For decades, governments around the world believed they had found a way to protect their most sensitive communications.
Presidents used it.
Diplomats used it.
Military commanders used it.
Intelligence officers used it.
Governments fighting wars used it.
Even the Vatican was reportedly among its customers.
The product was manufactured by a respected Swiss company called Crypto AG, and its reputation rested on something extremely valuable in international politics: trust.
Switzerland was synonymous with neutrality.
Crypto AG was selling encryption technology.
And governments believed the machines were protecting their secrets from precisely the sort of intelligence agencies that desperately wanted to obtain them.
They were wrong.
Behind the Swiss company was one of the most audacious intelligence operations of the Cold War.
The CIA and West Germany's intelligence service, the BND, secretly acquired Crypto AG in 1970. They then controlled the company, influenced the design of its encryption systems and used its products to obtain encrypted communications from governments around the world.
The operation was eventually known as Rubicon.
The CIA's own history reportedly described it in extraordinary terms:
“The intelligence coup of the century.”
And perhaps the most astonishing part was that the victims were not simply being spied on.
They were paying for the equipment that made the spying possible.
The perfect cover was Switzerland
Crypto AG's greatest asset wasn't necessarily its machines.
It was its identity.
The company was Swiss.
For governments shopping for secure communications equipment during the Cold War, that mattered enormously.
The United States and Soviet Union were locked in an existential confrontation. Countries that did not want their communications passing through American or Soviet-controlled technology needed alternatives.
A Swiss company offered something psychologically powerful:
neutrality.
Crypto AG had been founded by Swedish-born cryptographer Boris Hagelin and became an important manufacturer of encryption equipment. By the late 1960s, the company had customers across the world.
That made it enormously attractive to intelligence agencies.
The CIA and BND eventually agreed to secretly purchase the company.
The price was approximately $5.75 million, split between the two intelligence services. The ownership was hidden through a complicated corporate structure involving Liechtenstein and bearer shares, making it extremely difficult for outsiders to determine who actually controlled Crypto AG.
The deception was extraordinary.
The customers believed they were buying Swiss encryption.
They were actually buying equipment controlled by foreign intelligence services.
It wasn't simply a “backdoor”
This distinction is important.
The operation wasn't necessarily about installing a crude mechanism that allowed an intelligence officer to press a button and read every message.
The intelligence agencies wanted something much more sophisticated.
They wanted the encryption to appear secure while containing weaknesses that could be exploited by people who understood how the system had been constructed.
CIA and NSA personnel worked with Crypto AG to influence the technology.
The CIA history described the vulnerabilities as needing to be difficult to detect through normal statistical testing and capable of being explained away as implementation problems or human error if discovered.
In other words, the objective was plausible deniability.
A government buying the machine should have no obvious reason to conclude that its security had been deliberately compromised.
That was the genius of the operation.
The weakness had to be strong enough for American and German cryptanalysts to exploit, but subtle enough that customers wouldn't discover what had happened.
The customers included America's enemies — and its friends
This is where the operation becomes even more extraordinary.
Crypto AG equipment reportedly went to more than 120 countries.
Among its customers were Iran, India, Pakistan, military governments in Latin America and other states across Africa, Asia and the Middle East. Even the Vatican was reportedly a customer.
The operation wasn't limited to America's obvious enemies.
The CIA and BND also obtained communications from countries that were formally aligned with the West.
This created tension inside the operation itself.
The Germans were reportedly uncomfortable with the extent to which the Americans wanted to target allies.
The CIA's priority was different.
From Washington's perspective, encrypted foreign communications were intelligence regardless of whether the country was formally friendly.
That meant NATO membership was not necessarily protection from surveillance.
The numbers are staggering
At the operation's height, Crypto AG wasn't some obscure intelligence source.
It became one of the major pipelines through which Western intelligence obtained foreign communications.
According to the CIA history, Crypto AG equipment accounted for roughly 40 percent of the diplomatic cables and other foreign-government transmissions that NSA cryptanalysts decoded and exploited at certain points in the 1980s.
Think about what that means.
A company selling encryption machines had effectively become an enormous collection point for international intelligence.
The governments using those machines believed they were reducing the possibility of interception.
Instead, their choice of security equipment was helping the NSA and CIA obtain the information.
This was not merely surveillance.
It was supply-chain intelligence.
