Savannah Wins Permit for $2.3bn Hutchinson Island Terminal: 3.5m TEU, 12m Port Capacity

The Analysis

Spread the map out and find the bend of the Savannah River, the stretch of water where Garden City sits with its long line of cranes facing the channel. Ten years ago this was a story about depth: who had the water to take the big ships after the Panama gates were widened, and who would win the East Coast. That argument has been settled, and Savannah won a good part of it. The argument now is a different one, and it is being conducted in a file room rather than on the water.

On 14 October, at its State of the Port event, the Georgia Ports Authority said the US Army Corps of Engineers had approved its permit for the Savannah Container Terminal after a three-year review. Three big-ship berths on 350 acres at Hutchinson Island, $2.3bn, 3.5m TEU of annual capacity, taking the port from 7.75m TEU to about 12m. Construction starts in January 2028 and finishes in the mid-2030s. In the same season, Savannah worked 504,015 TEU in September, a record month, up 3.7% on a year earlier.

Read those two facts next to each other and the shape of the thing appears. The permit is real and the money is real, and neither of them moves a single container before 2028. Between now and then the port runs on what it already has, and what it already has just posted its busiest month on record.

Start with what actually moved on 14 October, because the language around it invites a mistake. The Corps approved a permit. It did not open a berth. A permit is a legal instrument: it says the authority may build, under conditions, inside a window. Three years of review ended that day, and the review is the part worth thinking about, because $2.3bn is not the scarce input in this story. Money of that size has been available to American ports for a decade. What has not been available is permission on a timetable, and the three years it took to get this one is the number that should interest anyone whose cargo moves through Georgia.

The bones of the project are conventional and well chosen: 350 acres on Hutchinson Island, three berths sized for the big ships, 3.5m TEU of annual capacity. Savannah's stated capacity today is 7.75m TEU; the announcement puts the port at about 12m once the new terminal is finished. Construction starts in January 2028. Completion is given as the mid-2030s, which is not a date but a range, and a wide one for a project that has to build a terminal, a channel interface, and whatever road and rail it takes to feed three new berths.

Notice that 7.75 plus 3.5 is 11.25, not 12. The extra three-quarters of a million TEU is not explained by the new berths. It has to come from the existing Garden City side running better than it runs today: denser yard, faster gate, fewer rehandles. That is a reasonable assumption and it is still an assumption. About 750,000 TEU of the headline number, roughly 6% of it, rides on improvements to the terminal that is already straining rather than on the one that has not been built.

Now set the announcement against the traffic it is meant to serve. Savannah worked 504,015 TEU in September, a monthly record, up 3.7% on a year earlier. Multiply by twelve and you get about 6.05m TEU. That multiplication overstates, because September sits in the peak import season and no port runs its best month twelve times; read 6.05m as a run-rate rather than a forecast and you have the useful version of it. Against 7.75m of stated capacity, a 6.05m run-rate is about 78%. This is not a port with room to spare. It is a port already running close to the number it quotes, in the month it quotes it. And design capacity is never the same as working capacity, because berths, yard and gate do not peak on the same day, and a terminal is usually congested well before its annual design figure is reached.

Here is the account I would actually run. Take 6.05m as the base and 7.75m as the ceiling; closing that gap needs 28.1% more volume. At September's own growth rate of 3.7% a year, that takes about 6.8 years, which puts the ceiling somewhere in 2033. At 6% growth, which is what a strong import year plus a shift of cargo from the West Coast would produce, it takes about 4.3 years, so 2031. At 2%, it takes about 12.5 years, so not until 2039. Three growth assumptions, three answers ten years apart. The project's own completion date, the mid-2030s, sits right on top of the middle one. The whole thing compresses into one line: the new berths are planned to arrive at almost exactly the moment the old ones run out, with nothing in between.

That coincidence is not a planning failure, it is the nature of the thing. Ports do not build capacity early, because empty berths are expensive and the political case for them is hard to make. They build when the forecast says they must, and the forecast is always made some years before the concrete. Which means the margin between running out and catching up is structurally thin, and anything that pushes on either side of it turns a tight fit into a gap. Demand running hotter than 3.7% pushes from one side; completion slipping past the middle of the decade pushes from the other. A gap of two years at this port would not appear as a headline. It would appear as berth waiting, yard density and dwell, which is to say it would appear on your invoice.

