Cities built on compressible fill or soft soil face long-term structural risks, as demonstrated by San Francisco's Salesforce Tower and Millennium Tower, where buildings are experiencing differential settlement and tilting due to foundation piles not reaching bedrock, combined with factors like sea level rise and the Hayward fault's seismic threat, highlighting the need for thorough geotechnical surveys and transparent communication about ground stability in urban development.
Deep Dive
Prerequisite Knowledge
- No data available.
Where to go next
- No data available.
Deep Dive
SAN FRANCISCO ALERT! The Sinking Skyscraper Crisis Just Took a Terrifying TurnAdded:
People are living in this thing while this work is going on. Uh temporary >> right now in May 2026. The ground beneath San Francisco is moving, not shaking, not rumbling, moving slowly, silently, and with the kind of quiet inevitability that keeps structural engineers awake at 3:00 in the morning, staring at spreadsheets they wish said something different. One of the most recognizable skylines in America is sinking into the mud it was built on.
The buildings are tilting. The numbers are getting worse and the warning signs that were ignored for years have just crossed a threshold that scientists are calling with no exaggeration terrifying.
Hi, my name is Daniel and this is Natural Disasters. The Salesforce Tower situation.
Salesforce Tower, 326 m tall, the tallest building in San Francisco, a glass and steel monument to the tech boom that reshaped this city over the last two decades. When it opened in 2017, it was celebrated as an engineering triumph, a symbol of ambition, a declaration that San Francisco was competing with New York, with Chicago, with the great skylines of the world. Here is what the press releases did not mention. The building was going up in a city where the ground does not stay still. In May 2026, monitoring data from sensors embedded in the foundation and the surrounding soil has confirmed what structural engineers have been quietly tracking for months.
Salesforce Tower is experiencing measurable ground settlement. The numbers, depending on which monitoring firm you ask, range from concerning to alarming. Differential settlement, meaning the building is not sinking evenly but tilting, one side moving faster than the other, has been detected at rates that exceed the tolerances written into the original engineering specifications.
Now, before you panic and cancel your flight to San Francisco, and look, I am not here to tell you what to do, but maybe do not book a room on the 47th floor just yet, let me give you some context. Settlement in tall buildings on soft soil is not unusual. buildings sink. That is a known phenomenon. What engineers design for is controlled, predictable, uniform settlement. What they design against is differential settlement. The kind where one corner drops faster than another, where stress concentrates in the structural frame in ways the original calculations did not account for. And that is what the sensors are showing right now.
Horizontal displacement measurements taken between January and April 2026 show a cumulative shift that has crossed the threshold from what engineers call within acceptable parameters into what engineers call requiring immediate further investigation. That phrase requiring immediate further investigation is the structural engineering equivalent of a doctor saying let us schedule some additional tests in a very calm voice while writing something alarming in your file. The city of San Francisco's Department of Building Inspection has acknowledged that monitoring is ongoing. They have not acknowledged what the monitoring is finding. That gap between what is being measured and what is being said is where this story lives. The Millennium Tower blueprint for disaster. To understand what is happening right now, you need to understand what happened first. And that means talking about Millennium Tower.
Millennium Tower opened in 2009. 58 stories, 419 luxury condominiums. Units selling for between 1 and $10 million to buyers who were told they were purchasing a piece of San Francisco's most prestigious new address. The building won architectural awards. It was profiled in magazines. It was by every visible measure a triumph. Then it started sinking. By 2016, residents noticed doors that no longer closed properly. cracks appearing in walls, elevator shafts showing signs of stress.
When independent engineers were finally brought in, the findings were stunning.
Millennium Tower had sunk approximately 41 cm into the ground, roughly 16 in, and had tilted approximately 61 cm off vertical at the roof line. To put that in perspective, you are talking about a 58story building leaning at a measurable angle that you could detect just by standing inside it and trusting your own sense of balance. The cause, when it was eventually untangled through years of litigation that involved the developer, the city, the neighboring transit construction project, and more lawyers than I care to count, came down to one fundamental problem. The building's foundation piles did not reach bedrock.
They stopped in sand and clay. and the sand and clay subjected to the weight of 400,000 tons of concrete and steel and the cumulative drainage effects of nearby construction compressed slowly, steadily, inevitably. The fix, a perimeter pile retrofit that was supposed to stabilize the structure, was completed in 2023 at a cost of approximately $100 million. And here is the part that nobody really wants to talk about. The fix stopped the tilting from getting worse. It did not reverse it. Millennium Tower is still leaning.
