measuring.

Put the network around it, because a terminal means nothing on its own. Asia to North America moves by three broad paths: all-water through Panama to the east and Gulf coasts, landbridge through the Pacific Northwest with rail onward, and direct calls at the big southern California gateway. Vancouver competes in the second of those, against Prince Rupert, against Seattle and Tacoma, and against the American intermodal leg out of Los Angeles. It competes on how short the corridor is into the middle of the continent and on how well that corridor behaves in February. Three berths and 2.4 million containers of capacity are a bet that the Pacific Northwest path keeps taking share from the all-water path, and it is a bet placed nine years before it pays off.

The last time this map moved on a comparable scale was the Panama Canal expansion in 2016, which let larger ships reach the east coast directly and forced every west coast gateway to justify its rail leg. The west coast kept its share, but the justification changed: it stopped being about geography and became about how many days the intermodal leg actually takes. That is the same test Roberts Bank will face in 2035. A gateway does not win because it built berths; it wins because the whole door, from quay to inland ramp, behaves predictably. Everything in this announcement sits on the quay side of that sentence and nothing in it sits on the ramp side.

Prince Rupert is the other half of the story and it rarely gets into the coverage. That port has spare capacity now, a shorter run into the American midwest than Vancouver has, and a reason to price sharply against a rival that cannot add a berth until the next decade. For a shipper, a Canadian gateway with two doors competing is worth more than one large door, and that competition starts today rather than in 2035. If you take one thing from this into a rate conversation, take that one: the threat of capacity nine years out is already a lever on the capacity that exists now, and you can pull it this quarter.

A listing is also a political object, and that cuts both ways. On the helpful side, the authorisations are settled up front and the four years are real. On the other side, a project lifted onto a national interest list becomes visible, and visible projects attract opponents who had no reason to organise while it was a line in a port authority budget. Environmental review, the fisheries and habitat questions around Roberts Bank, and the standing argument about who pays for the rail works all now have a spotlight on them. Nine years is a long time for a spotlight. This is why the mid-2027 investment decision matters more than the listing: a decision is a commitment, a listing is a process.

Is this a pulse or a turn. A pulse is a local capacity addition that the network absorbs and forgets, and new berths at an existing gateway are usually exactly that. A turn is when the map reorganises, and I have seen one of these on this coast already. Prince Rupert was a pulp and grain port until a container terminal opened there in the second half of the 2000s, with a rail corridor built to feed it. Within a decade the Pacific Northwest had a new door and Vancouver stopped being the only Canadian answer. That was a turn, and it happened because the berths and the corridor arrived together.

Roberts Bank Terminal 2 will be a turn on the same terms and a pulse if they do not. The condition is specific and testable, which is the useful part: watch whether the rail corridor and the inland intermodal yards get the same treatment the terminal just received. If they do, Canada is building a second door with its own plumbing, which is a decade-scale change in how North America is served from Asia. If they do not, this is a very expensive extra quay.

What is genuinely structural here is the intent rather than the concrete. For thirty years the working assumption in North American gateway planning was that Canadian cargo could move through American ports and American rail and the other way round, because the border was a formality. Once that assumption is priced as a risk, the network starts to want redundancy: two doors per country, each with its own corridor, so that a policy shock on one side does not close the continent. That is a different map from the one we drew in 2015. It is also expensive, slow, and it produces capacity that looks redundant in calm years. Every port authority in the world is currently making the same bet, and some of that capacity will end up underused. That is what it costs for a map to change.

Do the arithmetic on it, with the assumptions stated. Assume you are a mid-size importer bringing 600 containers a year into western Canada, fifty a month, railed inland. Assume the new berths take average berth waiting at Vancouver down by one day once they are running. Assume one day of a container's time, counting demurrage, storage and the knock-on to your delivery promise, is worth 150 Canadian dollars. That is 90,000 Canadian dollars a year, and it starts in 2035.

