cYpher.camp All articles

AI & business · Field notes

Who actually makes money from AI?

Follow the money through 43 AI businesses: what they sell, who pays, their revenue and profit, and where cYpher.camp fits.

By Keenan Bennett 20 min read

Written
Research
Updated

Two companies can ride the same AI boom and end up on opposite sides of the bill.

NVIDIA reported US$120.07 billion in net income for its fiscal year ending January 25, 2026. CoreWeave reported a US$1.17 billion net loss for the year ending December 31, 2025. One sells the computing machinery. The other buys and operates it, then rents the capacity to customers.

Both belong in an AI headline. They do not belong in the same mental bucket. The question that interests me is who keeps the money after everyone else gets paid.

This matters to me as the founder of cYpher.camp. We sit closer to the customer, at the application and workflow layer. AI Studio brings together tools for images, video, app building, and research. Our hosted cYpher.claw agents support recurring work using the customer's instructions, connected tools, and approval boundaries.

We are not a chip manufacturer or a frontier-model lab. Customers pay through plans and AI credits to use those capabilities. The value we have to earn is turning access to AI into work someone can actually use, while paying for the technology underneath it. Having access to a powerful model is the starting point, not the whole business.

Here's an illustrative example, not a customer receipt: a café owner uses AI Studio to draft a flyer for a weekend event. The owner checks the offer, date, and artwork before using it. The useful result is an approved flyer, not merely a successful generation.

The owner's payment supports the application. The application pays for generation; the provider in turn needs computing capacity. Behind that capacity are chips, memory, factories, power, and cooling. Some providers rent the equipment; others own it. Those purchases happen on different schedules, not as a neat split of each payment.

And whatever remains after one model call is not automatically our profit. Hosting, storage, support, development, and the rest of running a business still have to be paid. That is the thread I want to follow through this map.

Below are 43 companies and branded businesses, grouped by the job they do. For each: what it sells, who pays, the financial evidence, and why its position matters. Some sell AI. Some sell what AI needs. Some use AI to strengthen a business they already have.

A quick translation of the numbers: revenue is sales; net income is the accounting bottom line after expenses and taxes. Operating income excludes items such as interest and tax. Free cash flow usually means operating cash flow minus capital spending. Those are different answers to 'how much money are they making?'.

A revenue run rate annualizes a recent pace; it is not a completed year's sales. Funding, valuation, and backlog are not revenue or profit. Private-company figures below distinguish reported estimates from company statements, and say when profit could not be verified.

Figures are rounded and researched through September 18, 2026. Fiscal periods differ and are dated beside each number. Totals are company-wide unless a segment is named, not AI-only earnings. Currencies stay original: US$ U.S. dollars, € euros, NT$ Taiwan dollars, RMB Chinese yuan, and ₩ Korean won.

Stock labels identify the listed company, not a recommendation to buy it. Parent-company shares are labeled as such; they are not a separate listing for the product. ADS means American depositary share, a US-traded security representing shares in a foreign company. Selected home-market listings are included too. Private companies have no public stock ticker. Listings checked through September 18, 2026.

1. The machines and factories behind the chips

Before a model can answer a question, someone has to make the machinery that makes the chip. This is the furthest part of the map from the customer, but a new app cannot bypass it just by being easier to use.

ASML — (Nasdaq: ASML; Euronext Amsterdam: ASML)

Builds the extraordinarily precise machines that print chip designs onto silicon. Chip factories buy the equipment, upgrades, and servicing. ASML is not competing with ChatGPT; it supplies tools needed far upstream. Reproducing that engineering is much harder than launching another chatbot.

FY2025 (ended 2025-12-31): €32.67 billion revenue; €9.61 billion net income.

  • source https://www.asml.com/en/news/press-releases/2026/q4-2025-financial-results

TSMC — (NYSE: TSM ADS; TWSE: 2330)

Runs the factories that turn other companies’ chip designs into working processors, then packages advanced chips together. Designers pay for manufacturing. It can benefit when different customers win, but factories cost enormous amounts to build and operate. NVIDIA designs chips; TSMC manufactures them. Those are different businesses.