Ten years ago this map looked different, and the difference is worth holding in mind. Savannah was one of several East Coast ports hoping the widened Panama gates would deliver them the big ships, and the bet was on depth and on being the first call after the transit. It paid. What the port is doing now is not the same bet. It is a bet that the share it won can be held for another fifteen years, against Charleston and Norfolk on the same coast and against a West Coast that has spent two years trying to win its cargo back. A terminal that opens in the mid-2030s is a statement about 2045, because berth financing and concession horizons run on that scale. That is the right way to read the size of it: 3.5m TEU is not an answer to this year's congestion, it is a claim on what the East Coast's share will be in twenty years.

Why now, then? Partly because a three-year review ends on the Corps's calendar, not the port's. Partly because Savannah has a case to make and a record month to make it with. But the deeper reason is that the binding constraint on American port capacity has moved. For most of the last twenty years the constraint was money and dredging: raise the bond, deepen the channel, and the cranes follow. Today the channel at Savannah is deep enough for the ships that matter, and the money for this project is committed. What took three years was the review, and what can move the start date is not a bond market but a condition, an appeal, or a seasonal restriction written into the permit. That is the ten-year change in this network: steel is no longer the slow part.

There is a structural detail that deserves more attention than it is getting. Hutchinson Island is across the river from Garden City. This is not three more berths bolted onto the end of an existing quay; it is a second terminal cluster on the other side of the water from the largest single container terminal in North America. Two clusters means the landside has to be built twice: road access, gate capacity, bridge capacity, and the internal shuttle that moves boxes between the two banks. Ports that grow by adding a disconnected terminal almost always find that the new constraint is not the berth but the ground behind it. The pattern repeats everywhere, because terminals separated by water or distance turn drayage into the scarce resource, and the berths you built sit idle while trucks queue at the gate. Three big-ship berths will discharge boxes faster than any road network designed for one terminal can carry them away.

The inland side is where a port of this ambition usually meets its limit, and it is not what the announcement addresses. Savannah's case has always rested on something more than berths: putting a box on a train quickly, and reaching the population centres of the south-east and the mid-west within a day or two of the ship. Doubling the gate does not double that ability, because rail capacity, crew, chassis and inland terminal space move on their own clocks, and none of them is owned by the port. Ten years ago the binding constraint on an East Coast gateway was the channel. Now it is the network behind the gate. Anyone planning around Savannah's growth should be asking about that side with the same intensity they are asking about the cranes.

There is one more thing in the geometry. Savannah is the southern anchor of the East Coast's big-ship range, and Charleston and Norfolk are short sailing distances away but long landside distances away. A ship can be switched between them in a day; a distribution network cannot. So the decision that matters to a shipper is not which port the carrier calls at, it is where the inland footprint sits relative to two or three possible ports. The terminal at Hutchinson Island will change the first question in ten years, and it will not touch the second one at all.

So here is the thing the coverage is not saying, and it is why I would slow down before treating this as good news. The permit has been read as capacity, and it is not capacity. It is permission to start building in fifteen months for delivery in roughly a decade. Between those two dates Savannah runs on what it has, and what it has just posted its busiest month on record. Every decision an importer or exporter takes in the next five years, which gateway to route through, where to put the inland distribution centre, how long a rail commitment to sign, whether a three-day door delivery can be promised to a customer, is taken against a ceiling of 7.75m TEU rather than against 12m. The 12m figure is a number belonging to the 2030s being used to underwrite decisions taken in 2026. Hoping for it is fine. Planning on it is not.

The date itself is softer than it looks, too. A Corps permit of this kind carries conditions: mitigation, monitoring, restrictions on when work in the water may happen. Construction starts in January 2028 is a plan, and plans of this size move. A one-year slip is unremarkable. A two-year slip puts first concrete in 2030 and first box in the second half of the decade, which is four or five years after Savannah would have met its present ceiling on trend growth. If one sentence goes into the planning meeting, let it be this one: the risk is not that the terminal fails to get built, it is that it arrives later than it is needed, and later is the ordinary case.