It will always be leaning. The question that the retrofit answered was not can we fix this, but rather can we stop it from getting worse. And even that answer came with qualifications that the residents association would prefer you not read too carefully. But here is what Millennium Tower actually revealed.
beyond its own walls and its own litigation. It revealed something about the ground beneath San Francisco that the city's development boom of the 2010s had been cheerfully ignoring. The soil that San Francisco's newest, tallest, most expensive buildings were being constructed on was not stable ground. It was fill. It was bay mud. It was material deposited over decades of landfill operations that turned the shallow edges of San Francisco Bay into developable real estate. and in doing so created a geological ticking clock that nobody wanted to put on the project timeline. Millennium Tower was not a one-off failure. It was a preview. And in 2026, the preview is starting to look like the feature presentation. The geology beneath the city. Here is something that the San Francisco Tourism Board would prefer not to be the opening sentence of a travel brochure.
Significant portions of San Francisco are built on landfill that was dumped into San Francisco Bay between the 1850s and the midentth century. When the gold rush hit in 1849, the city exploded in size almost overnight. Thousands of ships arrived in the bay and were simply abandoned when their crews deserted to go find gold. Those ships were buried.
Streets were built over them. And as the city grew, it kept pushing outward into the bay, using rubble, debris, garbage, and whatever else was available to create new land where there had been water. The result, and I want you to really let this sink in, no pun intended, is that some of San Francisco's most valuable real estate is sitting on what is essentially a mixture of old ships, Victorian era trash, bay mud, and sand. the neighborhood called South of Market, where Millennium Tower stands, where Salesforce Tower stands, where billions of dollars of tech era real estate has been developed over the last 25 years, sits on some of the softest, most compressible, most seismically vulnerable soil in any major American city. Geotechnical surveys describe the subsurface conditions in this area as young bay mud over loose sand over older bay mud. The water table is high. The soil is saturated. And critically, the soil has what engineers call a lowbearing capacity, meaning it cannot support heavy loads without compressing over time. Now, in normal circumstances, this is manageable. You design deep foundations. You drive piles down to bedrock. You engineer around the problem. The problem is that bedrock beneath parts of San Francisco's south of Market District sits between 60 and 90 m below the surface. That is a long way down. And not every building that went up during the development boom of the 2010s paid the cost, financial and logistical of going all the way down to bedrock. Some stopped in the dense sand layers above it, betting that those layers were competent enough to carry the load. That bet is now being called.
Additionally, sea level rise, currently tracking at approximately 3.2 mm per year in the San Francisco Bay area, is doing something insidious to this already compromised soil. Rising groundwater levels are increasing poor water pressure within the soil column.
In plain terms, the ground is getting wetter from below. Wetter soil compresses more easily under load. The buildings above it are not getting lighter. The math is going in exactly the wrong direction. the numbers that are breaking scientists. Let us talk about what the sensors are actually showing. Because the data coming out of ground monitoring stations across San Francisco's south of market district in the first quarter of 2026 is not the kind of data that makes geotechnical engineers want to come into work on a Monday morning. It is the kind of data that makes them pour a second coffee, close their office door, and call their colleagues in very quiet voices. Insar technology interferometric synthetic aperture radar uses satellite imagery to measure ground surface movement down to the millimeter not the centimeter the millimeter. It is extraordinarily precise. It covers enormous areas simultaneously. And it does not care about press releases or non-disclosure agreements. The satellites just watch the ground and report what they see. And what they are seeing over the south of Market Corridor between January and April 2026 is not good. Localized subsidance rates near Salesforce Tower and the Transbay Transit Center have accelerated measurably compared to the baseline data collected between 2022 and 2023. We are talking about vertical ground movement of between 8 and 14 mm over a 4-month window. I know that sounds small. I know 8 mm sounds like nothing. Here is the context that makes it not nothing. Annualized, those readings translate to a subsidance rate of between 24 and 42 mm per year. The foundation engineering specifications for structures in this corridor were designed around anticipated long-term settlement rates significantly below that range. The ground is moving faster than the engineers planned for. Not slightly faster, meaningfully faster.
The kind of faster that causes people who understand what the numbers mean to use words like accelerating and anomalous in the same sentence without anyone in the room laughing. Then there are the tilt meters. Tilt meters are sensors embedded in structures that measure angular deviation from vertical.