FY2025 (ended 2025-12-31): NT$3.81 trillion revenue; NT$1.70 trillion consolidated net income, under IFRS as issued by the IASB.

  • source https://www.sec.gov/Archives/edgar/data/1046179/000162828026025362/tsm-20251231.htm

2. The processors, memory, and complete systems

The finished server contains several businesses, not one. Compare NVIDIA with Dell below: both sell into AI infrastructure, but their company-wide results show very different amounts left after expenses. That does not isolate AI margins; it does show why a server order and a chip order are not the same economic prize.

NVIDIA — (Nasdaq: NVDA)

Sells AI processors, networking, systems, and the software ecosystem around them. Clouds and companies buy this machinery to train models and run them for users. Its advantage is not just a fast chip: developers and infrastructure teams have built around its tools. That makes switching a project, not a simple shopping decision.

FY2026 (ended 2026-01-25): US$215.94 billion revenue; US$120.07 billion net income. Data Center revenue was US$193.70 billion, a category broader than AI alone.

  • source https://nvidianews.nvidia.com/news/nvidia-announces-financial-results-for-fourth-quarter-and-fiscal-2026

AMD — (Nasdaq: AMD)

Sells competing AI accelerators and server processors, alongside PC and other chips. Buyers want performance, supply, and alternatives to relying on one vendor. AMD has an opportunity when its complete hardware-and-software setup meets the job; a cheaper chip alone does not settle the purchase.

FY2025 (ended 2025-12-27): US$34.64 billion revenue; US$4.34 billion net income.

  • source https://ir.amd.com/news-events/press-releases/detail/1276/amd-reports-fourth-quarter-and-full-year-2025-financial-results

Broadcom — (Nasdaq: AVGO)

Helps big cloud companies build custom AI chips and sells networking that connects computing clusters. It also owns a large infrastructure-software business. Some buyers want processors designed for their own workloads rather than only buying general-purpose accelerators.

FY2025 (ended 2025-11-02): US$63.89 billion revenue; US$23.13 billion net income.

  • source https://investors.broadcom.com/news-releases/news-release-details/broadcom-inc-announces-fourth-quarter-and-fiscal-year-2025

Arm — (Nasdaq: ARM ADS)

Licenses processor designs and collects royalties when customers ship chips using its technology. That puts it inside products sold by many other companies. It also announced its own data-center CPU in 2026, so calling it only a licensing business now misses its expansion into silicon.

FY2026 (ended 2026-03-31): US$4.92 billion revenue; US$904.00 million net income.

  • source https://investors.arm.com/node/8141/html

SK hynix — (KRX: 000660)

Makes memory, including the high-bandwidth memory that feeds AI accelerators. A powerful processor waiting for data is wasted capacity. Buyers pay for memory that keeps up. Manufacturing difficulty can support pricing, but memory remains a cyclical, investment-heavy business.

FY2025 (ended 2025-12-31): ₩97.15 trillion revenue; ₩42.95 trillion consolidated net profit. Audited K-IFRS figures.

  • source https://mis-prod-koce-homepage-cdn-01-blob-ep.azureedge.net/web/attach/121822586381465204.pdf
  • Listing / ownership https://www.skhynix.com/ir/UI-FR-IR01/

Micron — (Nasdaq: MU)

Sells memory and storage chips used in servers and devices. AI needs somewhere to keep both its working data and its stored data. More demand helps, but supply cycles and factory spending still matter; being essential does not make profits steady.

FY2025 (ended 2025-08-28): US$37.38 billion revenue; US$8.54 billion net income.

  • source https://micron.gcs-web.com/news-releases/news-release-details/micron-technology-inc-reports-results-fourth-quarter-and-full-8

Samsung Electronics — (KRX: 005930)

Makes memory, manufactures chips, and sells devices. It has several chances to earn from AI: supplying a component, making someone else's design, or selling the finished phone. The challenge is that success in one does not automatically make the others profitable. These totals cover the broader group.

