BREAKING: San Francisco’s Famous Leaning Tower Is Still Sinking

本站添加: 2026-06-02

[00:00:00]San Francisco's most infamous skyscraper is sinking again. And now engineers say the foundation itself may be bending.

[00:00:09]After nearly $100 million in repairs, Millennium Tower is still moving. When Millennium Tower opened in 2009, it was marketed as the future of luxury living in San Francisco. The 58story skyscraper featured private spas, floor to-seeiling glass walls, valet service, and pen houses worth more than $10 million.

[00:00:35]But beneath the polished glass exterior, the building had already begun sinking into the soft ground below. Over time, the tower dropped 18 in and tilted 26 in off vertical at the top. A massive engineering rescue project was launched to stabilize it, costing nearly $100 million.

[00:00:58]After 9 years of work, the system corrected the tilt by only about 1 in far less than engineers originally predicted. It corrected the tilt by exactly 1 in a quarter of what its own designer had forecast. So, how does a $750 million luxury tower built with today's technology end up sinking, leaning, and now bending into the ground.

[00:01:25]But first, let's go back to when this building was supposed to be perfect.

[00:01:30]This is Millennium Tower, 58 stories, 419 luxury condos, located right in the heart of San Francisco's financial district at 301 Mission Street. When it opened in 2009, it was a statement. Floor to ceiling windows, a private spa, valet parking, interiors designed by Yabu Pushelberg.

[00:02:00]Pen houses sold for over $10 million.

[00:02:04]Smaller units traded at $2 to5 million.

[00:02:08]Joe Montana, the legendary 49ers quarterback, bought in. The family of former Secretary of State Warren Christopher owned property there. Tech executives, hedge fund principles, professional athletes, the kind of people who could buy literally anywhere in the world chose this address. The original engineers predicted the building would settle 6 in over its entire 100-year lifetime. 6 in total.

[00:02:40]It did that and three times more in the first 7 years. And here's the thing.

[00:02:47]San Francisco's south of market district sits on reclaimed shoreline. Two centuries ago, that land was part of San Francisco Bay. Developers filled it in over decades.

[00:03:01]So when you drill down underneath 301 Mission Street, you do not hit rock right away. Under the tower sits layer after layer of unstable ground, artificial fill, soft bay mud, compressible clay still consolidating decades later. Bedrock lies nearly 200 ft below the surface. Millennium Tower never reached it. Almost every super tall building in San Francisco is anchored into that bedrock. Salesforce tower just two blocks away sits on piles that go the full 200 ft down. That is how you build a skyscraper on questionable ground. You drill all the way to something that does not move.

[00:03:49]Millennium Tower did not do that. When the developer Mission Street Development, an entity of Millennium Partners, broke ground in the mid 2000s.

[00:03:59]The structural engineer of record was D.

[00:04:01]Simone consulting engineers. The geotechnical engineer was Treadwell and RO, a firm later renamed Langun. The foundation they specified was a roughly 10 ft thick concrete mat resting on approximately 950 reinforced concrete friction piles. And those piles only went 60 to 90 ft down. They stopped in the Kulma sand layer. They did not reach bedrock.

[00:04:31]Now, here is the difference between the two methods because this is the entire story right here.

[00:04:38]A pile that reaches bedrock transfers the building's weight to a hard immovable layer. It is like setting a chair on a concrete floor. The load goes straight down. No give, no surprise.

[00:04:53]Friction piles are different. They hold the building up by squeezing against the soil along their length. It is like sticking a pencil into a tub of wet sand and trusting the sand to grip it tight enough to hold a brick on top. In theory, with the right calculations, friction piles work fine. They are cheaper. They are faster to install. And the design team's models projected only 6 in of settlement over the building's entire lifetime.

[00:05:26]There was just one number they underestimated, the weight. Millennium Tower was built primarily of reinforced concrete instead of the lighter steel framing most San Francisco highrises use. When it topped out, it weighed 686,000 tons. That made it one of the heaviest residential structures in the entire city. And it was sitting on friction piles on still consolidating clay on filledin bay mud. The ground started reacting before the building even opened. The weight compressed the sand.