Iran: one of the biggest prizes
One of the operation's most important targets was Iran.
After the 1979 Iranian Revolution and the subsequent hostage crisis, American intelligence desperately wanted information about the new Iranian government.
Iran used Crypto AG equipment.
That gave American intelligence an extraordinary window into Tehran's communications.
The intelligence value was enormous.
Rather than attempting to penetrate every Iranian communication network from the outside, the United States had helped ensure that some of the equipment carrying those communications was compromised from the beginning.
That is the difference between breaking encryption and controlling the encryption manufacturer.
The second can be vastly more powerful.
The Falklands War produced another extraordinary example
Argentina also relied on Crypto AG equipment.
During the 1982 Falklands War, Argentina became suspicious that its encrypted communications might have been compromised.
The concern was serious enough that Crypto personnel were dispatched to Argentina.
But instead of revealing the real vulnerability, the company reportedly reassured Argentine officials that the main equipment they had purchased remained secure and suggested that British intelligence had probably broken an older system.
The deception worked.
Argentina continued buying Crypto equipment.
Meanwhile, the intelligence operation continued.
That episode illustrates something fundamental about Rubicon:
The intelligence operation depended not merely on technical weaknesses but on maintaining the customer's confidence.
Once customers stopped believing Crypto's security claims, the entire business model would collapse.
The operation survived because almost nobody knew the truth
Even inside Crypto AG, knowledge of the arrangement was tightly compartmentalised.
The CIA and BND did not simply tell the company's hundreds of employees:
“Your machines are compromised and we are spying on your customers.”
Only a tiny number of people understood the real arrangement.
The ownership itself was concealed.
A Liechtenstein law firm helped construct the corporate structure used to hide the intelligence agencies' control. The CIA maintained a covert operational presence in Munich, while the BND handled parts of the financial and corporate arrangements.
The CIA and BND even divided the company's profits.
So the intelligence agencies were doing something particularly audacious:
They were spying on foreign governments and making money from those governments at the same time.
But the secret almost collapsed
Rubicon wasn't flawless.
There were people inside and outside Crypto AG who became suspicious.
Engineers noticed problems.
Customers occasionally complained.
Foreign intelligence services sometimes suspected that the equipment wasn't quite what it appeared to be.
In 1977, an engineer named Peter Frutiger reportedly repaired vulnerabilities in Syrian equipment, causing concern within the intelligence operation because the machines suddenly became harder for NSA analysts to exploit.
The CIA history indicates that officials believed Frutiger had figured out too much about the secret arrangement.
The greatest danger wasn't necessarily a brilliant spy discovering the operation.
It was an engineer asking:
“Why does this machine behave like this?”
That is an important lesson for modern cybersecurity.
Backdoors don't always remain secret because they're technologically perfect.
They remain secret because the people who encounter them don't know what they're looking at.
Then came the Iranian arrest
One of the most dangerous episodes occurred in 1992.
Hans Bühler, a Crypto AG salesman, was arrested by Iranian authorities while doing business in Iran.
He had no knowledge of the secret intelligence arrangement.
The Iranian authorities suspected something was wrong.
Bühler was detained for months before eventually being released.
The episode became a major security crisis for the operation because it threatened to expose Crypto AG's true nature. The CIA history reportedly regarded the incident as the most serious security breach in the programme's history.
The irony is remarkable.
The salesman was not a spy.
He didn't know that the company he worked for was secretly controlled by intelligence agencies.
He was simply selling encryption equipment.
But the company was carrying one of the most consequential secrets in the intelligence world.
West Germany eventually walked away
The Cold War was changing.
The Soviet Union collapsed.
Europe was changing.
The strategic environment that had helped justify the operation was disappearing.
And the BND eventually decided to leave.
West Germany sold its interest in Crypto AG in the early 1990s, reportedly to the CIA.
The CIA then continued controlling the company alone for roughly another quarter-century.
This is where one important correction to the common version of the story is necessary.
It is not quite accurate to say that the CIA remained jointly involved with West German intelligence until 2018.
The BND exited around 1993.
The CIA continued its ownership until approximately 2018, when Crypto AG's remaining assets were divided and sold.
So the operation effectively had two phases:
1970–1993: CIA + BND
1993–2018: CIA control
That makes the story even more extraordinary.
