Meta Seeks $97.5bn and Anthropic Files for IPO as AI’s Bill Moves to Investors

Meta launched share sales worth up to $97.5bn on 26 October to pay for its AI build-out. The package has three parts. There’s $25bn of common stock priced at $625 a share, $12.5bn of convertible preferred shares, and a $60bn programme of gradual sales from early 2027. Its prospectus supplement says Nvidia, AMD and Broadcom have offered to buy up to $5bn of the stock.

Meta’s free cash flow was positive at $10.63bn in the third quarter of 2025. In the same quarter this year it was negative by $2.12bn, the first shortfall in more than a decade. Meanwhile, Meta has raised the bottom of its 2026 capital spending forecast to $135bn, and its future spending commitments now total about $825bn.

Existing shareholders in the owner of Facebook, Instagram and WhatsApp will see each new share dilute their stake. Meta’s AI build-out now draws on outside investors’ money as well as its own profits.

Chipmakers buying into their customer

Nvidia said it would put $3bn into Meta’s shares. As a result, a chip supplier is helping to fund one of its customers. The IMF’s October stability report warned that “weaker or less durable investment and profitability in artificial intelligence could lead to a correction”. It added that corrections “could be amplified by rising leverage and concentrated exposures.”

Alphabet lifted its own equity offerings to $84.75bn in June. On 26 October, SpaceX said an Apollo affiliate would buy AI chips and servers to lease to it. The deal is worth about $40bn in principal. Spending on this scale already shows up in the world growth figures.

Investors in Meta’s 2027 programme will also get less say. Those sales start once the company’s new non-voting Class C shares begin trading, which Meta expects in the first quarter of 2027.

A listing priced on expectations

Anthropic, the maker of the Claude AI models, made its IPO filing public on 26 October. It plans to list on Nasdaq under the ticker ANT. Revenue reached $16.3bn in the first half of 2026, against $1.16bn a year earlier.

Its net loss for those six months was $80.7bn. However, $72.4bn of that came from revaluing convertible notes and similar instruments as the company’s own valuation rose. On an adjusted basis, Anthropic made an operating profit of $1.6bn. It held $131.7bn in cash and short-term investments at the end of June.

A funding round in May valued Anthropic at about $965bn, and Reuters has reported that the listing could value it above $2tn. For comparison, SpaceX raised $75bn in its June IPO, more than double Saudi Aramco’s previous record of $29.4bn.

Buyers of Anthropic’s listed shares won’t control the company either. A Long-Term Benefit Trust elects a majority of the board, and a single founders’ share carries 50.1% of the votes on most matters.

Payroll as a source of savings

On the same day, Microsoft offered a voluntary exit package to staff at level 70 and below in the US, UK, Canada and India. About 60,000 people are eligible, a figure Microsoft confirmed to TechCrunch.

That’s roughly 27% of the 223,000 people Microsoft employed full time in June, according to its annual report. A year earlier it had 228,000. Staff in data centres, sales and marketing are excluded. Earlier this year, more than 30% of the roughly 8,750 workers offered a retirement package took it.

Microsoft’s own post doesn’t mention AI, and Bloomberg describes the offer as its latest cost cut. Across the US, employers cited AI for more than 120,000 job cuts in the first nine months of 2026.

What would test the shift

Taken together, these deals move more of the cost and risk of AI onto shareholders and staff. Anthropic’s next filing will set a price range, and Meta’s $60bn programme will run into 2027. Both will show how much more capital investors will supply before AI revenue covers the spending. Meta’s next quarterly results will show whether its free cash flow turns positive before the $60bn programme starts.