---
title: "Which county in America has the most Superfund sites?"
topic: "Silicon Valley’s Superfund sites"
answer: "Santa Clara County, California: Silicon Valley. It has 21 sites on EPA’s list, against 13 around Houston, and electronics companies left 17 of them. Solvents leaked from buried tanks, sumps and pipes at the chip plants. The cleanup has run for more than 40 years, and EPA expects it to take many more decades."
url: "https://nearfieldnotes.com/notes/silicon-valley-superfund/"
site: "Near Field Notes"
author: "Michael Walker"
note: 31
family: "History"
published: "2026-11-10"
revised: "2026-11-10"
revision: 1
video: "https://nearfieldnotes.com/videos/silicon-valley-superfund/"
---

Note 31 · History · Rev 1 · [Watch the video · 3:20](https://nearfieldnotes.com/videos/silicon-valley-superfund/)

Silicon Valley’s Superfund sites

# Which county in America has the most Superfund sites?

**Short answer:** Santa Clara County, California: Silicon Valley. It has 21 sites on EPA’s list, against 13 around Houston, and electronics companies left 17 of them. Solvents leaked from buried tanks, sumps and pipes at the chip plants. The cleanup has run for more than 40 years, and EPA expects it to take many more decades.

**Note 31’s opening figure: Superfund sites by county.** The seven counties with the most

![Note 31’s opening figure: Superfund sites by county](https://nearfieldnotes.com/notes/silicon-valley-superfund/fig-opening.png)

*What the figure shows:* Bars of the counties with the most sites on EPA’s Superfund list: Santa Clara, California 21, Los Angeles, California 18, Montgomery, Pennsylvania 17, Middlesex, New Jersey 15, Nassau, New York 14, Middlesex, Massachusetts 14, Harris, Texas 13. Harris County is Houston. Santa Clara County is Silicon Valley, and electronics companies left 17 of its 21.

**Key**

- Left by electronics companies
- Other sites

Sites on EPA’s National Priorities List on 30 September 2026. Proposed and deleted sites are not counted.

## The valley before chips

*Santa Clara Valley, 1948 to 1953*

Superfund is the federal program for the country’s worst hazardous-waste sites, and its National Priorities List names them. Seventy years ago nothing here belonged on such a list. This was the Valley of Heart’s Delight, with about 106,000 acres of fruit and nut trees around 1940, more than half of them prunes.

In 1956 William Shockley opened a transistor lab in an old fruit-packing shed in Mountain View. Eight of his staff left to start Fairchild Semiconductor in 1957. Fairchild’s people went on to start Intel in 1968 and AMD in 1969, and the chip plants went up where the orchards had been.

**Fig. 31: The valley in 1948 and 1953.**

![Fig. 31: The valley in 1948 and 1953](https://nearfieldnotes.com/notes/silicon-valley-superfund/fig-31.png)

*What the figure shows:* Figure 31: the valley on USGS maps of 1948 and 1953, from Palo Alto to San Jose, most of it orchard. Dots mark 20 of today’s 21 Superfund sites, between Palo Alto and San Jose. A square marks Shockley’s transistor lab of 1956 in Mountain View.

Two USGS sheets: 1948 west of the seam and 1953 east of it. Green dots are orchards, and pink is built-up land. 20 of the 21 sites fall on this map. The other, Fairchild Semiconductor, South San Jose, is past its southeast corner.

## Why chip plants kept solvents

*The chemistry*

Making a chip means cleaning a silicon wafer again and again, because one speck of grease can ruin it. The plants wanted solvents that cut grease and would not catch fire, and chlorinated solvents did both. The main one was trichloroethylene, or TCE. From the 1960s many plants moved to a close relative, TCA.

**Figure: What each chemical did.**

Fig. 31b

| Chemical | Job in the plant | Found in the ground at |
|---|---|---|
| TCE | Degreasing wafers and parts | Fairchild, Raytheon and Intel in Mountain View |
| TCA | Cleaning and degreasing, in place of TCE | Fairchild in South San Jose, IBM |
| Freon 113 | Cleaning, and removing flux | Applied Materials, Intel in Santa Clara |
| PCE | Cleaning | AMD, Applied Materials, Intersil |
| Acetone, alcohol, xylene | Photoresist work and cleaning | The same waste tanks as the solvents above |
| Acids | Etching and cleaning | Neutralizing tanks and sumps |

Jobs from EPA’s 1995 notebook on the electronics industry. Places from EPA’s listing of each site. TCE is trichloroethylene, TCA is 1,1,1-trichloroethane and PCE is tetrachloroethylene.