In plain terms, they measure how much a building is leaning. Several south of market highrises have tilt meters installed as part of their ongoing structural monitoring programs. And in the months between late 2025 and early 2026, multiple tilt meter readings on multiple buildings triggered automatic alerts to the engineering firms responsible for monitoring those structures. automatic alerts, meaning the numbers crossed a threshold that the engineers themselves had pre-programmed into the system as the point at which someone needs to be notified immediately. Now, here is where I have to be straight with you. The specific numbers from those alerts have not been made public. The building owners have not released them. The city has not released them. What has been confirmed through sources speaking to structural engineering trade publications rather than through any official statement is that the alerts were genuine. They were not instrument malfunctions. They were not software errors. The buildings were doing something that the sensors were designed to detect and the sensors detected it. And then there are the voids. ground penetrating radar surveys conducted in March 2026 near the Transbay Transit Center, the transit hub that opened in 2018, that has its own extensively documented history of settlement problems and that sits directly adjacent to Salesforce Tower, detected what the survey teams are formally describing as anomalous subsurface voids in the fill layer beneath the structure. voids, empty spaces, hollows in the ground beneath a building that processes 25,000 passengers every single day in the middle of a city of 870,000 people on soil that was never bedrock to begin with. I will give you a moment with that. The Hayward fault factor.
Everything we have talked about so far, the sinking, the tilting, the soft soil, the accelerating subsidence rates, the voids, all of that is the situation in calm conditions. No earthquake, no major seismic event, just the slow, steady, grinding physics of heavy buildings on compressible ground. Now add an earthquake. The Hayward fault runs for approximately 72 miles through the East Bay directly across San Francisco Bay from the downtown skyscraper corridor.
It passes beneath the cities of Berkeley, Oakland, Fremont, and Hayward, densely populated urban areas that sit directly on or immediately adjacent to the fault trace. Seismologists at the United States Geological Survey have been calling the Hayward fault the most dangerous urban fault in the United States for years. Not the San Andreas fault, which gets all the press, the Hayward. Here is why. The San Andreas runs largely through less densely populated terrain for much of its length. The Hayward runs directly beneath and through one of the most densely populated areas in the American West. and it is in seismological terms overdue. The last major Hayward fault rupture occurred on October 21st, 1868, magnitude 6.8. In 1868, Oakland was a small town. San Francisco had not yet built most of what it would build on landfill. The shaking was catastrophic for the era, but the exposure, the number of people and structures in the hazard zone was a fraction of what it is today. Current USGS probability estimates place the likelihood of a magnitude 6.7 or greater Hayward fault earthquake at approximately 33% within the next 30 years. That is the scientist being measured and methodical. Some researchers put the number higher. What happens to a building that is already experiencing differential settlement?
Already leaning, already showing sensor alerts, already sitting above anomalous subsurface voids. When a magnitude 6.7 earthquake shakes the region, the answer, and every structural engineer who has looked at this scenario will tell you this is that you do not know.
The uncertainty is not reassuring. A building that is already stressed, already tilted, already compromised at the foundation level has less margin.
Less margin means less tolerance for additional load. Less tolerance means a lower threshold for catastrophic failure. Liqufaction is the other word that comes up in every conversation about San Francisco and earthquakes and it needs to be said plainly. When saturated loose sand is subjected to strong shaking, it loses its structural integrity and behaves like a liquid. The fill beneath south of market is composed largely of material with high liquefaction susceptibility. In a major Hayward event, parts of San Francisco's most developed, most expensive, most densely occupied real estate corridor could quite literally turn into soup.
The human stakes. I want to step back from the data for a moment because it is very easy to spend 40 minutes talking about millimeters and fault probabilities and Insar readings and tilt meter alerts and forget that none of this is actually about millimeters.
It is about people, real ones. and I think they deserve a section of this video that treats them like that. So, let us talk about who is actually in these buildings. On a typical weekday in May 2026, approximately 55,000 people work in San Francisco's South of Market district. 55,000.
That is roughly the population of Burbank, California, showing up every morning, taking the elevator, sitting down at their desks, and going about their day on top of soil that the sensors are telling us is moving faster than anyone planned for. Salesforce Tower alone hosts around 8,000 people during business hours. 8,000 people in a single building on a single block above a single patch of ground that is recording subsidance rates outside its own engineering tolerances. The Transbay Transit Center, the hub sitting directly above what surveyors are now calling anomalous subsurface voids, moves approximately 25,000 passengers through its corridors every day. Commuters, students, elderly residents going to medical appointments, families with strollers navigating the concourse, 25,000 people passing daily through a structure built on fill above hollows in the ground that until March 2026, nobody knew were there. Not one of those people has been officially told. And I want to sit with that for a second because the argument I hear most often when this topic comes up and I do talk to people about this, this is not a purely theoretical exercise for me. Is that we should not alarm people unnecessarily.
