Anthropic Spends $1.7bn on Staff in Three Months as IPO Set to Create New...

Anthropic's rapid expansion has turned employment at the artificial intelligence company into an unusually lucrative opportunity, with personnel expenses exceeding $1.7 billion in the third quarter of 2026 as the company prepares for an initial public offering later this month.

The spending covered approximately 5,300 employees between July and September, translating into an average personnel cost of more than $322,000 per worker over the three-month period. The figure, disclosed in an amended registration statement filed ahead of the IPO, illustrates the financial commitment Anthropic has made to attracting and retaining specialized talent as competition for AI engineers, researchers and other technical professionals intensifies.

According to BI, which first reported the figures, the company's personnel expenses include salaries, bonuses, employee benefits, employer payroll taxes and stock-based compensation. The largest component was equity compensation, which accounted for $1.1 billion of the third-quarter total.

For the first nine months of 2026, Anthropic's personnel costs reached $4.8 billion, according to the filing. The scale of the expenditure highlights the premium the company places on its workforce, particularly as it competes with established technology groups and other AI developers for employees capable of building and improving increasingly sophisticated models.

Anthropic's upcoming listing could create a new group of wealthy employees by allowing investors to place a public-market value on the company's equity. Workers who received stock options during its earlier growth stages could see substantial gains if the IPO valuation supports the expectations built into its private-market funding rounds.

The potential windfall also illustrates how compensation in the AI industry extends beyond cash salaries. Equity awards have become a major recruitment and retention tool, allowing companies to offer employees a share of future corporate value while competing for a relatively limited pool of specialized talent.

The $1.1 billion in stock-based compensation recognized during the third quarter represents a substantial portion of Anthropic's reported personnel costs. Unlike cash salaries and bonuses, stock-based compensation is an accounting expense that records the value of equity awards granted to employees over the relevant period.

The figure does not mean Anthropic paid $1.1 billion in cash to its workforce during the quarter. Instead, it captures the accounting cost of employee equity compensation, which can fluctuate with the valuation of the company and the timing of awards.

Anthropic disclosed that it had more than $32 billion in stock-based compensation expenses that had not yet been recognized in its accounts as of the end of September. That amount could feed into personnel expenses over subsequent periods as the awards become eligible for accounting recognition.

A successful IPO could further change the economics of employee compensation. Public trading would provide a market price for Anthropic shares, potentially increasing the value of existing employee awards and making equity compensation more visible to investors assessing the company's operating costs.

The distinction between cash compensation and stock-based expenses will therefore be essential as public-market investors examine Anthropic's profitability. While equity awards do not require the same immediate cash outlay as salaries, they carry an economic cost for existing shareholders through dilution when shares are issued.

Anthropic's personnel spending has also grown alongside its workforce. The company increased its headcount fivefold between the end of 2024 and September 2026, reflecting the scale of recruitment required to support its expanding AI research, product development and commercial operations.

Yet the spending has become more manageable relative to revenue. In 2023, Anthropic spent more than twice as much on personnel as it generated in revenue. During the first nine months of 2026, personnel expenses represented about 15% of total revenue.

That change indicates the company has achieved substantial operating leverage as sales have expanded faster than its workforce-related costs. It is an important measure for investors because it suggests that Anthropic's growth has not required personnel spending to rise at the same pace as revenue.

However, the ratio does not capture every cost associated with developing and delivering AI services. Computing infrastructure, data-center capacity, model training and inference expenses remain important parts of the company's cost structure. A lower personnel-to-revenue ratio alone cannot establish the company's overall profitability or the sustainability of its business model.

Small Workforce, High Revenue Per Employee

Anthropic's relatively lean workforce is another feature likely to attract attention ahead of its listing. Despite employing about 5,300 people at the end of September, the company reported $6.6 million in annual revenue per employee.

That compares with estimates of approximately $3.2 million per employee at Meta and $2.5 million at Google, according to the figures cited in the report. Meta employs about 67,000 people, while Broadcom has more than 30,000 employees. Both companies have market values of roughly $1.6 trillion.

The comparison places Anthropic among a group of technology businesses generating substantial revenue with relatively small workforces. Its own filing argues that this approach gives it an advantage, stating that "Talent density matters more than talent mass."

The argument reflects a broader change in the economics of technology companies. Traditional large technology businesses often require extensive workforces to operate global platforms, manage advertising systems, and support diverse product portfolios. AI companies, by contrast, can generate significant revenue through software products and automated services without expanding employee numbers at the same rate.

However, revenue per employee is not a complete measure of productivity. It does not account for capital investment, computing costs, research expenditure, or the external infrastructure required to operate AI systems.

The company's IPO will place these financial characteristics under closer scrutiny. Investors will need to assess whether its revenue growth can support the cost of developing more capable models, maintaining computing capacity, and retaining highly paid specialists.

The listing is expected to turn years of equity compensation into substantial personal wealth for employees.