Chlorinated solvents

### They cut grease, they do not burn, and they sink

**Field marks.** TCE, TCA and PCE dissolve grease and do not catch fire. All three are heavier than water. TCE weighs 1.46 times as much.

**Look-alike.** TCA took over from TCE as the safer choice. It was later phased out for damaging the ozone layer.

**Range.** EPA and the International Agency for Research on Cancer both class TCE as a cause of cancer in people, on the evidence for kidney cancer. The drinking-water limit is 5 parts per billion for TCE and 200 for TCA.

**Habitat.** Local fire codes had flammable chemicals stored underground, and the solvents and their waste went into buried tanks too. The tanks had no leak detectors.

From the street the plants showed lawns and no smokestacks. Janet Gray Hayes, San Jose’s mayor in those years, said later: “There was no doubt in my mind that this was a clean industry.”

## 1981: a tank leaks in South San Jose

*Fairchild Semiconductor, 101 Bernal Road*

In late 1981 Fairchild found that a buried waste-solvent tank at its South San Jose plant had failed. The tank held 5,940 gallons. About 60,000 gallons of solvent and water had gone into the ground, because the plant kept refilling it for at least a year and a half. The solvent was mostly TCA.

**Fig. 31c: The 1981 leak.**

![Fig. 31c: The 1981 leak](https://nearfieldnotes.com/notes/silicon-valley-superfund/fig-31c.png)

*What the figure shows:* Figure 31c: a cutaway of the 1981 leak, not to scale. A buried 5,940-gallon waste-solvent tank under Fairchild’s South San Jose plant leaks down to the water table. The solvent moves with the groundwater to drinking-water well GO-13, about 1,800 ft away.

A drawing, not to scale.

**Fig. 31d: The leak in two numbers.**

![Fig. 31d: The leak in two numbers](https://nearfieldnotes.com/notes/silicon-valley-superfund/fig-31d.png)

*What the figure shows:* Figure 31d: the 1981 leak in two numbers. The tank held 5,940 gallons, and about 60,000 gallons of solvent and water leaked out. Well GO-13 tested at 5,800 parts per billion of TCA against a limit of 200.

Gallons from EPA’s 2024 review and the National Research Council’s 1988 report. The well was tested in December 1981.

The 1981 leak

### About ten tankfuls went into the ground

**Field marks.** About 1,800 ft away, drinking-water well GO-13 tested at 5,800 parts per billion of TCA, 29 times the state’s limit. The water company shut it on 7 December 1981.

**Look-alike.** This leak was TCA. The TCE sites are in Mountain View, about 19 miles to the northwest.

**Range.** The plant made chips for six years, from 1977 to 1983. A wall was built underground around the source in 1986, and the site is still on the list.

**Habitat.** It was not the first. IBM had reported a solvent leak at its disk-drive plant nearby in October 1980, and Hewlett-Packard found a leaking tank in Palo Alto in July 1981.

Neighbors in Los Paseos, across Bernal Road from the plant, counted birth defects and miscarriages among themselves and took the count to the city. California’s health department studied it. It found a real excess in that neighborhood in 1980 and 1981: about twice the miscarriages and three times the birth defects of a similar neighborhood on other water. Its later studies modeled where the well’s water had gone and concluded that the leak was unlikely to be the cause. A lawsuit by 530 residents was settled in 1986, and the terms were sealed.

## Everyone else checks

*Santa Clara Valley, 1982 to 1985*

After Fairchild, the regional water board asked the valley’s other plants about their tanks. A survey counted 1,692 underground tanks at 388 sites. Of the 96 sites with solvent tanks, 75 had contaminated the ground. “We were batting 1,000,” the water board’s Wil Bruhns said later. “Anyone who looked for leaks found them.”

**Fig. 31e: Sites with solvent tanks.**

![Fig. 31e: Sites with solvent tanks](https://nearfieldnotes.com/notes/silicon-valley-superfund/fig-31e.png)

*What the figure shows:* Figure 31e: 96 squares, one for each site with underground solvent tanks in the 1980s survey. 75 are filled: contamination was found there. 21 are empty.

One square for each of the 96 sites. From Olivieri and others, 1985.

The tank survey

### 75 of 96 had leaked

**Field marks.** The leaks came from tanks, from acid-neutralizing sumps and from pipes. At Intel’s first plant, acid had eaten through the bottoms of the sewer pipes.

