Oh, quite, quite a few!

• I suppose Java must be called out as, likely, by far the most pervasive language, with which most services have been written, across both AWS and other organizations.
• Kotlin is seeing a lot of adoption, which I’m personally both pushing for and happy about, as a clear replacement to Java. Most of my new code is written in Kotlin, these days!
• For some, specific use cases, we are seeing Rust being adopted, more and more.
• I’m pretty sure there’s still a fairly large C++ codebase, although, sadly, I doubt that it’s being heavily invested into anymore.
• To my delight, C# is being used more and more, though, sadly, I haven’t written a line of it in Amazon.
• We have a huge codebase of tools and other applications that are written in Python. Many a Jupyter notebook is written in Python, for diverse use cases, some, of course, involving Machine Learning. Various forms of automation, including dealing with AWS resources (i.e. boto) also comes to mind, as well as Lambda function code, or even Redshift’s UDF’s.
• We use TLA+ for formal verification of our algorithms, particularly in AWS. I happen to use it quite a bit and really enjoy it.
• We have a huge codebase of tools and other specific applications that are written in Ruby.
• As our engineers (SDE’s) are also heavily involved in operations, everyone is quite familiar with the standard Linux toolkit and capable of writing Shell scripts in, say, Zsh or Bash.
• We do use SQL to interact with databases, such as Redshift - my personal favorite.
• I’ve personally used Scala, though I haven’t seen it being used a lot, outside of Spark(ML).
• There are a few appearances of Clojure, here and there, mainly related to Jepsen tests, in my experience.
• Perl is still alive and well and commonly occuring in some tools and scripts, as well as some even some internal Web apps, I believe.

I’m pretty sure there are other, sporadic instances of packages being written in Go, Haskell, ML, R and many others, but I personally haven’t seen many of those. Perhaps it’s also worth mentioning that we use a lot of JSON, XML and YAML as representation formats for various types of data.

Disclaimer: This answer is not authoritative or complete in any way, shape or form; it’s based solely on the observations of a single developer who’s worked in only a few groups.

There are few mainstream languages that aren’t being used to develop software at Amazon. Generally people try to use the right tool for the job.

Java is by far the most common language for Amazon software. I don’t know how much of the code base it makes, but it’s definitely a large majority. A huge part of our software is web services, and AFAIK Java was the first mainstream language that combined excellent support for web services with the stability and maintainability that is needed for a large code base. There are alternatives now — Golang, C#, and others — but at this point Java is pretty well established as the Amazon standard, so switching would require a very compelling reason.

Clojure and Scala have seen their popularity wax and wane with time; there’s always a vocal minority that advocates for functional programming and it tends to be people who can actually deliver on the promise. I wish I could say I was one of them, but I’m not.

Kotlin, another JVM language, is gaining popularity and even some “institutional support” lately. It’s obviously becoming very popular in Android, which Amazon has quite a few people building apps for.
I played around a bit with Kotlin and think it’s a very good replacement for Java. Unlike most other languages in this list, Kotlin has a realistic (though still small IMO) chance to unseat Java as the Queen of the Amazons in the foreseeable future, because the transition can be gradual.

While we’re on the subject of JVM languages, Groovy is also used occasionally, but AFAIK not for anything serious (other than in the context of Gradle).

C++ is used in some places, although it’s mostly niche at this point: low-level and system code, machine learning, graphics, and other performance-sensitive or hardware-specific stuff.

C# has a niche, mostly for Windows software, but I can’t tell you much about that.

iPhone languages — ObjectiveC and Swift — are obviously use to develop iPhone apps. I’m personally not aware of any other use, but I wouldn’t be surprised if Swift has some adherents outside that world.

Javascript is, as we all know, the lingua franca of the client world, and as such is found in pretty much every internal or external website built at Amazon. I won’t even try to go into all the various platforms being used (the jQuerys, Reacts, and Angulars of the world), but you can pretty much assume that if it’s popular, someone at Amazon is using it.

Unsurprisingly, there’s also quite a bit of server-side Javascript being written. I think it’s an abomination, but those darned kids keep ridin’ their darned scooters all over my lawn and don’t seem to give a darned hoot about my yammerin’.