FY2025 (ended 2025-12-31): ₩333.61 trillion revenue; ₩45.21 trillion net income.

  • source https://images.samsung.com/is/content/samsung/assets/global/ir/docs/2025_con_quarter04_all.pdf
  • Listing / ownership https://www.samsung.com/global/ir/stock-information/listing-Info/

Dell — (NYSE: DELL)

Turns processors, memory, storage, and networking into complete systems businesses can deploy. Customers buy servers, support, and other hardware. It earns its place by integrating and delivering the machinery, not by inventing every chip inside it.

FY2026 (ended 2026-01-30): US$113.54 billion revenue; US$5.94 billion net income. Within that total, AI-optimized-server revenue was US$24.68 billion.

  • source https://www.dell.com/en-my/dt/corporate/newsroom/announcements/detailpage.press-releases~usa~2026~2~dell-technologies-delivers-fourth-quarter-and-full-year-fiscal-2026-results.htm
  • Listing / ownership https://investors.delltechnologies.com/stock-information

Supermicro — (Nasdaq: SMCI)

Builds servers, racks, and cooling systems. Customers need usable installations, not a pile of expensive components. Fast system integration can win orders, but hardware sales carry manufacturing, working-capital, and execution costs.

FY2026 (ended 2026-06-30): US$39.06 billion revenue; US$2.23 billion consolidated net income. These are the audited financial-statement figures; the auditor separately issued an adverse opinion on internal controls.

  • source https://d18rn0p25nwr6d.cloudfront.net/CIK-0001375365/136c312a-7976-4aee-8013-dedb2763f464.pdf

3. The power and cooling nobody can skip

The digital part runs into a physical limit here. More processors are useful only if a facility can power and cool them. These suppliers can benefit without having to predict which assistant people will prefer.

Vertiv — (NYSE: VRT)

Sells power and cooling equipment and services for data centers. It is not an electricity utility. More computing creates heat and power-delivery problems, and operators pay to solve them. A chip cannot earn anyone money while the facility cannot safely run it.

FY2025 (ended 2025-12-31): US$10.23 billion revenue; US$1.33 billion net income.

  • source https://www.prnewswire.com/news-releases/vertiv-reports-strong-fourth-quarter-with-organic-orders-growth-of-252-and-diluted-eps-growth-of-200-adjusted-diluted-eps-37-302684536.html

Schneider Electric — (Euronext Paris: SU)

Supplies electrical distribution, energy management, automation, and data-center systems. Buildings and industrial customers also buy its products. It participates in the physical buildout behind AI, but its entire industrial business is not an AI revenue stream.

FY2025 (ended 2025-12-31): €40.15 billion revenue; €4.16 billion net income, group share.

  • source https://www.se.com/ww/en/assets/564/document/528239/presentation-fy-results-2025.pdf
  • Listing / ownership https://live.euronext.com/en/search_instruments/SU

4. The companies renting out the computing power

Now the equipment becomes a service: customers rent computing instead of building a facility. The operator takes on the bills that renting lets the customer avoid. CoreWeave's loss and Oracle's negative free cash flow make this more than an accounting footnote.

CoreWeave — (Nasdaq: CRWV)

Rents GPU computing capacity to AI labs and other customers. It operates the equipment; it does not manufacture NVIDIA’s chips. Customers pay to use infrastructure instead of building it themselves. The catch is financing: equipment, facilities, depreciation, and interest can consume the revenue. Its loss makes that distinction very real.

FY2025 (ended 2025-12-31): US$5.13 billion revenue; US$1.17 billion net loss.