**Look-alike.** A sound tank could still leak. Crews overfilled tanks, drained delivery hoses onto the ground and mixed chemicals that attacked a tank’s walls.

**Range.** In 1983 the county’s cities adopted a storage ordinance for hazardous materials, and California passed the country’s first law on underground tanks. New tanks needed a second wall and leak monitoring.

**Habitat.** On 15 October 1984 EPA proposed 19 South Bay sites for the Superfund list in a single day.

## The 21 sites

*Santa Clara County, 2026*

The list holds 21 sites today, and most of them follow US-101 from Palo Alto to Santa Clara. Electronics factories account for 17, most of them chip plants. The other 4 are a Navy air station, a transformer plant, a drum recycler and a group of asbestos landfills.

Los Angeles County is next with 18. The count says how many sites EPA has listed. It is not a measure of how polluted a place is.

**Fig. 31f: Where the sites are.**

![Fig. 31f: Where the sites are](https://nearfieldnotes.com/notes/silicon-valley-superfund/fig-31f.png)

*What the figure shows:* Figure 31f: an aerial photograph of the Santa Clara Valley with the county’s 21 Superfund sites marked and numbered as in the table. They follow US-101 from Palo Alto through Mountain View, Sunnyvale and Santa Clara, with two more in San Jose. 17 were left by electronics companies.

Numbers match the table below. Places from EPA’s list, on USGS aerial photographs.

**Figure: Each site, then and now.**

Fig. 31g

| No. | Site | City | Listed | Then | Now |
|---|---|---|---|---|---|
| 1 | [Hewlett-Packard, 620–640 Page Mill Road](https://cumulis.epa.gov/supercpad/cursites/csitinfo.cfm?id=0902134) | Palo Alto | 1990 | Optoelectronics plant, 1962 to 1986 | An office building with a vapor barrier and a ventilated garage |
| 2 | [CTS Printex](https://cumulis.epa.gov/supercpad/cursites/csitinfo.cfm?id=0901206) | Mountain View | 1990 | Printed circuit boards, 1966 to 1985 | Townhomes, finished in 2010, with vapor controls |
| 3 | [Teledyne Semiconductor](https://cumulis.epa.gov/supercpad/cursites/csitinfo.cfm?id=0901134) | Mountain View | 1987 | Chip plant from 1962 | Offices and labs |
| 4 | [Spectra-Physics](https://cumulis.epa.gov/supercpad/cursites/csitinfo.cfm?id=0900252) | Mountain View | 1991 | Lasers and electronics from 1961 | A church and offices |
| 5 | [Fairchild Semiconductor, Mountain View](https://cumulis.epa.gov/supercpad/cursites/csitinfo.cfm?id=0901680) | Mountain View | 1991 | Fairchild’s chip plants | Google’s former Quad campus, unoccupied in 2025 |
| 6 | [Raytheon](https://cumulis.epa.gov/supercpad/cursites/csitinfo.cfm?id=0901202) | Mountain View | 1986 | Semiconductor plant, 1961 to 1997 | An office campus leased to OpenAI in 2026 |
| 7 | [Intel, Mountain View](https://cumulis.epa.gov/supercpad/cursites/csitinfo.cfm?id=0901482) | Mountain View | 1986 | Intel’s first chip plant, 1968 to 1981 | Older office and lab buildings |
| 8 | [Moffett Naval Air Station](https://cumulis.epa.gov/supercpad/cursites/csitinfo.cfm?id=0902734) | Mountain View, Sunnyvale | 1987 | Navy air station | NASA Ames and its research park |
| 9 | [Westinghouse Electric](https://cumulis.epa.gov/supercpad/cursites/csitinfo.cfm?id=0900956) | Sunnyvale | 1986 | Transformer plant | Still a factory, now Northrop Grumman’s |
| 10 | [Advanced Micro Devices, Building 915](https://cumulis.epa.gov/supercpad/cursites/csitinfo.cfm?id=0902708) | Sunnyvale | 1990 | Chip fab and development center, 1974 to 2009 | The Vale: about 450 townhomes, with a vapor barrier and vents under every building |
| 11 | [TRW Microwave, Building 825](https://cumulis.epa.gov/supercpad/cursites/csitinfo.cfm?id=0901181) | Sunnyvale | 1990 | Microwave and semiconductor parts, 1968 to 1993 | A lab and office building leased to Apple |
| 12 | [Advanced Micro Devices, 901 and 902 Thompson Place](https://cumulis.epa.gov/supercpad/cursites/csitinfo.cfm?id=0901389) | Sunnyvale | 1986 | Chip fab, 1969 to 1992 | A self-storage building |
| 13 | [Monolithic Memories](https://cumulis.epa.gov/supercpad/cursites/csitinfo.cfm?id=0901398) | Sunnyvale | 1987 | Chip fab, 1970 to 1987 | Commercial buildings, one a daycare |
| 14 | [National Semiconductor](https://cumulis.epa.gov/supercpad/cursites/csitinfo.cfm?id=0901326) | Santa Clara | 1987 | National’s chip campus from 1967 | Texas Instruments offices and labs |
| 15 | [Intel Magnetics](https://cumulis.epa.gov/supercpad/cursites/csitinfo.cfm?id=0901661) | Santa Clara | 1986 | Bubble memory, 1978 to 1987 | Offices |
| 16 | [Applied Materials, Building 1](https://cumulis.epa.gov/supercpad/cursites/csitinfo.cfm?id=0901344) | Santa Clara | 1987 | Chip-equipment plant, 1974 to 2003 | Offices |
| 17 | [Synertek, Building 1](https://cumulis.epa.gov/supercpad/cursites/csitinfo.cfm?id=0902620) | Santa Clara | 1989 | Chip fab from 1974, closed 1985 | Vacant in 2022 |
| 18 | [Intersil and Siemens Components](https://cumulis.epa.gov/supercpad/cursites/csitinfo.cfm?id=0901325) | Cupertino | 1990 | Chip and LED plants, 1968 to 1995 | Offices used by Apple and Panasonic, and medical offices |
| 19 | [South Bay Asbestos Area](https://cumulis.epa.gov/supercpad/cursites/csitinfo.cfm?id=0902250) | San Jose (Alviso) | 1986 | Landfills of asbestos-cement pipe waste | Offices, hotels and a mobile-home park on capped landfill |
| 20 | [Lorentz Barrel & Drum](https://cumulis.epa.gov/supercpad/cursites/csitinfo.cfm?id=0901287) | San Jose | 1989 | Drum recycler, 1947 to 1987 | Parking and a recycling plant |
| 21 | [Fairchild Semiconductor, South San Jose](https://cumulis.epa.gov/supercpad/cursites/csitinfo.cfm?id=0901685) | San Jose | 1989 | Device plant, 1977 to 1983 | A shopping center |