There’s also quite a bit of Javascript alternatives like TypeScript and CoffeeScript, though AFAICT they’re not yet at critical mass. That’s not really my field, though, so take it with a big grain of salt.

Python is also gaining popularity, especially with the rise of Data Science and AI. It’s also used pretty widely for general scripting.

(There’s also some R, especially in research, but it’s nowhere near Python AFAIK).

Ruby is found in various places around the company, for scripting as well as building websites and even web services. I do feel like its (Amazon) heyday was several years ago though.
(In particular, there is one important internal tool written in Ruby that requires people who use it to write some of their own Ruby, and many of us are still cursing that decision, which I’m sure seemed very reasonable at the time.)

Perl, yes that Perl, is still used occasionally. In the somewhat-olden days almost the entire Amazon website was built in Perl/Mason

I’m sure there are many more that I’m either not remembering or simply not aware of. Amazon is a huge company, with many thousands of developers that are largely empowered to make most of their own decisions about stuff like this, so I’m sure that it’s a veritable zoo out there.

Footnotes

One of the biggest Internet businesses in the world, Amazon, uses a wide range of technologies to power its extensive and intricate infrastructure. There is no exception when it comes to Amazon backend language that controls data processing, storage, and communication. Amazon's internal systems used a number of programming languages, each one chosen for a particular function to guarantee effective and reliable operations. I may not have the most recent information, but I can give some of the widely utilized languages that Amazon was using at the time.

• Java: Java is the most widely used programming language at Amazon. It is used for a variety of back-end and server-side development tasks, as well as for building the Amazon Web Services (AWS) platform. Java is a popular choice for Amazon because it is a robust and well-established language that is supported by a large community of developers.
• Python: Python is another popular programming language at Amazon. It is widely used for data analysis, machine learning, and other scientific computing tasks. Python is a relatively easy language to learn, which makes it a good choice for engineers who are not familiar with other programming languages.
• C++: C++ is a powerful programming language that is used for performance-sensitive applications. C++ is often used for applications that require low-level access to the system or that need to handle large amounts of data.
• Other languages: Amazon also uses a variety of other programming languages, including JavaScript, Ruby, and Go. These languages are used for a variety of specific tasks, such as developing front-end web applications, building internal tools, and working with big data.

Read More: Angular vs React - Which Reigns Supreme?

It's important to keep in mind that technological decisions can change quickly, and since my previous update, Amazon might have added or adopted newer languages. The individual needs of each project determine the company's approach to technology, which emphasizes scalability, fault tolerance, and high availability in order to provide the best possible client experience.

For its software development, Amazon largely employs a range of web app programming languages. Performance, scalability, maintainability, and compatibility with Amazon's infrastructure are among the criteria used to select these languages. The following are some of the programming languages that Amazon frequently employs:

JavaScript: JavaScript is the most popular programming language for web development, and it is used extensively on Amazon's website. It is used for both client-side and server-side development, and it is also used to power many of Amazon's APIs.

Java: Java is another popular programming language that is used for a wide variety of applications, including web development. It is a compiled language, which means that it is converted into machine code before it is executed. This makes Java a more efficient language than JavaScript, but it also makes it more difficult to learn.

C++: C++ is a powerful programming language that is often used for high-performance applications. C++ is also a very versatile language, and it can be used for a wide variety of tasks, including web development.

Python: Python is a popular programming language that is known for its simplicity and readability. It is a scripting language, which means that it is interpreted rather than compiled. You can hire python developers easily because they are readily available. This makes Python a more flexible language than Java or C++, but it also makes it slower.

Perl: Perl is a powerful programming language that is often used for text processing and system administration. It is a compiled language, and it is known for its flexibility and expressiveness. Perl is not as widely used as some of the other languages on this list, but it is still a valuable tool for certain tasks.

It is important to note that because of Amazon's extensive and diversified technological stack, different teams, projects, and use cases may require different programming languages. In order to be on the cutting edge of innovation, Amazon also encourages its developers to continuously research and master new programming languages and technologies.