That the situation is being monitored, that engineers are on top of it, that responsible communication means waiting until you have the full picture before you say anything publicly. I understand that argument. I genuinely do. Panic is not useful. Incomplete information can cause more harm than no information. I am not dismissing that concern. But here is the other side of it. There are people in the south of market corridor who own property there. Who signed 30-year mortgages on condominiums and buildings that are sitting on soil that is compressing at rates the foundation engineers did not design for. There are renters who chose buildings based on price and location and safety ratings that were calculated before the 2026 monitoring data existed. There are workers who have no idea that the tilt meter in their building triggered an automatic engineering alert sometime in the last 6 months. These people cannot make informed decisions about their own lives if the information being gathered about the ground beneath their feet is being held in engineering reports that the public cannot access. I genuinely debated putting this next part in the script. I am including it because I think you deserve the full picture.
Several structural engineers who spoke to trade publications in the first quarter of 2026, not on the record, and I understand exactly why not, use the phrase generational liability to describe what is developing in San Francisco's high-rise corridor.
Generational liability. The decisions made during the development boom of the 2010s will outlast the people who made them. The costs will be paid by people who had no vote on the matter. And the problem is not coming. In May 2026, it is already here. Nobody is evacuating south of market. Nobody is shutting down Salesforce Tower. Maybe that is correct.
Maybe the engineers reviewing this data have concluded the risk is manageable.
And I simply do not have access to what they know. But I will tell you this much. The sensors are running. The data exists. And the people living above it deserve to know what it says. The engineering coverup question. Let me be careful here because this section is where responsible journalism and sensationalism are standing very close together and I want to make sure I am on the right side of that line. I am not going to tell you there is a cover up.
What I am going to tell you is that there is a pattern of delayed disclosure around foundation and settlement issues in San Francisco's high-rise sector that is well documented and that raises legitimate questions about what building owners, developers, and the city know, when they know it, and when the public finds out. The Millennium Tower situation is the clearest example.
Residents reported problems starting around 2014. Engineers hired by the building's homeowners association confirmed significant settlement in 2016. The developer did not proactively disclose the issue. The city did not issue any public warning. The residents found out because they hired their own engineers. The full scope of the problem, including the debate over what caused it and who was responsible, took years of litigation to partially resolve. Settlement agreements included non-disclosure provisions that prevented certain parties from speaking publicly about what they knew and when. Now, I am not suggesting that every developer in San Francisco is concealing structural problems. I am suggesting that the legal and financial incentives around disclosing foundation issues in buildings worth hundreds of millions of dollars are not necessarily aligned with rapid, transparent, proactive public communication. When a building is worth $400 million on the open market and disclosure of a settlement problem could reduce that value by 40% overnight, the pressure to manage the narrative, to wait, to study, to commission additional reports, to consult with lawyers before communicating with residents is enormous. What is different in 2026 is that the monitoring technology has gotten better. Insar satellite data is publicly available. Independent researchers can and do analyze it. The information is harder to contain than it was in 2014. The ground, in a very literal sense, is harder to hide.
Several San Francisco supervisors have, as of May 2026, called for a citywide independent audit of foundation monitoring data for all buildings over 20 stories in the south of Market District. The audit has not yet been commissioned. The call has been made.
The silence since then has been loud.
The domino effect. Here is where the story gets bigger than any one building.
The south of market district is not just home to a handful of high-profile skyscrapers. It is home to dozens of large commercial and residential structures built over the last three decades. Many of them on the same compressible fill. Many of them with foundation designs that were approved under building codes that have since been revised. And some of them, this is documented, not speculation, with foundation systems that were considered marginal even at the time of construction. When Millennium Tower settlement became public, independent engineers began reviewing the foundation records of neighboring buildings. What they found was not reassuring. Multiple structures within a fiveb block radius of Millennium Tower were identified as having foundation systems with limited redundancy, meaning systems that were designed to work adequately under expected load conditions, but that had little engineering margin for additional stress. Settlement does not stay in one place. When a large building sinks unevenly, it changes the stress distribution in the soil around it. It can accelerate settlement in adjacent structures. It can redirect groundwater flow in ways that affect neighboring foundations. It can, under the right conditions, create a progressive failure scenario where one building's movement destabilizes the soil condition for the building next to it, which in turn affects the building next to that.
Geotechnical engineers call this group effects, the polite version. The less polite version is dominoes. In May 2026, InSAR data covering a 12b block area in south of Market shows that the elevated subsidance rates are not confined to the footprints of individual buildings. They are distributed across the area in patterns consistent with regional soil compression. Meaning the soil itself across a broad area is compressing not just beneath the buildings between them around them. The water table fluctuations tied to regional drought and groundwater extraction patterns are contributing. Sea level pressure on the bay mud layer is contributing. The accumulated weight of 25 years of high-rise construction is contributing.