Each name links to EPA’s page for the site. Listed is the year the site went on the list for good. Now is from each site’s latest EPA review or record. The list once held two more sites, since deleted: Intel’s Santa Clara III plant in 2019 and JASCO Chemical in 2020.

## A cleanup that cannot finish

*Middlefield-Ellis-Whisman, Mountain View*

The largest cleanup is in Mountain View, where Fairchild, Raytheon and Intel ran chip plants a few blocks apart and their solvent merged into one plume. The companies dug out 36,000 cubic yards of soil, built four walls underground, and began pumping the groundwater up to clean it.

They have pumped about 8 billion gallons, enough to fill 12,000 Olympic pools, and recovered about 126,000 lb of solvent.

**Fig. 31h: Under Mountain View.**

![Fig. 31h: Under Mountain View](https://nearfieldnotes.com/notes/silicon-valley-superfund/fig-31h.png)

*What the figure shows:* Figure 31h: the ground under the Mountain View site, to scale in depth. Water-bearing zones: A from 0 to 45 ft, B1 from 50 to 75 ft, B2 from 75 to 110 ft, B3 from 120 to 160 ft, C from 200 to 240 ft. A clay layer lies from 160 to 200 ft. Solvent is mainly in the top three zones, down to 110 ft. Underground walls reach 40 ft and 100 ft, and wells pump from the zones that hold solvent.

Depths to scale, from EPA’s 2024 review. Widths are not to scale. Three of the four walls reach 40 ft, and Raytheon’s reaches 100 ft.

Pumping worked well at first and less well each year. TCE is heavier than water, so it sinks through sand until it meets clay. It soaks into the clay and seeps back out for decades after the sand beside it is clean.

**Fig. 31i: What the pumps bring up.**

![Fig. 31i: What the pumps bring up](https://nearfieldnotes.com/notes/silicon-valley-superfund/fig-31i.png)

*What the figure shows:* Figure 31i: pounds of solvent recovered per million gallons pumped at the Mountain View site, by period. Start to 2008: 21. 2009 to 2013: 9.9. 2014 to 2018: 9.4. 2019 to 2023: 6.7 to 8.8.