  • source https://cdn.prod.website-files.com/62ba1fb86485b6d5029975c4/69e821077498160ba1c2c201_2025+CoreWeave+Annual+Report.pdf
  • Listing / ownership https://www.nasdaqtrader.com/dynamic/SymDir/nasdaqlisted.txt

Microsoft — (Nasdaq: MSFT)

Gets paid at several layers: Azure computing, business software, and Copilot products. It already has relationships with the companies buying AI. That distribution matters. An employer can add AI to an existing software relationship instead of introducing an entirely new supplier.

FY2026 (ended 2026-06-30): US$331.84 billion revenue; US$133.75 billion net income. Azure revenue exceeded US$100 billion; Azure-only net profit is not given.

  • source https://www.microsoft.com/en-us/Investor/earnings/FY-2026-Q4/press-release-webcast
  • Listing / ownership https://www.nasdaqtrader.com/dynamic/SymDir/nasdaqlisted.txt

Amazon/AWS — (parent Amazon — Nasdaq: AMZN)

AWS rents computing and sells managed AI services. Amazon also develops its own AI chips and models. AWS can earn money serving different model providers and enterprise customers. Amazon’s overall numbers also include shopping, logistics, advertising, and much more.

FY2025 (ended 2025-12-31): US$716.90 billion revenue; US$77.70 billion net income. AWS alone: US$128.70 billion revenue and US$45.60 billion operating income, not net profit; AWS includes non-AI computing.

  • source https://www.aboutamazon.com/news/company-news/amazon-earnings-q4-2025-report
  • Listing / ownership https://www.nasdaqtrader.com/dynamic/SymDir/nasdaqlisted.txt

Alphabet/Google — (parent Alphabet — Nasdaq: GOOGL Class A / GOOG Class C)

Owns Google, whose AI business spans Gemini, custom TPU chips, cloud services, search, and workplace software. Advertisers, businesses, developers, and subscribers pay different parts of the company. Google can sell AI directly and use it to protect or improve businesses it already has.

FY2025 (ended 2025-12-31): US$402.84 billion revenue; US$132.17 billion net income.

  • source 1 https://s206.q4cdn.com/479360582/files/doc_financials/2025/q4/2025q4-alphabet-earnings-release.pdf
  • source 2 https://cloud.google.com/tpu/docs/intro-to-tpu
  • Listing / ownership https://www.nasdaqtrader.com/dynamic/SymDir/nasdaqlisted.txt

Oracle — (NYSE: ORCL)

Sells databases, business software, and rented cloud infrastructure, including AI computing. Existing enterprise relationships and demand for capacity support the business. But signed future commitments are not this year’s sales, and accounting profit does not mean the data-center buildout funds itself.

FY2026 (ended 2026-05-31): US$67.36 billion revenue; US$17.09 billion consolidated net income. Cloud infrastructure revenue was US$18.10 billion; company free cash flow was negative US$23.70 billion.

  • source 1 https://s23.q4cdn.com/440135859/files/doc_earnings/2026/q4/supplemental-info/Q426_Form8K_Exhibit99-1_Earnings_Release_Tables-Final.pdf
  • source 2 https://www.oracle.com/news/announcement/q4fy26-earnings-release-2026-06-10/
  • Listing / ownership https://www.nasdaqtrader.com/dynamic/SymDir/otherlisted.txt

5. The giants that can make AI part of another business

Apple and Meta have another route. AI can help them sell a device or an advertisement. A free assistant is not necessarily an unmonetized assistant; the payment may arrive somewhere else in the business.

Apple — (Nasdaq: AAPL)

Sells devices and services, with AI integrated into its hardware and operating systems. Some AI runs on the device; some uses cloud services. Apple does not need every customer to buy a separate chatbot subscription for AI to help its business. It can make the device more useful and worth keeping.

FY2025 (ended 2025-09-27): US$416.16 billion revenue; US$112.01 billion net income.