The city built a skyline on mud. The mud has been patient. In 2026, the patience appears to be running out. 12 city blocks of the most valuable real estate in Northern California are showing signs of ground movement that no single building's engineering team is responsible for solving. And nobody has yet explained who is the city's response. On April 22nd, 2026, the San Francisco Department of Building Inspection issued a statement regarding foundation monitoring in the south of Market District. I want to read you the key line from that statement because I think it is instructive. The statement said, "Routine monitoring of structures in areas of known geotechnical sensitivity is ongoing and the department is committed to ensuring the safety of all occupied buildings.
Routine monitoring, known geotechnical sensitivity, committed to ensuring safety." That statement was issued 3 weeks after independent satellite data showed subsidance rates in the south of market district running at levels that structural engineers described to trade publications as significantly elevated.
3 weeks after sensors on at least two south of market highrises triggered automated engineering alerts and one week after a city council member publicly requested an independent foundation audit. To be completely fair, and I try to be fair on this channel, even when I am also being alarming because you deserve both. The city is not doing nothing. Monitoring is genuinely happening. Engineers are genuinely looking at data. The building inspection department does not have unlimited resources and is responding to a situation that is complex and not yet fully characterized. But here's the thing that I keep coming back to. The people who live and work in these buildings are not being told what the monitoring is finding. The data that would allow San Francisco residents to make informed decisions about their own safety. Data that exists, that is being collected, that engineers are actively reviewing, is not being shared publicly in any form that is actually useful.
Routine monitoring, known geotechnical sensitivity, committed to ensuring safety. Those are words. In San Francisco right now, the ground is doing something that deserves more than words.
One week after this story first broke into mainstream consciousness, San Francisco's skyline looks the same from the outside. Salesforce Tower still catches the afternoon light. The bay is still beautiful. The sourdough is still excellent. I want to be clear, I am not casting any aspersions on the sourdough.
The sourdough is blameless in all of this. But beneath the surface, the numbers keep moving. The soil keeps compressing. The sensors keep recording.
And a city that built its newest, most ambitious chapter on landfill and bay mud, and the engineering optimism of a real estate boom is beginning to confront the geological invoice for those decisions. San Francisco is not alone. Miami is built on poorest limestone that is dissolving. New Orleans sits below sea level and sinks a little further every year. Jakarta sank so severely that Indonesia moved its capital. Houston subsided dramatically due to groundwater extraction. The list of major cities built on geologically problematic ground is not a short list.
It is in fact most of the great coastal cities of the world. What San Francisco represents in May 2026 is a specific and urgent version of a global question that the 21st century is going to force every one of those cities to answer. What happens when the engineering assumptions of the 20th century meet the geological and climatic realities of the 21st? What happens when the ground moves faster than the model said it would? The skyscrapers are still standing. For now, they are still standing, but the ground beneath them is telling a story that the buildings themselves cannot hide forever. The sensors are running. The satellites are watching. And somewhere beneath the gleaming towers of south of Market, 40 meters down in the dark, the mud is moving.
Related Videos
U.S. Military Just Flexed The Most Dangerous Aircraft Ever Built The F-47
MaxAfterburnerusa
11K viewsβ’2026-05-29
Heating Staying On On The Hottest Day Of The Year
PlumbLikeTom
507 viewsβ’2026-05-29
λ°μ ν¨μ¨μ λμ΄λ νμκ΄ μΆμ μμ€ν μ κΈ°μ μ μ리 #곡ν #곡μ #νμκ΄ #μκ³ λ¦¬μ¦ #μ¬μμλμ§
μ°νμ₯κΈ°μ
2K viewsβ’2026-05-29
μ§κ΄ λ° κ³‘κ΄ λ°°κ΄ κ²°ν© κ³ μ μμ #worker #process #fabrication #pipework #clamp
μλμ΄μ΄
2K viewsβ’2026-05-30
Wire To Wire Connection Trick | Strong And Secure Electrical Joint #shortvideo #wireworks
ElectricianTips-b1h
5K viewsβ’2026-06-02
Peterborough to Newark Northgate Driver's Eye View aboard an InterCity 225 - East Coast Main Line
TrainsTrainsTrains
822 viewsβ’2026-05-31
AI turbine design: hypersonic cooling leap #shorts #ai #hypersonic
bobbby_rn
671 viewsβ’2026-05-31
How Far Can A Tomahawk Missile Actually Travel?
WarCurious
13K viewsβ’2026-05-28