Solvent removed divided by water pumped, from EPA’s reviews. The light part of the last bar is the spread between the two totals EPA gives for 2019 to 2023.

Pump and treat

### Each gallon brings up less

**Field marks.** Through 2008 a million gallons brought up about 21 lb of solvent. In 2024 the three main systems pumped 141 million gallons for 919 lb, about 6.5 lb per million.

**Look-alike.** Cleaner water at the pump is not a clean aquifer. The solvent that remains is held in clay, where pumping reaches it slowest.

**Range.** 89 wells feed eight treatment systems at about 424 gallons a minute. NASA has also planted more than 1,000 poplars north of the highway to draw the plume up.

**Habitat.** The water now carries PFAS as well. One system’s intake holds PFOS at 30 to 40 times the 2024 drinking-water limit.

The worst well read 1,000,000 parts per billion of TCE in the 1980s, close to the most that water can dissolve. In 2022 the worst read 8,800. That is a fall of more than 99%, and it is still 1,760 times the limit.

**Fig. 31j: The worst well, then and now.**

![Fig. 31j: The worst well, then and now](https://nearfieldnotes.com/notes/silicon-valley-superfund/fig-31j.png)

*What the figure shows:* Figure 31j: TCE in groundwater at the Mountain View site on a log scale. The worst well read 1,000,000 parts per billion in the 1980s and 8,800 in 2022. The limit is 5.

A log scale: each step is ten times the one before. From EPA’s 1989 plan and 2024 review.

**Fig. 31k: How long.**

![Fig. 31k: How long](https://nearfieldnotes.com/notes/silicon-valley-superfund/fig-31k.png)

*What the figure shows:* Figure 31k: the Mountain View cleanup against the years. Pumping has run from 1982 to 2026. The 1989 plan allowed as long as 46 years, to 2035, or into the indefinite future. EPA’s 2019 review found the improving wells would not reach the goal within 100 years, which is past 2119.

Each estimate starts in the year it was made. A dotted line means the estimate has no end date.

The schedule

### No estimate has an end date

**Field marks.** The 1989 plan said the shallow pumping “may be in operation for as long as 46 years or into the indefinite future.”

**Look-alike.** No official source says hundreds of years. EPA’s 2019 review fitted trend lines to the wells and found that the improving ones “will not reach cleanup goals within 100 years under current conditions.”

**Range.** In that review 11 of 18 wells sampled showed TCE that was not falling. The 2024 review says the remedy will not reach its goal “for many more decades” and calls for new methods.

**Habitat.** The 2024 federal ban on TCE allows cleanup sites 50 more years to dispose of the water they pump.

## Up through the floor

*Vapor intrusion*

Solvent in shallow groundwater also evaporates. The vapor rises through the soil and into whatever stands on top, through cracks and pipe openings in the floor slab. In Mountain View it comes from the top 5 to 35 ft of the plume. The usual fix is a fan that pulls the vapor from under the slab and vents it above the roof.

**Fig. 31m: How vapor gets in.**

![Fig. 31m: How vapor gets in](https://nearfieldnotes.com/notes/silicon-valley-superfund/fig-31m.png)

*What the figure shows:* Figure 31m: a cutaway of a building over solvent in shallow groundwater, not to scale. The solvent evaporates and rises through the soil. On the left, with no controls, the vapor comes up through the floor slab. On the right, a fan pulls it from under the slab and vents it above the roof.

A drawing, not to scale.

**Fig. 31l: Three plants, one plume.**

![Fig. 31l: Three plants, one plume](https://nearfieldnotes.com/notes/silicon-valley-superfund/fig-31l.png)

*What the figure shows:* Figure 31l: an aerial photograph of the Mountain View site. Three old chip plants, Fairchild’s, Raytheon’s and Intel’s, sit south of US-101. EPA’s boundary for the solvent in the groundwater runs north from them, under the highway and onto Moffett Field. It covers about 0.74 square miles.

EPA’s boundary for the Mountain View site covers 0.74 square miles. Groundwater here flows north, toward the Bay.

What stands on top

### Offices, homes and schools now sit over the plumes

**Field marks.** In November 2012, after a change to the ventilation in two Google buildings on Fairchild’s old plant, TCE in the indoor air rose above EPA’s level for offices. Fans under the slabs have run since 2014.

**Look-alike.** Google’s campus there stood unoccupied in 2025. In March 2026 OpenAI leased the old Raytheon plant next door, which sits inside the 100 ft underground wall.