[00:06:02]The sand pressed on the clay below.

[00:06:06]The clay began its slow squeeze. Water left the tiny spaces between particles.

[00:06:12]The particles packed tighter. The surface dropped.

[00:06:16]By the time it opened in 2009, it was already moving faster than predicted.

[00:06:22]Then between roughly 2010 and 2014, the Transbay Transit Center began construction less than 100 ft away. That excavation required massive dewatering.

[00:06:36]Crews pumped millions of gallons of groundwater out of the pit every day.

[00:06:42]The water table across the entire block dropped in the clay layers. The effective stress rose because the water no longer helped buoy the soil particles. The clay consolidated faster.

[00:06:57]It squeezed like a sponge under a boot.

[00:07:00]The tower started sinking immediately.

[00:07:04]The rate doubled, then tripled. By 2014, residents on the upper floors began noticing things. Doors that would not close. Cracks at the corners of the drywall.

[00:07:17]Vertical lines inside the apartments that no longer matched up with the city skyline outside the windows.

[00:07:25]The building was sinking and it was not sinking evenly. On the sixth floor, an attorney with a marble. On August 2nd, 2016, a sixthfloor homeowner named Jerry Dodson walked into the lobby of Millennium Tower with a marble in his pocket and met an investigative reporter from NBC Bay Area named Jackson Vanv.

[00:07:50]Together, they filmed the marble rolling on its own across the floor toward the Northwest.

[00:07:56]That video went everywhere. The moment went viral and the story has not stopped breaking since. The official numbers came out shortly after.

[00:08:07]By 2016, Millennium Tower had already sunk 16 in into the ground, more than a foot, more than double what engineers had predicted. And it was not sinking evenly. The northwest corner had sunk more than the rest, which meant the building was tilting. approximately 14 in of lean at the top and getting worse.

[00:08:33]The problem was leaning. To put that in perspective, the leaning tower of Pisa leans about 12 ft at the top, but it took centuries to get there. Millennium Tower moved 14 in in just a few years.

[00:08:49]The rate was not slowing down. It was accelerating. The lawsuits flew in every direction. The Homeowners Association sued Mission Street Development. Mission Street Development sued the Transbay Joint Powers Authority, blaming the dewatering next door. Mission Street also sued its own engineers, Treadwell and Rolo and Desimone. The engineers denied responsibility. The Transbay team fired back with their own data, showing the tower was already sinking before they ever broke ground.

[00:09:25]Property values collapsed. Units that had sold for $5 million could not find buyers at $3 million.

[00:09:34]Some owners were underwater on their mortgages within months.

[00:09:38]Pen houses that had traded for over $10 million dropped 30 to 50%.

[00:09:45]Smaller units lost nearly half their value. By 2018, a multi-party settlement framework worth roughly $30 million had been hammered out.

[00:09:56]But the real question was not who was going to pay. The real question was, can this thing actually be fixed?

[00:10:05]The challenge was unlike anything modern engineering had really attempted. You cannot rebuild a foundation underneath the 58story occupied skyscraper. You cannot ask 400 families to leave, demolish the building down to ground level, and start over. You cannot excavate underneath a standing tower without risking catastrophic collapse.

[00:10:30]So, whatever fix you came up with had to work around the building that was already there. The man hired to design the rescue was one of the most respected forensic structural engineers in America. His name is Ron Hamburger. He is a senior principal at Simpson Goomperts and Hager.

[00:10:50]After September 11th, 2001, he led the structural investigation of the World Trade Center collapse.

[00:10:58]If anyone in this country could solve this problem, it was him. His original plan announced in 2019 called for 52 micro piles drilled all the way to bedrock around the entire perimeter of the building.

[00:11:14]The work began in November 2020 and almost immediately something went wrong.

