Home / Research / AI Company Profitability Statistics

Market research · October 23, 2026

AI Company Profitability Statistics: 2026 Report

By Darin ten Bruggencate, Founder, AI Industry Reviews

Companies that build artificial intelligence products have raised and spent enormous sums over the past three years, on chips, data centers, researchers and the electricity to run them. Revenue at many of these businesses has grown quickly, and some of the largest private labs now take in billions of dollars a quarter. Spending has grown just as fast, and in several cases faster, so a company can double its sales and still lose more money than it did the year before.

Profit has been much harder to find, so we looked at which AI companies make money and how much.

Our analysts collected income statement figures for 15 publicly traded AI-focused companies and eight large technology and chip companies from their SEC 10-K and 10-Q filings. We measured revenue, operating income and net income for each company's latest fiscal year and latest reported quarter. We then calculated operating and net profit margins. For OpenAI, Anthropic and xAI, which do not file their own public financial statements, we gathered the revenue and loss figures reported from audited statements, IPO prospectuses and SpaceX's SEC filings. Only 3 of the 15 AI-focused companies, Palantir, UiPath and Innodata, posted a net profit in their latest fiscal year.

In the following sections, we compare profit margins across publicly traded AI companies, track how those margins changed from 2021 to 2026, set them against Big Tech and AI chipmakers, and lay out the reported losses at the three largest private AI labs.

Profitability of Publicly Traded AI Companies

This section shows revenue, operating income and net income for 15 publicly traded companies whose main business is AI software, AI infrastructure or AI-driven products, using each company's most recent full fiscal year. Operating income is what a company earns from its business before interest, taxes and investment gains. Net income is the final profit or loss after all of those items.

AI-focused publicly traded companies, latest fiscal year (US$ millions)

Source: AI Industry Reviews analysis of SEC 10-K filings. n/m = not meaningful; margins are shown only for companies with revenue of $50 million or more. Margins are calculated from unrounded figures in company filings.

• The 15 companies brought in a combined $16.57 billion in revenue and a combined net loss of $2.14 billion. Palantir's $1.63 billion net profit offsets much of that total. Without Palantir, the other 14 companies lost a combined $3.77 billion.
• Among the 13 companies with at least $50 million in revenue, the median net margin was -19.3% and the median operating margin was -13.8%.
• CoreWeave came close to breaking even on operations, with a $46 million operating loss on $5.13 billion of revenue, but its $1.15 billion in interest expense on debt pushed its net loss to $1.17 billion. Interest costs can outweigh operating results at AI companies that borrow heavily to build data centers.

Profitability in the Latest Reported Quarter

This section uses each company's most recent quarterly filing, which for most companies covers April through June 2026. Quarterly results show whether the annual picture has shifted since the last 10-K. It adds Cerebras Systems, which went public in 2026 and has not yet filed a 10-K.

AI-focused publicly traded companies, latest reported quarter (US$ millions)

Source: AI Industry Reviews analysis of SEC 10-Q filings. Operating margins are shown only for companies with quarterly revenue of $10 million or more. Innodata's quarterly filing data did not include an operating income figure.

• Four of the 15 companies with an operating income figure earned an operating profit in their latest quarter: Palantir, UiPath, Symbotic and Cerence. Six of the 16 companies reported a net profit.
• Palantir's operating margin rose to 47.1% in the quarter ended June 30, 2026, from 31.6% for full-year 2025.
• Tempus AI reported a $5.6 million net profit in the quarter while its operations still lost $75.9 million, so its net figure depended on items outside its core business.

How AI Company Margins Changed from 2021 to 2026

This section follows the operating margin of five publicly traded AI companies across their last five or six fiscal years, using the same 10-K data. We chose companies with a long enough filing history to show a trend and margins that fit on one scale.

Operating margin by fiscal year, selected AI companies

Source: AI Industry Reviews analysis of SEC 10-K filings. Fiscal years follow each company's own labels. UiPath's fiscal year ends January 31 and C3.ai's ends April 30. n/a means the year was not reported in the structured filing data or has not ended.

• Palantir moved from a -26.7% operating margin in 2021 to 31.6% in 2025, and UiPath reached an operating profit in fiscal 2026 for the first time in the years shown.
• C3.ai went the other way. Its revenue fell from $389.1 million in fiscal 2025 to $250.3 million in fiscal 2026 while its operating loss grew to $498.5 million.
• Tempus AI's margin improved in 2023, fell back to -99.7% in 2024 and improved again to -19.9% in 2025. Palantir and Innodata, the two companies with an operating profit in each year from 2023 to 2025, improved their margins in each of those years.