**Range.** In Sunnyvale, AMD’s Building 915 is now about 450 townhomes, with a vapor barrier and vents under every building. Nearby, one plume lies under more than 500 homes and three school campuses. In Cupertino, Apple has offices on the old Intersil lot, across Tantau Avenue from Apple Park.

**Habitat.** The South San Jose plant that leaked in 1981 is a shopping center.

EPA treats TCE vapor as urgent because of one short-term risk: heart defects, which can form in about three weeks early in a pregnancy. Its regional office asks for a fix within weeks when a home reads 2 micrograms per cubic meter, and within days at 6. For an eight-hour workday the levels are 8 and 24.

The long-term levels in Mountain View are lower still: 1 for homes and 5 for offices. News reports of EPA’s 2012 samples in the Google buildings put most stations under 5, about a dozen between 5 and 30, and one at 120.

**Fig. 31n: TCE in indoor air.**

![Fig. 31n: TCE in indoor air](https://nearfieldnotes.com/notes/silicon-valley-superfund/fig-31n.png)

*What the figure shows:* Figure 31n: TCE in indoor air, in micrograms per cubic meter. EPA’s long-term level at the Mountain View site is 1 for homes and 5 for offices. The highest reading reported in Google’s Quad buildings in late 2012 was 120.

Levels from EPA’s 2010 decision for the Mountain View site. The reading is the highest of EPA’s samples as CBS reported them in February 2013.

## Ten thousand wells

*Santa Clara County, 2026*

All of it is watched through wells. The state’s records hold 10,002 monitoring wells in Santa Clara County, and 2,406 had a water level measured in 2024 or later. EPA’s own sites keep hundreds more that the state’s files leave out.

The companies still pay. Fairchild’s share goes to Schlumberger, an oilfield-services company that bought Fairchild in 1979, sold it in 1987 and kept the cleanup. Raytheon’s goes to RTX and National Semiconductor’s to Texas Instruments. Intel pays its own, and the Navy and NASA pay for Moffett Field.

The tap water is a separate matter. Nobody drinks the shallow groundwater. EPA’s 2024 review says drinking water in Mountain View comes from the Hetch Hetchy reservoir and local sources, and is treated to meet state and federal standards. A thick layer of clay lies between the shallow plumes and the deep aquifers that supply wells.

**Fig. 31o: Every monitoring well on record.**

![Fig. 31o: Every monitoring well on record](https://nearfieldnotes.com/notes/silicon-valley-superfund/fig-31o.png)

*What the figure shows:* Figure 31o: an aerial photograph of Santa Clara County with the 10,405 monitoring-well points in the state’s GeoTracker records. They crowd the valley floor from Palo Alto to San Jose, with a second cluster near Morgan Hill. 1,884 had a water level measured in 2024 or later and are drawn in yellow.

The 10,405 well points with a position in the State Water Board’s GeoTracker files. 1,884 of the wells measured since 2024 have a position and are drawn in yellow.

## Is there a site near you?

*Anywhere in the United States*

Santa Clara County has the most listed sites, and it is not the only place with them. EPA and the states publish where every site is, and the records are free to search.

The story in dates

1. 1956 Shockley’s lab opens in a fruit-packing shed in Mountain View
2. 1957 Fairchild Semiconductor starts
3. 1968 Intel starts, with a plant in Mountain View
4. 1971 A trade paper names the place Silicon Valley
5. 1980 IBM reports a solvent leak in South San Jose
6. 1981 Fairchild’s tank leak is found, and well GO-13 is shut
7. 1983 The county’s storage ordinance and California’s tank law
8. 1984 EPA proposes 19 South Bay sites in one day
9. 1989 The Mountain View plan: 46 years, or the indefinite future
10. 2012 TCE vapor is found in Google’s buildings
11. 2019 EPA’s review: not within 100 years
12. 2024 EPA issues a ban on TCE. Parts of it have since been delayed.
13. 2026 OpenAI leases the old Raytheon plant

Key to a site near you

### Is there a site under my building?

1. 1a [EPA’s Superfund search](https://www.epa.gov/superfund/search-superfund-sites-where-you-live), by address Federal sites
2. 1b The state’s cleanup map. In California it is [GeoTracker](https://geotracker.waterboards.ca.gov/map/) State cases and their wells
3. 1c Your water supplier’s yearly quality report What is in the tap water

Each Superfund site’s page lists its five-year reviews. A review says what was found, what is being done, and who is paying.