[00:11:20]By the summer of 2021, monitoring data showed the tower was sinking faster during the rescue than it had been before. The drilling itself, the vibrations, and the redistribution of stress around the existing friction pile system were destabilizing the very foundation the new piles were trying to save. The northwest corner sank an additional 2 in during construction.

[00:11:47]That was not supposed to happen. The fix was supposed to stop the movement. It was making it worse. The project paused.

[00:11:56]The engineers went back to the drawings.

[00:11:59]The redesign that emerged in 2022 was radically smaller. Instead of 52 piles around the entire perimeter, there would only be 18. Instead of all four sides of the building, the work would target only the north and the west. The two faces the tower was tilting toward. The piles would go 275 ft down this time all the way to bedrock.

[00:12:27]A new collar would tie them into the existing concrete mat.

[00:12:32]Pile installation finished in May 2023.

[00:12:37]Load transfer and monitoring continued for two more years. And here is where the story becomes uncomfortable.

[00:12:45]Ron Hamburger had publicly forecast that the engaged piles would correct the tilt by about 4 in in the first 6 months of operation.

[00:12:55]In August 2025, when the fix was officially declared complete, the actual measured correction was about 1 in 1 in 1 in. That's roughly 25% of what the designer of the Pix had predicted. The most respected forensic structural engineer in America designed a $100 million rescue. It hit a quarter of its target. And then the story got worse.

[00:13:20]Late in 2025 and going into 2026, C. Bay Area's continuing investigation still anchored by Jackson Vandervan.

[00:13:34]The same reporter who broke the story back in 2016, surfaced something the rescue engineers had not anticipated.

[00:13:43]The tower's central footprint, the part of the foundation mat that sits between the original friction piles and is no longer being asked to do the same work as the newly engaged perimeter.

[00:13:56]On the north and west appears to be settling in a different pattern than the edges. In plain English, here is what that means. The edges of the building are now being held up by the new bedrock piles on the north and west sides. But the middle of the building, the central section of the mat foundation, is still resting on those old friction piles in the Kulma sand. And those friction piles are still sitting on top of clay that is still consolidating.

[00:14:26]So now the edges are being held, but the middle is dishing.

[00:14:32]The leading hypothesis from engineers studying the data is that strengthening only two sides of the building transferred the load paths in a way that allows the unsupported portion of the mat to keep settling into the old bay clay underneath it. The fix did not fail the way most people would expect. It worked on the corners they targeted, but it may have created a brand new problem in the part of the foundation they did not touch. The tower is no longer just sinking. The tower is no longer just leaning. The foundation itself may now be bending.

[00:15:10]This is a 645 ft residential tower with hundreds of people living inside it. The foundation is no longer behaving like a single connected slab.

[00:15:23]Different parts of it are moving in different ways. The marble that Jerry Dodson rolled across his floor in 2016 still rolls today.

[00:15:33]Doors that were patched have stopped closing again. Cracks that were sealed have reopened.

[00:15:40]Tape lines on the floors of common areas track movement that residents can see with their own eyes. These were the most expensive condos in San Francisco.

[00:15:51]Today, the people living in them watch marbles roll for entertainment.

[00:15:57]Some residents stayed. They believed the fix worked. They trusted the engineers.

[00:16:04]They decided the building was now safer than it had ever been because at least part of it is finally anchored to bedrock, the way it should have been from the start.

[00:16:14]Others left.

[00:16:16]They sold at massive losses 30, 40, sometimes 50% below what they paid.

[00:16:23]A unit that traded for $5 million in 2015 is now worth $2 to $3 million, if you can find a buyer at all.

[00:16:33]Even after the August 2025 fix was declared complete, the discount has not closed.

[00:16:41]Buyers know exactly what they are buying and some are still there stuck. They cannot sell because nobody will pay what they need to break even. They cannot leave because they cannot absorb the financial hit.

[00:16:56]So they live inside a building they no longer trust, paying taxes on values that no longer exist.