  • source 1 https://www.apple.com/newsroom/pdfs/fy2025-q4/FY25_Q4_Consolidated_Financial_Statements.pdf
  • source 2 https://www.apple.com/newsroom/2025/06/apple-intelligence-gets-even-more-powerful-with-new-capabilities-across-apple-devices/
  • Listing / ownership https://www.nasdaqtrader.com/dynamic/SymDir/nasdaqlisted.txt

Meta — (Nasdaq: META)

Advertisers pay for access to people using Facebook, Instagram, and its wider ecosystem. AI supports recommendations, advertising, and consumer products; Meta also releases Llama models. A free assistant can serve an advertising business. That is different from a lab depending on paid model calls.

FY2025 (ended 2025-12-31): US$200.97 billion revenue; US$60.46 billion net income.

  • source 1 https://www.prnewswire.com/news-releases/meta-reports-fourth-quarter-and-full-year-2025-results-302673127.html
  • source 2 https://ai.meta.com/blog/llama-4-multimodal-intelligence/
  • Listing / ownership https://www.nasdaqtrader.com/dynamic/SymDir/nasdaqlisted.txt

6. The companies building the models

This is the layer most people mean when they say 'AI company.' But rapidly growing model sales leave two bills unanswered: serving today's customers and building tomorrow's model. A run-rate headline tells us the pace of demand, not what the owners keep.

OpenAI — (private; no public stock ticker)

Makes the models behind ChatGPT and sells subscriptions, developer access, and enterprise products, including coding tools. It has both the engine and a recognizable front door. But every paid customer also brings computing costs, and building the next generation requires more investment.

Bloomberg reported on August 13, 2026 that annualized revenue exceeded US$40 billion, citing people familiar with the matter. This is a reported pace, not completed-year sales. Net profit not verified.

Anthropic — (private; no public stock ticker)

Makes Claude and sells subscriptions, developer access, and enterprise tools such as Claude Code. Repeated coding and business work can create recurring demand. The important question is not only how much customers spend, but what remains after serving them and funding research.

CNBC reported that Anthropic told investors its annualized revenue run rate reached US$65 billion at July-end 2026. This is not full-year revenue or profit. Net profit not verified.

  • CNBC reporting https://www.cnbc.com/2026/08/17/anthropic-says-annualized-revenue-climbed-to-65-billion-in-july.html
  • Listing / ownership https://forgeglobal.com/anthropic_ipo/

SpaceX/xAI — (Nasdaq: SPCX; xAI’s parent)

SpaceX acquired xAI in February 2026 and completed its Cursor acquisition in August. Customers now pay the group for launches, Starlink connectivity, Grok, X, rented AI computing, and coding software. Owning more layers creates more places to charge, but it also combines very different costs. Scale alone does not tell us which activity earns its keep.

Six months ended 2026-06-30, not a full year: US$12.51 billion revenue and US$4.82 billion consolidated net loss, from its unaudited GAAP statements. The AI segment generated US$3.38 billion revenue, including advertising, subscriptions, and infrastructure; that is not Grok-only sales. These results predate the completed Cursor acquisition.

  • source 1 https://s21.q4cdn.com/184289198/files/doc_financials/2026/q2/SpaceX-Reports-Second-Quarter-2026-Results.pdf
  • source 2 https://x.ai/news/xai-joins-spacex
  • source 3 https://cursor.com/blog/joining-spacex

DeepSeek — (private; no public stock ticker)

Offers model access and downloadable model weights. Developers can pay for hosted use or operate eligible models themselves. This creates another option in the market and can pressure pricing. A low model-training cost or an estimated serving margin is not the company’s bottom-line profit.

Company revenue and net profit not verified. Unknown does not mean zero.

Alibaba/Qwen — (parent Alibaba — NYSE: BABA ADS; HKEX: 9988)

Alibaba combines commerce and cloud computing with the Qwen model family. Qwen can attract developers and businesses that then pay Alibaba for hosting and related services. That makes a model useful to a much larger business even when downloading its weights is free.

Fiscal year (ended 2026-03-31): RMB1023.67 billion revenue; RMB102.13 billion net income. Cloud segment revenue was RMB158.13 billion, not Qwen-only sales.