#### Conditions

- **The count:** Sites with the status “NPL Site” in EPA’s National Priorities List layer, as edited on 30 September 2026 and counted on 9 November 2026. A site that spans two counties counts in each.
- **Electronics:** A site counts as electronics if a maker of chips, circuit boards, lasers, microwave parts or chip equipment ran it, by EPA’s listing.
- **Then and now:** From each site’s EPA listing and its latest five-year review or record, read in November 2026
- **The leak:** EPA’s 2024 review of the South San Jose site and the National Research Council’s 1988 report
- **Mountain View:** EPA’s 1989 plan and its 2014, 2019 and 2024 reviews. The total pumped through 2023 is 8.0 to 8.3 billion gallons, because the 2024 review gives two totals for 2019 to 2023.
- **Wells:** State Water Board GeoTracker files for the county, built on 8 November 2026
- **Units:** In water, 1 part per billion is 1 microgram per liter. An Olympic pool is taken as 660,000 gallons.
- **Maps:** Aerial photographs from the USGS base map, as served in November 2026, with the Bay and the ocean drawn from the USGS National Hydrography Dataset. Site boundaries from EPA, US-101 from OpenStreetMap, and the USGS Palo Alto sheet of 1948 and San Jose sheet of 1953, both at 1:62,500.

Sourced, not measured. Every number here comes from a public record, and nothing was simulated.

#### Limits

- A count of listed sites is not a ranking of pollution. It also reflects how closely a place was studied in the 1980s.
- The 60,000 gallons is solvent and water together. No source gives the solvent alone with confidence.
- The state’s studies found that the cluster of birth defects was real and that the leak was unlikely to be its cause. They did not find another cause.
- The 120 reading in the Google buildings is from a news report of EPA’s samples. EPA’s own record says only that readings were above its level for offices.
- The fire-code reason for burying the tanks rests on one historian’s account and one news report.
- What stands on a site can change. The table’s Now column is as of November 2026.
- The state’s files leave out most wells at the sites EPA leads, so the county’s true count is higher.

#### Files

- refs/sc_npl_sites.csv Every site, with its EPA number, dates and links
- refs/sites_now.csv What stands on each site, with its source
- refs/npl_county_ranking.csv The count for every county
- data/note_facts.json Each number the figures draw, with its source