Discount it at 6 per cent over nine years and the present value of that first year's saving is about 53,000 Canadian dollars; treat the saving as running forever and the whole stream is worth roughly 890,000 in present-day terms. Now set that beside something you can do in 2027. Shift 30 per cent of the volume, 180 containers, to Prince Rupert or to a United States west coast port with a cross-border move, and assume it saves 200 Canadian dollars a container against your present corridor.

That is 36,000 Canadian dollars a year starting next year, with no discounting and no construction risk, and across the nine years to 2035 it puts more than 320,000 Canadian dollars in your pocket before the first container is worked at Roberts Bank. The lesson is dull and true: distant capacity is worth less than near-term routing, even when the distant capacity is much bigger.

There is a second calculation worth doing, on the capital side. At 2,458 Canadian dollars of capital per container of annual capacity, and at a rough assumption of eight to ten container lifts per berth slot per year, the project is betting on throughput per berth around 800,000 TEU, which is at the aggressive end for a three-berth layout. I am inferring that, not quoting it. If the terminal lands nearer the 600,000 a berth figure that operators tend to plan against, the 2.4 million becomes an ultimate capacity reached years after 2035 rather than a day-one number. Ask the port authority which of the two it is: opening-day capacity or design capacity. The answer changes what you put in your 2035 model by a factor of about a third.

What would make me wrong. If the United States and Canada settle their trade relationship and the border stops being priced as a risk, the sovereignty argument weakens and a mid-2027 investment decision becomes easier to slip, and a project that has already spent years in environmental and legal process can always spend one more. If the corridor is treated as a separate and later problem, the berth capacity lands into a bottleneck and the effective gain is a fraction of 2.4 million. If Asia to North America growth disappoints through the 2030s, the new capacity arrives into a soft market and becomes a bargaining chip against Prince Rupert rather than new national capability, in which case your freight rates may gain more from the competition between two Canadian gateways than from the capacity itself. And there is one condition that would flip this into something bigger than I have described: a second national interest listing, for the rail corridor and the inland yards. If that happens, I will rewrite this piece as a turn.

What this means for a routing decision you make this quarter. Nothing about 2035 should enter a contract you sign in 2027. Plan on the assumption that Vancouver container berth capacity is flat until the middle of the next decade, because that is what flat means when the step arrives in 2035. Treat Prince Rupert as the live alternative today, not as a future option, since its spare capacity exists now and the Roberts Bank capacity does not. Ask your rail provider for a written service commitment for the winter months rather than an annual average transit time, because winter is when the corridor tightens and averages hide exactly that. Keep at least one United States west coast routing alive with a cross-border plan, because redundancy is what this whole policy is really buying and you can buy your own slice of it for the cost of qualifying a second door.

The mistakes I expect people to make with this news. Treating a listing as a commitment, when it is a decision about process and not a cheque. Putting 2035 capacity into a 2029 network model. Assuming the four years saved come off the end of construction rather than off the front of the approvals. Forgetting that Prince Rupert exists and that its spare capacity is available this year. Booking everything through Vancouver on the strength of news about a terminal that does not open for nine years. And the quiet one: assuming that because Canada announced this, Canadian gateway dwell times improve over the next few years. They will not. If anything the next few years get busier, because nothing new is being built while the corridor keeps carrying more.

One more thing for anyone whose volume is seasonal rather than steady. A step change in capacity landing in a single year does nothing for the eight years before it and nothing for the peak weeks inside those years, and peak is where west coast gateways actually fail. Winter weather on the corridor, a labour dispute at one door, or a week of vessels bunching after a typhoon season in Asia will hurt you long before 2035 and long after it. So build the plan around the peaks you can measure this year rather than the capacity you have been promised for the next decade, and hold some slack in the December to February window every single year, whether or not a terminal is being built anywhere.

Ports are slow objects. A causeway built in the 1960s still sets what can happen at Roberts Bank today, and a decision taken in 2026 sets what can happen in 2035, by which time most of the people reading this will be doing something else. The map changes at the speed of concrete, not at the speed of announcements. What I take from 30 October is not the 2.4 million containers. It is that Canada has decided it wants a second door with its own plumbing, and decisions like that tend to outlast the governments that make them.

↑ Back to the key points

— By Tomás Ferreira