Big Tech and AI Chipmaker Profitability

This section compares the AI-focused companies with eight large technology and semiconductor companies that sell the chips, cloud capacity and software platforms AI products run on, alongside many other products.

Large technology and chip companies, latest fiscal year (US$ millions)

Source: AI Industry Reviews analysis of SEC 10-K filings.

• The eight companies earned a combined $568.7 billion in net income on $2.03 trillion of revenue. Their median operating margin was 36.0%, against -13.8% for the median AI-focused company.
• Nvidia's operating margin was 60.4% in fiscal 2026, the highest in the group. Its $120.1 billion in net income was more than seven times the combined revenue of the 15 AI-focused companies.
• Net income at these companies can also move with the value of their investments. Alphabet's $112.2 billion net income in the second quarter of 2026 included $99.0 billion of net gains on equity securities, according to its earnings release, so most of that quarter's profit came from investment gains.

OpenAI, Anthropic and xAI Losses

This section brings together the revenue and loss figures reported for the three largest private AI labs. OpenAI and Anthropic do not publish financial statements, so these figures come from audited statements and draft prospectuses reviewed by news outlets. xAI's figures come from SpaceX, which bought xAI and reports it in a segment of its SEC filings. That segment also includes the X social network and, since August 14, 2026, the coding software company Cursor.

Reported revenue and losses at private AI labs (US$ billions)

Source: AI Industry Reviews compilation of the cited reports and filings. OpenAI quarterly operating losses include stock-based compensation. SpaceX AI segment figures for 2025 quarters are the prior-year comparisons in its 2026 10-Q filings.

• OpenAI's operating loss equaled 160% of its revenue in 2025, down from 237% in 2024, as reported by Ars Technica. Its losses still grew in dollars, and the 2026 quarters reported so far show the operating loss widening from $9.3 billion to $12.3 billion.
• Most of the gap between operating and net losses at OpenAI and Anthropic comes from non-cash accounting charges tied to investor financing. Anthropic's 2025 net loss included a charge of roughly $34 billion, and its first-half 2026 net loss included a $72 billion charge, according to the reports.
• Anthropic's revenue rose to $11.6 billion in the second quarter of 2026 with a small operating profit, The Wall Street Journal reported, and Anthropic told investors that third-quarter revenue reached $16.8 billion and was profitable on an adjusted basis, according to Reuters as cited by Quartz. Neither profit figure was given in dollars, and Anthropic has not publicly defined the adjusted measure.

• SpaceX's AI segment moved from a $1.26 billion operating loss in the second quarter of 2026 to a $431 million operating profit in the third quarter, on revenue of $7.56 billion. Over the first nine months of 2026 the segment still lost $3.30 billion on operations.
• The third-quarter jump came after Cursor joined the segment in August and after Anthropic agreed in May to pay xAI $1.25 billion a month for computing capacity, as PitchBook reported. The filing does not break out how much each source contributed.

What Outside Research Says About AI Profitability

This section adds two outside reference points that help explain why losses at AI companies remain large even as revenue grows.

• Bain & Company's 2025 Global Technology Report estimated that about $2 trillion in annual revenue will be needed by 2030 to fund the computing power that AI demand requires, and that the industry could fall about $800 billion short of that even after expected savings from AI.
• OpenAI told investors it hopes to be profitable by 2030, Ars Technica reported. In November 2025, The Information reported that Anthropic's best-case outlook included $17 billion in cash flow in 2028. Both are company projections.

What These Numbers Mean for the AI Industry

The money in AI has so far gone mostly to the companies that supply it. Nvidia, Microsoft, Alphabet and the other large platforms earned hundreds of billions of dollars while most of the companies building AI products reported losses, and OpenAI, Anthropic and xAI together lost more than $35 billion on operations in 2025.

That split is starting to shift at the edges, with operating profits at Palantir and UiPath and profitable quarters in 2026 at Anthropic and SpaceX's AI segment. OpenAI and Anthropic have both filed draft IPO paperwork with the SEC, and a public listing would put their audited quarterly results in front of investors.

The larger question is whether revenue can catch up with the spending commitments these companies have made for data centers and chips. Anthropic disclosed $518 billion in future cloud and computing obligations in its September draft prospectus, and those costs will arrive whether or not demand keeps pace. So if you are comparing AI companies on profitability, look at operating income in the most recent quarters, because net income at both the labs and their largest investors now swings on accounting charges and gains.