  • source https://data.alibabagroup.com/ecms-files/1532295521/5b1cb883-8d00-4237-a148-6631cc12a5d2/Alibaba%20Group%20Announces%20March%20Quarter%202026%20and%20Fiscal%20Year%202026%20Results.pdf

Mistral AI — (private; no public stock ticker)

Sells model access, Le Chat subscriptions, and enterprise deployments, with downloadable weights for some models. European positioning and deployment choice offer a different pitch from a single closed hosted service. Freely available models can coexist with paid hosting and business support.

CNBC reported annualized revenue above US$400 million in January 2026. In a September interview, its CEO declined to give a revised achieved figure; US$1 billion remained an expectation. Net profit not verified.

Moonshot AI/Kimi — (private; no public stock ticker)

Builds Kimi and sells assistant subscriptions, developer access, and enterprise services. Model distribution can bring developers in; convenient hosted service can turn some of them into paying customers. A revenue target for later in the year is not money already earned.

TechCrunch, citing Bloomberg, implies an August 2026 annualized revenue run rate of US$1 billion: its reported year-end US$2 billion target was double the August pace. The target is not achieved revenue; net profit not verified.

Zhipu AI/GLM — (HKEX: 2513 — Knowledge Atlas)

Builds GLM models and sells cloud services and customer-controlled deployments. Also known as Knowledge Atlas Technology, it listed in Hong Kong in January 2026. Its reported loss is a reminder that growing model sales and building a profitable model company are different milestones.

Fiscal year (ended 2025-12-31): RMB724.33 million revenue; RMB4.72 billion net loss.

  • source https://www.hkexnews.hk/listedco/listconews/sehk/2026/0331/2026033101549.pdf

7. Chinese giants spanning several layers

A quick detour from layers to geography: these companies combine several of the businesses above. Keeping them together makes the map easier to read, but 'Chinese AI' is not one business model any more than 'American AI' is.

Huawei — (private; no public stock ticker)

Sells telecom equipment, devices, computing chips, cloud services, and other infrastructure. A business can buy several parts of its technology setup from Huawei rather than assemble separate suppliers. That breadth is the commercial advantage to understand here, not a claim that every part earns money from AI.

FY2025 (ended 2025-12-31): RMB880.94 billion revenue; RMB68.04 billion net income.

  • source https://www.huawei.com/en/annual-report/2025
  • Listing / ownership https://www.huawei.com/en/facts/question-answer/who-owns-huawei

Tencent — (HKEX: 0700)

Runs consumer platforms, gaming, advertising, payments, and enterprise services. AI can reach customers through businesses Tencent already operates. Like Meta, it has ways to monetize better software and customer experiences without making every user pay for a standalone assistant.

FY2025 (ended 2025-12-31): RMB751.77 billion revenue; RMB229.80 billion net income.

  • source https://static.www.tencent.com/uploads/2026/03/18/e6a646796d0d869acc76271c9ee1a6a5.pdf
  • Listing / ownership https://www.tencent.com/wp-content/uploads/2026/08/E700_IR.pdf

Baidu — (Nasdaq: BIDU ADS; HKEX: 9888)

Sells advertising and enterprise cloud services while developing models and autonomous driving. Its search and cloud businesses give it customers to introduce AI to. The unresolved commercial question is how much newer AI work adds, rather than simply replaces, an existing source of sales.

FY2025 (ended 2025-12-31): RMB129.08 billion revenue; RMB5.46 billion net income.

  • source https://ir.baidu.com/news-releases/news-release-details/baidu-announces-fourth-quarter-and-fiscal-year-2025-results
  • Listing / ownership https://www.nasdaqtrader.com/dynamic/SymDir/nasdaqlisted.txt

8. The apps, data platforms, and business tools

Here the customer pays for something closer to a finished job: code, research, a creative asset, or work inside a business. This is also where cYpher.camp fits. The test I care about is whether the product remains useful when its underlying model is also available to a competitor.

