cloud
in English: cloud computing
The cloud is computing capacity—storage, processing power, software—that you do not keep on your own machine but use via the internet, typically for a rental fee or subscription.
In plain terms
The cloud is not actually in the sky but in huge buildings filled with server rooms, known as data centers. A large data center can have tens of thousands of computers working, which you access via the internet. When your phone photos 'go up to the cloud', they end up in the provider's data centers, usually in multiple copies, so they are not lost if one machine fails.
It is like electricity service. You do not build a power plant in your garden; you plug the device into the socket and pay for what you consume. Similarly, with the cloud, you do not buy a server; you rent as much storage and computing power as you need at the moment.
Cloud computing is a prerequisite for artificial intelligence. Large language models and AI agents require powerful machines that cannot be operated at home or in an average office. When you chat with a chatbot, the response is almost always generated in a distant data center.
Compared to your own machine, the difference is that in the cloud, you do not handle maintenance, updates, or expansion. In exchange, you entrust your data to the provider, and without internet, you can only access what has a copy on your own device.
How it works
- The provider builds data centers filled with servers, storage, and networking equipment, and ensures power, cooling, and security.
- Software divides physical machines into many smaller "virtual machines," so many customers can use the same hardware at once without seeing each other's data.
- You request storage, computing power, or ready-made services over the internet – via a browser, application, or program – and receive them within minutes.
- If configured this way, the system automatically allocates more resources under load and reduces them during quieter periods; this is called elastic scaling.
- The provider measures consumption and bills based on usage or a monthly subscription.
An everyday example
You take a photo of your child with your phone, and by evening the image is already on your laptop, even though you did not copy it over. The phone uploaded the photo to the cloud, and the laptop downloaded it from there. The same happens when you ask an AI assistant: your query goes to a remote data center, and the response comes back from there.
Who is affected
A common misconception
Limits and risks
- Without internet or in case of provider outage, you cannot access your data and programs.
- Usage-based fees can increase unnoticed, especially with AI tasks where computing power is expensive.
- It can be difficult and costly to switch providers if all your systems are built on a single company's solutions (this is called vendor lock-in).
- With sensitive data, check which country stores them and under whose jurisdiction the provider falls, because foreign authorities can access data stored in Europe if the provider is foreign. It is also important whether this complies with data protection rules, such as the GDPR.
Similar, but not the same
- data center – The data center is the physical building with the machines, while the cloud is the service through which you use these machines remotely on a rental basis.
- helyben futtatott AI – In such cases, the model runs on your own computer or your company's server, so data does not leave the premises, but performance is limited by your own hardware.
- edge computing – Edge computing processes data close to its point of origin – for example, in a factory or on a device – rather than in a remote data center.
Related concepts
IT infrastructure, computing capacity
public cloud, private cloud, hybrid cloud, sovereign cloud, software as a service
data center, GPU, edge computing, API, helyben futtatott AI
A brief history
Cloud services in the modern sense began in 2006, when Amazon Web Services started offering rentable storage (S3) and computing capacity (EC2). The US standards institute, NIST, issued its most frequently cited definition in 2011.
Where things stand
By the end of 2026, the cloud will be the standard background for digital services and artificial intelligence, while major providers will primarily expand data centers needed for AI. In Europe, meanwhile, the question of keeping data under its own jurisdiction is becoming increasingly prominent.
In the news
- Google Cloud adds OpenAI models and Codex agents to Model GardenNov 6, 2026 · Code
- Microsoft scales back neocloud deals: winds down Oracle rental, turns down NscaleNov 4, 2026 · Agent Age
- Google to invest $15 billion in expanding data-center capacity in MalaysiaNov 4, 2026 · Ipar
- Nvidia orders cloud clients to sell at least 51% of their services to non-Chinese firmsNov 3, 2026 · Agent Age
- 'AI washing' becomes a credibility problem as leaders cite AI for layoffsNov 2, 2026 · Work & HR
- India: Karnataka to train 5,000 teachers in AI, Telangana offers 1 lakh students free AI training, Delhi launches AI programmeNov 2, 2026 · Education & Family
- Google.org gives Helium Health $2.5m for AI against maternal deaths in Nigeria; Clinics On Cloud raises $12.5mNov 2, 2026 · Health
- AI data center construction shifts to a manufacturing mode: faster build-out, more flexibilityOct 28, 2026 · Ipar
Also known as
cloud · cloud service · cloud computing · compute cloud · cloud computing · cloud · cloud computing · the cloud
Sources
This entry was machine-edited and has undergone machine expert review; the “In the news” section updates automatically from Agent Age stories.