[00:17:03]Watching sensors track every millimeter of movement. The building is monitored 24 hours a day, 7 days a week. Engineers review the data monthly. Reports go to the city on a regular schedule.

[00:17:18]Officially, the structure is sound. It is not going to fall. That is the line everyone gives. But this story is bigger than one building. Here is the real lesson. San Francisco is not the only city built on soft ground. Seattle has similar soil conditions. Parts of New York City are built on fill.

[00:17:42]Miami, New Orleans, Boston, Amsterdam, and large parts of coastal Asia all face the same fundamental challenge. You are putting heavy buildings on ground that wants to move. And the question every developer, every engineer, and every city inspector has to answer before construction starts is the same question Millennium Tower got wrong.

[00:18:07]How deep do you go?

[00:18:10]How much do you spend?

[00:18:13]What happens when you guess wrong? This is not the first time a major project has underestimated the ground it stood on. Consider New Orleans. In 2005, Hurricane Katrina brought a storm surge that tested the city's levey system.

[00:18:30]Those leveies were built on soft delta soils, much like the bay fill under San Francisco. The designers used sheet piles driven into the mud. They calculated the resistance. They thought it was enough. When the water pressed against the walls, the soil beneath gave way. The levies slid. Water poured into the city. Neighborhoods flooded to the rooftops. More than 1,800 people died. The economic damage ran into the tens of billions. The similarity is not in the water. It is in the assumption.

[00:19:08]In both cases, the engineers trusted their calculations more than the unpredictable behavior of soft water-bearing soils under sustained load. They built large. They built to last. And when the real test arrived, the ground moved in ways the models had not fully captured.

[00:19:28]Millennium Tower had no hurricane, just time, gravity, and the slow consolidation of clay, but the outcome was the same kind of wakeup call. Big structures on marginal ground carry hidden risks that only reveal themselves after the fact. a $350 million tower with a $100 million patch that did not work the way it was supposed to. Still standing, still settling, still leaning, and now bending on ground that was never solid in a city that has known for 9 years exactly what it built and what it cannot fix.

[00:20:14]And somewhere right now, an engineer is running the numbers on the next tower, looking at a budget and asking the same question that started this whole story.

[00:20:25]How much foundation is enough? What could go wrong from here?

[00:20:30]If another strong earthquake hits the bay area and the fault lines are active, the differential settlement could interact badly with seismic shaking.

[00:20:40]The bend in the foundation might concentrate stress.

[00:20:44]Cracks could propagate faster. The tilt could increase suddenly if the middle drops further during the shaking. The building is designed for earthquakes.

[00:20:56]But it was not designed for a foundation that is actively warping.

[00:21:01]Or if the consolidation in the center continues for another decade, the dishing could reach a threshold where the mat loses its ability to distribute loads evenly. The upper structure might experience twisting forces it was never meant to handle.

[00:21:18]Windows could fail. Interior partitions could shift permanently. The building would still stand, but it would be a different building than the one that opened in 2009.

[00:21:31]These are not predictions. They are possibilities the data now forces us to consider. The engineering community is not unanimous on what the latest measurements mean. Some say the 1in correction falls within the margin of error for such complex soil structure interaction.

[00:21:51]Soil is never perfectly uniform. Small variations in the clay layers can change outcomes.

[00:21:58]Others see the bending as evidence that the partial fix created a new failure mode. The load was shifted, not eliminated.

[00:22:07]The city of San Francisco continues to accept the monthly reports, no evacuation orders, no structural red flags, but the long-term performance of the central mat under continued differential movement is still being studied. No one can say with certainty what the next 5 years will show. The marble rolled across the floor in 2016. It still does.

[00:22:35]What does it take for a city to trust a building again after something like that? If you were the engineer signing off on the next tower on this same block, how deep would you drive the piles? The Millennium Tower still rises above the financial district. Its sensors record every shift. The clay beneath continues its slow work.

[00:22:58]In San Francisco, the lesson sits in plain sight at 301 Mission Street. The ground does not forget what it once was.