Amazon uses a variety of programming languages for its large-scale systems, as the company handles a diverse range of applications, from e-commerce to cloud services, data analytics, and artificial intelligence. Here are some of the main programming languages Amazon utilizes for its large-scale systems:

1. Java

• Usage: Java is one of the most widely used languages within Amazon, particularly for backend systems. It is used for building scalable, high-performance systems, such as the Amazon shopping platform, AWS services, and large-scale enterprise applications.
• Why it's used: Java is known for its portability, reliability, and scalability. It’s a good choice for handling high-traffic applications and enterprise-grade solutions due to its mature ecosystem and large library support.

2. C++

• Usage: Amazon uses C++ in areas where performance is critical, such as the development of low-latency services, and systems requiring high throughput, such as Amazon’s core e-commerce platform and Amazon Web Services (AWS).
• Why it's used: C++ provides fine control over system resources, which is useful for performance optimization, especially in large-scale systems like Amazon’s fulfillment network, real-time bidding systems, and hardware-related tasks.

3. Python

• Usage: Python is widely used for automation, machine learning, data analysis, and AWS-related tools. It’s often used in backend development for scripting, data processing, and AI projects, including Amazon's recommendation engines and analytics platforms.
• Why it's used: Python is a flexible, easy-to-learn language with a rich ecosystem of libraries, making it ideal for rapid development and handling data-intensive tasks, such as web scraping, data analytics, and machine learning.

4. Ruby

• Usage: Ruby, especially with the Ruby on Rails framework, is used for some of Amazon's internal tools and services. It's also used in certain aspects of their development environment, particularly for web applications and internal APIs.
• Why it's used: Ruby is known for its simplicity and rapid development capabilities, making it great for building web applications and prototypes quickly. While it's not the primary language for Amazon’s core systems, it has a place in their toolset for specific purposes.

5. JavaScript (Node.js)

• Usage: JavaScript is used heavily on the frontend for Amazon’s customer-facing websites, but Node.js (which uses JavaScript on the server-side) is also utilized for building scalable backend services, APIs, and microservices.
• Why it's used: JavaScript is essential for building dynamic, responsive user interfaces, while Node.js is great for handling I/O-heavy tasks like web requests, streaming data, and real-time applications. It helps Amazon manage its microservices architecture, particularly for lightweight, high-concurrency services.

6. Go (Golang)

• Usage: Go is used in some of Amazon’s infrastructure, especially in microservices and cloud-native applications. Amazon is known to use Go for services that need to scale efficiently and handle concurrent tasks.
• Why it's used: Go’s simplicity, performance, and support for concurrency make it ideal for cloud services and microservices that need to run efficiently at scale. Amazon's cloud infrastructure, including AWS services, can benefit from Go's performance.

7. PHP

• Usage: While PHP isn’t as prominent at Amazon compared to other languages, it is used in certain internal tools and legacy systems, especially in content management and simple web-based applications.
• Why it's used: PHP is easy to deploy and manage, and it's widely used for web development. Amazon may use PHP for specific tasks related to web content or integration with other technologies.

8. Scala

• Usage: Scala is used for big data processing and distributed systems. Amazon uses Scala in systems like Apache Spark for handling large data workloads, particularly in Amazon’s analytics services and recommendation engines.
• Why it's used: Scala runs on the JVM and is compatible with Java, but it also supports functional programming, making it suitable for distributed systems and big data processing, a crucial part of Amazon’s data infrastructure.

9. SQL and NoSQL

• Usage: Amazon uses a combination of SQL (for relational databases like Amazon RDS and Aurora) and NoSQL (like DynamoDB) to handle large-scale data storage and retrieval needs. While these aren't "programming languages" in the traditional sense, they are essential for managing Amazon’s vast amounts of data.
• Why it's used: SQL is used for structured data that requires complex queries, while NoSQL is used for flexible, scalable data models that support high-velocity, unstructured data, especially for systems like Amazon’s product catalog, customer orders, and reviews.

10. Kotlin

• Usage: Kotlin is increasingly being adopted by Amazon for Android development, particularly for building apps and mobile services that interact with Amazon’s ecosystem.
• Why it's used: Kotlin is a modern alternative to Java that offers more concise code and better interoperability with Java. It’s becoming a popular choice for Android apps, including those within Amazon’s app ecosystem.

