A renewed debate has emerged over who should control the development of artificial intelligence (AI).
asserted
who → renew → intelligence
The latest dispute follows an essay by Anthropic CEO Dario Amodei, who has argued that the development of increasingly powerful AI systems needs to slow down, calling for greater regulation.
asserted
development → follow → regulation
US President Donald Trump pushed back, writing on Truth Social that the US already has “tremendous CRIMINAL and REGULATORY power” over AI companies, framing safety concerns as a conspiracy that only benefits China.
asserted
that → push → China
He went on to say that whoever wins the AI race wins outright.
asserted
wins → go → race
As states grapple with how to regulate an industry developing at an unprecedented pace, Al Jazeera unpacks the companies driving the AI race and the trillions of dollars behind them.
asserted
Jazeera → grapple → them
Beneath them sits a much larger ecosystem of hardware, infrastructure, and software.
asserted
ecosystem → sit → hardware
These companies can be broadly grouped into three categories:
Hardware
At the foundation of the AI industry are the companies that design and manufacture the physical chips AI runs on, along with the equipment that keeps data centres powered and cooled.
asserted
that → group → centres
These include:
- Chip designers such as Nvidia, Marvell, AMD, Broadcom, and Intel.
- Chip makers like TSMC, which physically manufacture the designs.
- Infrastructure firms such as Arista Networks, Vertiv, and Eaton, which supply the power, cooling, and networking equipment data centres depend on.
asserted
centres → include → Networks
Cloud and compute
These are the companies that turn raw chips into usable computing power, delivering the storage and processing capacity that AI systems run on.
asserted
systems → turn → that
These include:
- Hyperscalers, including Alphabet (Google Cloud), Microsoft (Azure), Amazon (AWS) and Oracle, which operate large-scale cloud infrastructure, storing data and delivering networking globally.
- Neoclouds, such as CoreWeave, Lambda, Crusoe, Nebius, and Nscale, a newer wave of specialised providers built to supply the enormous computing power needed to train AI models.
Software and applications
These are the companies that create the AI products people and businesses actually use.
asserted
people → include → products
These include:
- Frontier AI labs, such as OpenAI, Anthropic, xAI, Google DeepMind and Meta AI develop large language models, with many also building their own compute infrastructure.
- Enterprise AI companies, such as Palantir, ServiceNow, Salesforce, and Snowflake, apply AI to automate business operations.
asserted
companies → develop → operations
The 10 largest public companies operating in the AI space – spanning AI chips, cloud infrastructure, and AI-driven software and services – have a combined market capitalisation of at least $25 trillion.
asserted
companies → operate → trillion
That’s bigger than the gross domestic product (GDP) of every country except the United States, and larger than the entire Chinese economy.
asserted
That → ’ → economy
The graphic below shows the 20 most valuable public companies with direct exposure to AI, ranked by market capitalisation.
asserted
graphic → show → capitalisation
Nvidia is the largest AI-focused public company, with a market capitalisation of roughly $5.1 trillion as of mid-September 2026.
asserted
Nvidia → focus → mid
It designs graphics processing units (GPUs) and AI accelerators – hardware components in a chip that speed up AI workloads – that power leading AI models.
asserted
that → design → models
It makes its money by selling these chips to hyperscalers and AI companies.
asserted
It → make → hyperscalers
Apple, while primarily a hardware company, has invested heavily in on-device AI features and its market cap of $4.86 trillion makes it one of the most valuable companies with any AI exposure.
asserted
it → invest → exposure
Its main AI product, Siri AI, launched in 2026 and runs on Google’s Gemini models under a licensing deal.
asserted
product → launch → deal
It makes the most advanced chips, designed by the likes of Nvidia, Broadcom, and AMD.
asserted
It → make → Nvidia
It makes its money by manufacturing these chips at enormous scale, and currently controls over 70 percent of the global foundry market.
asserted
It → make → market
Additionally, several private AI companies are valued in the billions of dollars, with some expected to go public in the coming months.
asserted
some → value → months
Anthropic, valued at about $965bn (May 2026), develops the Claude family of AI models.
uncertain
Anthropic → value → models
The company filed for an initial public offering (IPO) in June 2026 to sell stocks publicly, and is reportedly targeting a listing as early as October 2026.
uncertain
company → file → October
OpenAI, valued at about $852bn, develops the GPT models behind ChatGPT.
asserted
OpenAI → value → ChatGPT
OpenAI filed for an IPO in June 2026, but is now reportedly leaning toward a 2027 listing.
xAI, now merged into SpaceX, develops the Grok family of AI models.
uncertain
xAI → file → models
At the time of the February 2026 merger, xAI alone was valued at about $250bn.
asserted
xAI → value → 250bn
The Aporia
