Anthropic first took up residence at 500 Howard St. in San Francisco's then somewhat bedraggled SoMa District back in 2023, occupying a 10-story office building formerly used by digital communications firm Slack.
Three years later, some joke that the stretch should be renamed "Anthropic Row" as the artificial intelligence firm has gradually leased floors at three neighboring buildings, with plans to occupy at least two others.
The company has wasted no time channeling the record levels of investment flowing into the AI industry into its ever-expanding downtown San Francisco office empire. This week, the company inked two more leases on Howard Street, marking its third and fourth this year.
The AI giant said it was converting its 240,000-square-foot sublease at 500 Howard St., its current headquarters, into a long-term direct lease from landlord Heitman, a Chicago-based global real estate investment management firm.
It also inked a short-term lease at 405 Howard St. to accommodate another 350 desks ahead of the company's planned 2027 move to its new headquarters at 300 Howard St., where it signed a full building lease in January. Last week, Anthropic added a lease for another three floors at 400 Howard St., just across the street, to the mix.
The firm's growth over the past three years illustrates how quickly the artificial intelligence industry has reshaped real estate and the economy in the nation's tech capital.
"As a company rooted in San Francisco from the very beginning, we are proud to have a long-term presence in the Bay Area," said a company spokesperson in an email to CoStar News.
From startup to behemoth
The AI startup occupied less than 18,000 square feet at a more than century-old building at 394 Pacific Ave. in the city's fashionable Jackson Square neighborhood before it moved into Slack's former space at 500 Howard in late 2023.
The company was widely expected to vacate that space once the Slack lease expired in 2028, especially after Anthropic announced in late January that it was leasing an entire 25-story building nearby at 300 Howard St. as its new future home base.
But the fast-growing maker of the Claude chatbot is now planning to hang on to its current digs at 500 Howard St., also known as Foundry Square IV, in addition to its already significant presence in the SoMa neighborhood.
"The new lease, signed directly with the landlord, will help us plan for the future as we continue to scale our headcount in San Francisco," said the Anthropic spokesperson.
The new lease deals bring the 5-year-old firm a step closer to catching up to local competitor OpenAI's footprint of more than a million square feet in the city alone.
Anthropic also leases about 100,000 square feet at 505 Howard St. The new leases bring the firm's total San Francisco office quarters to more than 900,000 square feet, making it one of the city's largest office tenants.
At those sizes, the relative AI newcomers rival the former footprints of legacy tech players like Google and Facebook — now Meta — that dominated San Francisco's office market during the last tech boom.
Fundraising efforts
The 10-story property at 500 Howard St. was put on the crowded sublease market after enterprise software giant Salesforce bought Slack in mid-2021 for $27.7 billion. At the time, companies were taking advantage of a severe glut of office space following the shift to remote work, which had dragged down rents and valuations across the city.
The picture has again shifted over the past year. Though San Francisco's vacancy rate of nearly 22% is still high, rising demand from AI firms large and small has reignited leasing in the city's commercial center, pushing deals to their highest level since before the COVID-19 pandemic and sending that vacancy rate lower for the first time in years.
The growth comes as AI companies shattered fundraising records in the first three months of the year, bringing in $297 billion, according to data from Crunchbase, which tracks private fundraising. A visible portion of these dollars is flowing into commercial real estate in San Francisco and Silicon Valley, where the AI boom has ignited a flurry of leases after a long period of dormancy.
Anthropic’s ongoing expansion is a signal of the ambitions of the fast-growing company since it was started five years ago by a group of former OpenAI researchers headed by siblings Dario and Daniela Amodei. They argued that leading AI labs were moving too fast toward commercialization without enough emphasis on safety and control.
The company recently became wrapped up in a dispute with the Trump administration after Anthropic refused to loosen restrictions on how its technology could be used for mass surveillance and fully autonomous weapons. Late last month, a judge blocked the administration from labeling Anthropic a "supply chain risk," a move that aimed to block private government contractors from using the company's technology.
The feud has seemingly not hindered its fundraising efforts.
In February, Anthropic raised another $30 billion, raising the company’s valuation to $380 billion amid increasingly heavy speculation that the company will stage an initial public offering in the next year to 18 months.
Anthropic's Claude Code AI technology in particular has helped redefine computer programming, compressing work that took days for humans into hours and moving coders into supervisory roles rather than writing code line by line.