11. Rust

• Usage: Rust is sometimes used by Amazon in areas requiring high-performance, memory-safe applications. This language is seen in certain internal tools and systems where both speed and reliability are critical.
• Why it's used: Rust provides memory safety guarantees and is designed to run fast and efficiently, making it an excellent choice for building secure, high-performance systems.

EDIT1: I am answering how did it happen that Amazon the company came to predominantly use Java as their internal language. So I apologize if Java-lovers aren't getting their egos stroked with comments about how great Java itself is. I'm sure it's the best thing since sliced bread. If you're looking for "Oh please tell me once again how great Java is, and let's all rejoice that we can quote Amazon as an endorsement", that's not this answer.

EDIT2: I was deeply involved in the migration and one of the first teams to ever move off Perl onto Java. I was also later involved in the first effort to move off Java to Scala+Node. So I know what I'm talking about. And I was their Tier 1 resolver (basically the single person responsible when the search part of the site goes down on Black Friday) for two years. So I REALLY know what I'm talking about. I've put my money where my mouth is. I'm not speaking hypothetically about how production systems should/could/would run. I ran them.

A few reasons. For one, Java is very new to the company. There are a lot of emotions involved when you make a migration of that scale.

Over the years, Amazon has moved from C to Perl to Java. Three years ago, it would have been 99% Perl code, with some "hipsters" who were trying to bring in this new thing called Java. Today it's Java with some "hipsters" trying to bring in new things like Node.js or Scala.

When they had to move off of Perl, languages like Swift, Rust, Scala, Javascript, and what have you - they weren't exactly options. You never plan a migration based on what you think will happen five years down the line. You have to plan it based on what exists.

If they weren't open to other things, they wouldn't have built Lambda on Javascript. They know what exists. There are teams that would love nothing more than to move to newer and simpler things.

Having said that, I will say that the issues about middle management that were written by another commenter, do exist. But Java wasn't brought in due to those issues. Those issues happened as a result of Java becoming predominant.

Countless hours are wasted debating design patterns, or whether something should be static, or whether something should return an Optional<> or not. Countless refactoring efforts produce promotions. Countless issues can and do stall progress.

But look at Javascript. It brought in classes didn't it? The static-typing people have their way. What was once the simpler, better, cleaner language is morphing into objects and classical inheritance over time. Scala is still bogged by the legacy of existing Java libraries that need to be carried over, and a JVM that performs type erasure, or the necessity for creative methods to implement tail-recursion (since stack frames can't be replaced).

But don't kid yourself. Amazon has a lot of VERY smart people, and many of them have been recognizing the limitations of Java for a while now. Plenty of internal tools use Prolog-like systems. Plenty of developers are toying with Scala. Plenty of internal tools use Python or Ruby. Plenty of new tools like Lambda are Javascript. There is no senior-management-level mandate to use Java or stick with it. It is not a company-wide doctrine. There are plenty of people trying to win the war against Java, and a surprising number of them are succeeding.

Addendum: These battles at Amazon are won through data, not by politics or thought-experiments. So when I saw they're winning these battles, they're doing it be either decreasing costs, increasing reliability, increasing revenue or all the above at once.

Amazon offices, circa 2018.

Me: Hey, Will, does Sarah (my manager) give two shits about what we code in?

Will: Dude, write in assembly for all she cares. I submitted a code review in brainfuck just for shits n’ giggles a few months back. She made me redo it in something intelligible, but that’s beside the point…

Me: Yeah? Well if you can write in brainfuck, I think I’ll be alright with R.

Will: Yeah, you’ll be fine. Nevertheless, screw you and that elitist pile of garbage you call a language – I’m gonna stick with my baby, Python.

Me: Good Christ! You’re almost as bigoted as my great-great-uncle!

There’s great support for AWS in both languages, and no hard and fast rule about which one to use internally.

Amazon cares far more about the conclusions that you make and the speed/accuracy with which you can make them than an essentially arbitrary choice of language.

But I can make prettier plots in R :)

Disclaimer: the photo above is a random one that I found by searching 'old guy'. He has a cool beard, and is likely not actually bigoted.