#### Sources

- [EPA, National Priorities List sites with status](https://services.arcgis.com/cJ9YHowT8TU7DUyn/arcgis/rest/services/Superfund_National_Priorities_List_(NPL)_Sites_with_Status_Information/FeatureServer/0)
- [EPA, Superfund site boundaries](https://services.arcgis.com/cJ9YHowT8TU7DUyn/arcgis/rest/services/FAC_Superfund_Site_Boundaries_EPA_Public/FeatureServer/0)
- [EPA, seventh five-year review of Fairchild South San Jose, 2024](https://semspub.epa.gov/work/09/100038400.pdf)
- [National Research Council, Hazardous Waste Site Management, 1988, chapter 7](https://nap.nationalacademies.org/read/1063/chapter/9)
- [Deane and others, 1989, pregnancy outcomes near the leak](https://pubmed.ncbi.nlm.nih.gov/2705433/)
- [Swan and others, 1989, heart defects in Santa Clara County](https://pubmed.ncbi.nlm.nih.gov/2784935/)
- [Wrensch and others, 1990, the follow-up study](https://pubmed.ncbi.nlm.nih.gov/2296981/)
- [Washington Post, 5 July 1986, the settlement](https://www.washingtonpost.com/archive/national/1986/07/05/multimilliondollar-settlement/5de9dd16-0157-49bf-bc44-4b6d9c2301e9/)
- [Olivieri and others, 1985, the tank survey](https://doi.org/10.1061/(ASCE)0733-9496(1985)111:3(346))
- [San Francisco Chronicle, 30 January 2004, the valley’s toxic history](https://www.sfgate.com/bayarea/article/The-valley-s-toxic-history-IBM-trial-is-latest-2826844.php)
- [State Water Board, underground tank law timeline](https://www.waterboards.ca.gov/ust/single-walled-tanks-progress/timeline-faqs.html)
- [EPA, the Middlefield-Ellis-Whisman study area](https://www.epa.gov/superfund/mew-study-area)
- [EPA, 1989 record of decision for Mountain View](https://semspub.epa.gov/src/document/09/88167245)
- [EPA, 2010 amendment on vapor intrusion](https://semspub.epa.gov/src/document/09/1163792)
- [EPA, fourth five-year review of Mountain View, 2019](https://semspub.epa.gov/src/document/09/100018492)
- [EPA, fifth five-year review of Mountain View, 2024](https://semspub.epa.gov/src/document/09/100038556)
- [Geosyntec for Schlumberger, 2024 annual progress report](https://semspub.epa.gov/src/document/09/100040824)
- [EPA record on the former Fairchild buildings, 2025](https://semspub.epa.gov/src/document/09/100047981)
- [EPA Region 9, 2014 memo on TCE response levels](https://semspub.epa.gov/work/05/928381.pdf)
- [CBS San Francisco, 22 February 2013, vapor under Google’s buildings](https://www.cbsnews.com/sanfrancisco/news/toxic-vapors-seeping-up-from-underneath-google-buildings/)
- [NBC Bay Area and the Center for Investigative Reporting, 20 March 2013](https://www.nbcbayarea.com/news/local/google-employees-face-health-risks-from-superfund-sites-toxic-vapors/2049435/)
- [KREF and TMG Partners, 23 March 2026, the lease to OpenAI](https://investor.wedbush.com/wedbush/article/bizwire-2026-3-23-kref-and-tmg-partners-announce-lease-of-entire-350-380-ellis-office-campus-to-openai-in-mountain-view)
- [EPA, five-year review of the Sunnyvale Triple Site, 2024](https://semspub.epa.gov/src/document/09/100038437)
- [EPA, five-year review of AMD Building 915, 2024](https://semspub.epa.gov/src/document/09/100038406)
- [EPA, five-year review of Intersil and Siemens, 2020](https://semspub.epa.gov/work/09/100021555.pdf)
- [EPA, five-year review of Teledyne and Spectra-Physics, 2024](https://semspub.epa.gov/src/document/09/100038411)
- [EPA, Sector Notebook on the electronics and computer industry, 1995](https://www.clu-in.org/download/toolkit/elecmpsn.pdf)
- [EPA IRIS, trichloroethylene](https://iris.epa.gov/ChemicalLanding/&substance_nmbr=199)
- [IARC Monographs, volume 106](https://publications.iarc.who.int/130)
- [EPA, final rule on TCE, 89 FR 102568, December 2024](https://www.federalregister.gov/citation/89-FR-102568)
- [Lécuyer, 2017, on the valley’s chemicals and tanks](https://hal.science/hal-03990797)
- [Heppler, 2017, The Conversation](https://theconversation.com/how-silicon-valley-industry-polluted-the-sylvan-california-dream-85810)
- [Christian Science Monitor, 4 December 1984](https://www.csmonitor.com/1984/1204/120454.html)
- [Santa Clara County crop reports, 1940 to 1949](https://ageconsearch.umn.edu/record/99346/files/SANTA%20CLARA%201940-1949.pdf)
- [IEEE Milestone, Birthplace of Silicon Valley, 1956](https://ethw.org/Milestones:Birthplace_of_Silicon_Valley,_1956)
- [Computer History Museum, Who named Silicon Valley?](https://computerhistory.org/blog/who-named-silicon-valley/)
- [State Water Board, GeoTracker](https://geotracker.waterboards.ca.gov/map/)
- [Valley Water, Where Does Our Water Come From?](https://assets.valleywater.org/files/2022%20Where%20Does%20Our%20Water%20Come%20From.pdf)
- [USGS, historical topographic maps](https://ngmdb.usgs.gov/topoview/)
- [USGS, The National Map, aerial photographs (USDA flights)](https://basemap.nationalmap.gov/arcgis/rest/services/USGSImageryOnly/MapServer)
- [USGS, National Hydrography Dataset](https://www.usgs.gov/national-hydrography/national-hydrography-dataset)
- [OpenStreetMap contributors](https://www.openstreetmap.org/copyright)

What should the next note be? [Ask for one](https://nearfieldnotes.com/#ask-for-a-note)

---

Source: https://nearfieldnotes.com/notes/silicon-valley-superfund/

Cite as: Michael Walker, “Silicon Valley’s Superfund sites”, Near Field Notes, Note 31, Rev 1, 10 Nov 2026. https://nearfieldnotes.com/notes/silicon-valley-superfund/

Figures: free to reuse under CC BY 4.0 (https://nearfieldnotes.com/license/). Each one's image is linked under its name. Credit Michael Walker, Near Field Notes, and link the figure: this address plus #fig- and its number, like https://nearfieldnotes.com/notes/silicon-valley-superfund/#fig-31.