Cursor/Anysphere — (parent SpaceX — Nasdaq: SPCX)

Cursor, built by Anysphere and now part of SpaceX, sells developers an editor and agent workflow that understands their codebase and uses development tools. It integrates outside models and develops its own. Customers pay for the usable coding experience, not merely model access. Its August 14 acquisition announcement illustrates companies moving across layers.

Reuters reported roughly US$2.6 billion in annualized business-to-business revenue in June 2026, citing company data. That is a B2B subset and a pre-acquisition snapshot, not total-company annual sales. Standalone net profit not verified.

Venice — (private; no public stock ticker)

A smaller specialist, not a hyperscaler peer. It sells consumer subscriptions and developer access to multiple AI models, with privacy as a selling point. That is a real product distinction, not proof of zero data processing. Its policy includes third-party processing and temporary generated-video storage. Token value is not operating revenue.

July 1, 2026: its CEO told TechCrunch annualized revenue exceeded US$70 million and the company was profitable. No net-profit amount or audited accounting basis was verified.

Perplexity — (private; no public stock ticker)

Sells AI research, answers, and agent tools to consumers and businesses. Its opportunity is to become the place people start a question or a task. It still has to pay for computation, produce reliable answers, and navigate content rights.

Sacra estimates an August 2026 annualized revenue run rate of US$750 million. This is a third-party estimate, not audited annual sales. Net profit not verified.

Hugging Face — (private; no public stock ticker)

A hub where developers share and discover models, datasets, and tools. It earns through paid collaboration features, hosting, and computing services. It is not one giant model. A free download can build the ecosystem without generating a sale that day.

Company revenue and net profit not verified. Its paid products establish how it charges, not how much the company earns.

Databricks — (private; no public stock ticker)

Helps businesses organize, govern, and use their data for analytics and AI. Customers pay for platform usage. A model becomes more useful when it can work with the right business information, and that makes the data layer valuable in its own right.

Company disclosure, August 13, 2026: revenue run rate above US$7 billion. This includes the data business, not only AI. Net profit not verified; adjusted free cash flow is a different measure.

  • source https://www.databricks.com/company/newsroom/press-releases/databricks-grows-80-yoy-surpasses-7b-revenue-run-rate-scales
  • Listing / ownership https://forgeglobal.com/databricks_ipo/

Snowflake — (NYSE: SNOW)

Sells a cloud data platform with analytics and AI capabilities. Customers pay as they use it. It competes for the place where business data is stored and put to work. Usage growth can bring both additional revenue and additional infrastructure costs.

FY2026 (ended 2026-01-31): US$4.68 billion revenue; US$1.33 billion net loss.

  • source https://www.snowflake.com/en/news/press-releases/snowflake-reports-financial-results-for-the-fourth-quarter-and-full-year-of-fiscal-2026/
  • Listing / ownership https://www.nasdaqtrader.com/dynamic/SymDir/otherlisted.txt

Salesforce — (NYSE: CRM)

Sells customer-management software and AI capabilities such as Agentforce into sales and service workflows. It already has business relationships and the context of the work. Its opportunity is to make the system a customer already uses more useful, rather than persuade that customer to start over elsewhere.

FY2026 (ended 2026-01-31): US$41.53 billion revenue; US$7.46 billion net income.

  • source https://s205.q4cdn.com/626266368/files/doc_financials/2026/q4/CRM-Q4-FY26-Earnings-Press-Release.pdf
  • Listing / ownership https://www.nasdaqtrader.com/dynamic/SymDir/otherlisted.txt

ServiceNow — (NYSE: NOW)

Sells enterprise workflow software, including AI features. Businesses already use it to route requests and manage work. That gives it a place to add AI where the next action happens, rather than leave an employee copying a chatbot's answer into another system. Adoption still depends on fitting the customer's processes.

FY2025 (ended 2025-12-31): US$13.28 billion revenue; US$1.75 billion net income.