Three things would turn this conclusion over, and it is worth knowing which one you are watching. If import growth through Savannah settles below about 2.5%, the ceiling never binds before the new berths open, the project stops being about relief and becomes about market share, and the 12m number turns into a bet on taking volume from Charleston, Norfolk and the West Coast rather than a response to running out of room. If growth runs above 6%, or if cargo shifts east again because of West Coast labour trouble or another thin year of water in Panama, the ceiling arrives around 2031 and two to three years have to be run hot before relief lands.

If the start date slips past 2028, both of the other cases get worse by the length of the slip. The variable that decides all three is not in Georgia. It is whether the East Coast keeps the share of United States imports it won over the last decade, and that is decided by Panama's water and the West Coast's labour calendar, neither of which the Georgia Ports Authority controls.

For an importer or exporter the translation is unglamorous, and most of it is about not doing something. Do not re-plan against capacity that does not exist. Plan against 7.75m and against a port that is already running at a record monthly rate. What remains is a question of how much optionality to buy, and that is a number you can put on paper. Assume a shipper moving 40 boxes a month through Savannah, 480 a year, at an average cargo value of $60,000 a box, with 8 of those 40 boxes tied to retail delivery windows that carry a penalty. Assume a tight year at the port, where berth waiting plus yard density adds two days to gate-out against today.

Two days on the 20% of boxes that breach free time, at an assumed $150 a box a day in detention and demurrage, is 96 boxes times two days times $150, about $28,800 a year. The carrying cost of holding two extra days of inventory across 480 boxes at 8% annual cost is 480 times $60,000 times 0.08 times 2/365, roughly $12,600 a year. The conditional one: if a delivery window is missed once in the year and the 8 window boxes have to be air-freighted at an assumed $8,000 premium each, that is $64,000, and at a 40% chance of it happening the expected cost is about $25,600. Add them and the quantifiable exposure in a tight year is roughly $67,000.

Now price the optionality. Qualifying a second East Coast gateway, Charleston or Norfolk, with a pilot covering 10% of Georgia volume is 48 boxes a year. At an assumed inland drayage and handling differential of $300 a box, that is $14,400 a year. Four test shipments, two per gateway, would tell you whether the transit holds within a day of Savannah's and whether your drayage carrier can actually cover the lane. $14,400 against $67,000 of exposure, and most of what a pilot really costs is somebody's time rather than money. A pilot you never use is a cheap insurance policy. A gateway you have never tested is not an alternative at all, it is a sentence in a contingency document.

The traps are predictable. Taking a short headline and signing a long contract against it is the common one: a five-year inland lease, a rail commitment, a cross-dock agreement running past 2032, all justified by a terminal that opens after the contract expires. New capacity is not the same as cheaper capacity, either, because new terminals arrive with new fee structures and a capital cost to recover, and their first two years are operationally their worst, since a terminal learns its own yard by running it. Treating the permit as funding certainty is the third one: $2.3bn is the project's number, and when it is drawn and in what order decides the pace, which is something to ask about rather than assume.

What actually helps is smaller than it sounds. Write the ceiling date into the plan rather than the capacity date. Test one alternate gateway while nothing is wrong, because testing under pressure is how pilots fail. Keep the option to move the seasonal peak separately from the base, because the base is cheap to move and the peak is not, and the peak is the part that misses windows. And ask the carrier a question almost nobody asks: which Savannah berth window are you contracting for, and what happens to the rotation once berth waiting runs past three days. That answer tells you more about your 2031 than the 12m headline does.

Spread the map out again in ten years. The river will bend the way it bends, the channel will be where it is, and there will be cranes on Hutchinson Island with boxes moving underneath them. The interesting question is not whether they get built. It is what the decade between the permit and the first box costs the people who were told relief was coming. Ten years ago this bend of the river was the East Coast's answer to a question about depth. Now it is an answer with a date attached, and the date is the part worth writing down.

↑ Back to the key points

— By Tomás Ferreira