The United States continued the operation long after the original Cold War partnership had ended.
Why 2018 matters
In 2018, Crypto AG was broken up.
Its remaining assets were divided between successor companies.
The operation was effectively over.
But the world didn't know the real story yet.
That came in February 2020.
The Washington Post, German broadcaster ZDF and Swiss broadcaster SRF published investigations based on a classified CIA history and a German intelligence history. The documents revealed the extraordinary scale of the operation.
Suddenly, decades of Cold War history looked different.
Governments that had trusted Crypto AG had unknowingly been purchasing compromised technology.
The Swiss identity that had helped sell the machines had been part of the cover.
And the company had been secretly controlled by the intelligence services of two Western powers.
Even Switzerland knew more than it admitted
The story became even more complicated when Switzerland investigated the affair.
A Swiss parliamentary oversight body concluded in 2020 that the country's intelligence service had known since at least 1993 that foreign intelligence services were behind Crypto AG.
The Swiss authorities themselves had benefited from the arrangement.
The parliamentary investigation concluded that Swiss authorities bore political responsibility for aspects of the affair.
That adds another layer to the story.
Crypto AG wasn't simply a case of America secretly fooling everyone while Switzerland remained completely innocent.
Different intelligence services knew different pieces of the puzzle.
The public, meanwhile, knew almost nothing.
The lesson is bigger than the Cold War
This is why the Crypto AG affair remains relevant in the age of smartphones, cloud computing, artificial intelligence and encrypted messaging.
The central vulnerability wasn't simply encryption.
It was trust in the supplier.
A government can possess an encryption algorithm that is mathematically strong.
But if somebody controls the hardware implementing it, the software running it, the random-number generator producing its keys, the update mechanism or the supply chain, the mathematical strength of the encryption can become almost irrelevant.
Crypto AG demonstrated that decades before today's debates over hardware security, software supply chains and government-mandated access to encrypted communications.
The intelligence agencies didn't necessarily need to defeat mathematics.
They controlled the machinery around the mathematics.
That is a profoundly modern lesson.
The uncomfortable question: who can you trust?
The Crypto AG affair also destroyed one of the most powerful assumptions in international security:
That buying security from a neutral foreign supplier makes you safer.
It doesn't necessarily.
A foreign government may have legal authority over the company.
An intelligence agency may secretly influence it.
Employees may be compromised.
Ownership may be hidden.
Software may contain vulnerabilities.
Hardware may have been modified.
Updates may be manipulated.
And the customer may have absolutely no idea.
In Crypto AG's case, governments weren't merely trusting a company.
They were trusting a company that secretly belonged to the very intelligence agencies from which they were trying to protect themselves.
The intelligence coup of the century
That explains why the CIA's description of Rubicon was so striking.
It wasn't modest.
It wasn't bureaucratic.
The agency reportedly regarded the operation as “the intelligence coup of the century.”
And from an intelligence perspective, it is easy to understand why.
Imagine convincing more than a hundred governments to purchase equipment designed to protect their most sensitive communications — and then discovering that your own intelligence service can exploit that equipment.
Imagine doing it for decades.
Imagine collecting the intelligence while simultaneously earning money from the customers.
And imagine maintaining the secret long enough that the company continued operating until 2018.
That isn't simply a successful espionage operation.
It is a demonstration of the extraordinary power that comes from controlling technology, trust and supply chains simultaneously.
The Crypto AG warning for today's world
There is a temptation to look at Rubicon as an extraordinary relic of the Cold War.
That would be a mistake.
The technology has changed.
The principle hasn't.
Today, governments depend on foreign companies for cloud infrastructure, telecommunications equipment, cybersecurity products, chips, artificial-intelligence systems, operating systems and encrypted communications.
The question is no longer simply:
“Is this technology secure?”
It is:
“Who controls the company that built it?”
“Who controls the code?”
“Who controls the hardware?”
“Who controls the updates?”
“Who has access to the data?”
“And what happens if the company is secretly compromised by a government?”
Crypto AG provides a frightening historical answer.
For decades, governments around the world thought they had bought secrecy.
They had actually bought into somebody else's intelligence operation.
And that may be the most enduring lesson of Rubicon:
The greatest vulnerability in a security system isn't always the technology. Sometimes it is the trust placed in the people who control it.
Comments
Post a Comment