  • source https://s205.q4cdn.com/916135447/files/doc_downloads/annual-meeting-of-share-holders/2026/NOW-2025-Annual-Report-bookmarked.pdf
  • Listing / ownership https://www.nasdaqtrader.com/dynamic/SymDir/otherlisted.txt

Adobe — (Nasdaq: ADBE)

Sells creative, document, and marketing software, with generative AI inside established tools. Creators pay for the working environment as well as the output. Being where the project already lives is an advantage; new tools still have a chance if they make the finished result easier to produce.

FY2025 (ended 2025-11-28): US$23.77 billion revenue; US$7.13 billion net income.

  • source https://www.adobe.com/cc-shared/assets/investor-relations/pdfs/adbe-10k-fy25-final.pdf
  • Listing / ownership https://www.nasdaqtrader.com/dynamic/SymDir/nasdaqlisted.txt

Palantir — (Nasdaq: PLTR)

Connects organizational data to decisions and operational workflows for businesses and governments. Customers buy software and deployment capability. Its value proposition is making data useful in real operations, not simply providing a chat window.

FY2025 (ended 2025-12-31): US$4.48 billion revenue; US$1.63 billion net income.

  • source https://investors.palantir.com/files/2025%20FY%20PLTR%2010-K.pdf
  • Listing / ownership https://www.nasdaqtrader.com/dynamic/SymDir/nasdaqlisted.txt

IBM — (NYSE: IBM)

Sells enterprise software, consulting, and infrastructure, including AI tools. Companies pay for technology and help making it work inside existing organizations. Its AI book-of-business headlines include measures other than recognized revenue, so they should not replace the sales figure.

FY2025 (ended 2025-12-31): US$67.54 billion revenue; US$10.59 billion net income.

  • source https://newsroom.ibm.com/2026-01-28-IBM-RELEASES-FOURTH-QUARTER-RESULTS
  • Listing / ownership https://www.nasdaqtrader.com/dynamic/SymDir/otherlisted.txt

9. AI in the physical world

In the physical world, an impressive demonstration and a business with paying customers are still separate milestones. The existing business can fund the ambition without proving the ambition already pays for itself.

Tesla — (Nasdaq: TSLA)

Sells vehicles, energy products, and services while investing in autonomous driving and robotics. Its appeal in this map is the possibility of selling work performed in the physical world, not only software. The financial totals below describe the existing company; they do not establish a profitable robot business.

FY2025 (ended 2025-12-31): US$94.83 billion revenue; US$3.79 billion net income attributable to common shareholders.

  • source https://ir.tesla.com/_flysystem/s3/sec/000162828026003837/tsla-20260128-gen.pdf
  • Listing / ownership https://www.nasdaqtrader.com/dynamic/SymDir/nasdaqlisted.txt

Who gets to keep the money?

NVIDIA and CoreWeave brought us into this map from opposite sides of the equipment bill. The rest of the companies show how many other places there are to stand: factories, memory, power, rented computing, models, distribution, and useful applications.

What I take from this is not that one layer always wins. It is that each company needs a reason the next customer cannot easily replace it: engineering that is hard to reproduce, capacity that is hard to obtain, a tool people rely on, or a better way to finish work.

The café owner in our example does not need to care which chip helped make the flyer. But every business between the owner and that chip has to care what it costs to deliver its part. Adding all their revenues together would count portions of the same spending repeatedly, not measure one AI market.

For cYpher.camp, and anyone building something smaller than a model lab, the test is simple: would the customer still pay us if they could use the same model somewhere else? The answer has to be in the finished work, the checks around it, and the costs of delivering it.

The AI boom is not one business. It is a long line of businesses sending each other bills. The interesting part is who gives the next person enough value to keep getting paid.

Source note: This is a dated educational guide, not an investment recommendation or a ranking. Financial lines use reported GAAP/IFRS results where established, not adjusted earnings. Company disclosures and third-party estimates are labeled separately. Public earnings releases may be unaudited; historical figures can be restated. Figures do not isolate AI profitability unless explicitly stated.