5 Things to Know About AI This Week
A quick scan of what happened in AI this week.
1. Anthropic reports $11.5B quarter, its first with an operating profit
Anthropic told investors its preliminary Q2 2026 revenue topped $11.5 billion, up more than 14-fold from $787 million a year earlier and up from $4.73 billion in Q1. The company also reported positive adjusted operating income for the first time. Reports point to a potential IPO as soon as this fall, which would make Anthropic one of the first major private AI companies to go public.
Real estate angle: Brokerages and title companies are careful about building workflows on AI company models that go through billions of dollars a year, because a company that folds or gets acquired takes your workflow with it. A frontier lab covering its own costs is a point in favor of committing to a platform rather than renting one tool at a time. It is not a guarantee, but it is the kind of stability that lets you breathe a little easier.
Source: CNBC
2. Anthropic, Macquarie, and GIC form Theseus Infrastructure to build AI data centers
On August 10, Anthropic announced a partnership with Macquarie Asset Management and Singapore's sovereign wealth fund GIC to create Theseus Infrastructure, a platform that will develop, operate, and lease purpose-built data centers to Anthropic under long-term agreements. The initial focus is in the United States. Macquarie funds and GIC will own the platform and supply most of the equity. Anthropic also committed to covering electricity price increases that local consumers might otherwise face from these sites.
Real estate angle: Data centers are real estate transactions before they are anything else: land acquisition, zoning, utility servitudes, long-term ground leases, and title work at enormous scale. Institutional capital structuring these deals as lease-back transactions means more of this work is coming to markets that have cheap land and power. If your market is in the running for one of these sites, the commercial side of your practice should be paying attention.
Source: Macquarie Group
3. Meta open-sources Muse Glimmer, a 30B model that runs on a laptop
Meta released Muse Glimmer on August 10 under the Apache 2.0 license. It's a 30-billion-parameter multimodal model built for agentic tasks that runs locally on a single 24GB GPU through tools like Ollama, with no cloud connection required. Meta claims it outperforms Google's Gemma 4 31B in its size class.
Real estate angle: Local models matter for one reason: the data never leaves your machine. Real estate professionals handle nonpublic personal information, and a common objection to AI tools is that client data flows to a third-party server. A capable model running entirely on office hardware removes that objection for tasks like summarizing contracts or drafting correspondence. The tradeoff is that you own the security of that machine, so treat a local AI workstation like you treat your server.
Source: VentureBeat
4. Vonovia backs AI property management startup Immoly
Vonovia, Europe's largest residential landlord, announced on August 13 that it is supporting Berlin-based startup Immoly, which launched an AI-driven management platform for landlords. The platform handles tenant communication, document processing, and administrative workflows for residential property owners.
Real estate angle: When Europe's largest landlord backs an AI management platform, US property managers and the vendors selling to them will soon follow. The legal risk is in the automation: fair housing rules do not care whether a human or a model drafted the tenant response, and screening decisions made by AI still need a human to do a compliance review. If you work with landlords or property managers using these tools, ask who audits the output and who is liable when the platform gets something wrong.
Source: Vonovia
5. OpenAI's public S-1 is expected within weeks, with a September listing target
OpenAI filed its S-1 in June, and news this week indicates the public version could be released in late August, with a listing as early as mid-September. Reported pre-IPO financials show roughly $2 billion in monthly revenue against losses of about $1.22 for every $1 earned, at a private valuation of $852 billion. OpenAI has not committed to a timeline, and a 2027 listing remains possible.
Real estate angle: A public OpenAI means quarterly disclosures, which means we finally get numbers on what AI actually costs to run. That matters for anyone negotiating multi-year contracts with AI-powered vendors: pricing that looks generous today may be subsidized by venture capital that public shareholders will not tolerate. Build price escalation and exit clauses into vendor agreements accordingly.
Source: StartupHub.ai
The theme this week is money getting serious. Anthropic turned a profit, OpenAI is heading for public trading, and sovereign wealth funds are structuring data center deals like core infrastructure. The speculative phase of AI is giving way to balance sheets, lease agreements, and quarterly reporting. For real estate professionals, that means two things: the vendors selling you AI tools are becoming more stable, and the AI industry itself is becoming one of the largest real estate customers in the country.
- Jason