Scram News
Stocks

Alphabet (GOOGL) seeks $80bn for AI capex surge

Alphabet $80bn raise would fund AI infrastructure and compute as Berkshire Hathaway commits $10bn and shares fall 2 per cent after hours.

By Avery Lin4 min read
Server racks inside a data center representing AI infrastructure spending.

Alphabet (GOOGL) plans to raise as much as $80bn in equity for AI infrastructure and compute, including a $10bn Berkshire Hathaway (BRK.B) commitment, in one of the clearest signs yet that the Google parent’s AI buildout is moving into shareholders’ capital base. Shares fell about 2 per cent after hours on Monday.

Monday’s plan turns a capital-expenditure forecast into a direct financing request. In its investor announcement, Alphabet said it would sell stock through a private placement, public offerings and an at-the-market programme, with proceeds going to infrastructure and compute capacity.

The Berkshire placement gives the package its headline investor. Reuters reported that the group would buy $5bn of Alphabet Class A shares at $351.81 each and $5bn of Class C shares at $348.20, both below Monday’s closing levels. Another $30bn would come through underwritten offerings, with a $40bn ATM programme planned for the third quarter, according to Reuters.

“The company is experiencing strong demand for its AI solutions and services”
Alphabet, according to Reuters

Share issuance is now one measure of that demand. Reuters said Alphabet lifted its 2026 capital-spending forecast in April to $180bn to $190bn, a range that makes data centres, chips and power access balance-sheet questions rather than ordinary growth spending. Alphabet remains highly cash-generative. Even so, an $80bn raise suggests internal cash flow is no longer the simplest route for the next phase of the buildout.

Berkshire’s signal

Berkshire gives Alphabet a patient anchor, not a free pass on dilution. Bill Stone, chief investment officer at Glenview Trust, told Reuters the purchase suggested Greg Abel believed Alphabet could earn an acceptable return on AI capital spending.

The discount remains part of the story. Berkshire gets a carefully priced entry point and Alphabet gets an endorsement from one of the market’s most closely watched buyers. Public investors are being asked to take the larger piece: a $30bn marketed sale and a $40bn ATM programme that would expand the share count to help Google Cloud, search and model infrastructure keep pace with compute demand.

Traders did not treat Berkshire’s entry as a full offset. MarketWatch argued the transaction effectively asks shareholders to help pay an $80bn AI bill. SiliconAngle noted that Alphabet has raised more than $85bn of debt over the past year. The financing stack now includes debt markets, operating cash and equity capital at the same time.

AI spending test

For IPO-bound AI companies, Alphabet’s deal gives investors a live yardstick. If a business backed by Google’s advertising cash flow is willing to sell stock for compute, newer model developers will have to explain how much capital their systems will need after they reach public markets.

That debate is already less theoretical. Wired reported that Anthropic confidentially filed for an IPO that could rank among the largest market debuts, and The Hill also reported the filing with the Securities and Exchange Commission. As a mature platform company, Alphabet is putting an equity price on the same infrastructure race that newer AI groups are trying to monetize.

The point is not that every AI company will rush to sell shares. It is that the financing mix is widening as the compute cycle lengthens. Debt can bridge data-centre spending, cloud revenue can absorb part of it and anchor investors can signal confidence. Alphabet’s plan adds shareholder dilution to the set of tools large technology companies may use when model training and inference capacity outrun a normal annual budget.

For Alphabet, the test is return on invested capital. Management can argue that search, advertising tools, cloud services and Gemini products all require heavier compute, and that underbuilding would be costlier than dilution. Shareholders have a more direct question after Monday’s announcement: whether $80bn of new equity protects those franchises, or only makes the AI spending curve harder to outrun.

AlphabetAnthropicBerkshire HathawaySundar Pichai

Avery Lin

Markets editor covering US equities, single-name stocks and quarterly earnings. Reports from New York.

Related