5 Things to Know About AI This Week
A quick scan of what happened in AI this week.
1. OpenAI Filed Confidential IPO Papers Targeting a $1 Trillion Valuation
OpenAI submitted a confidential draft registration statement to the SEC on May 22, 2026. The company is targeting a September listing at a valuation between $852 billion and $1 trillion. OpenAI is generating roughly $2 billion per month in revenue and has around 50 million consumer subscribers and 9 million business users. Despite the top-line numbers, the company lost $1.22 for every $1 earned in Q1 2026.
Real estate angle: If OpenAI goes public in September, the pricing models for ChatGPT, API access, and enterprise tools will become important to shareholders. Tools you use today could see pricing shifts or feature restructuring soon after an IPO. This is a good time to audit which AI subscriptions your operation actually uses and at what price point, before a public market debut changes the numbers.
Source: CNBC
2. Google Launched Its First 24/7 Consumer AI Agent at I/O 2026
At its annual developer conference on May 19, Google announced Gemini Spark, a general-purpose AI agent that runs continuously and reasons across connected apps. The company also cut its Google AI Ultra subscription from $250 to $100 per month and released Gemini 3.5 Flash, a lighter model priced at roughly one-third the cost of comparable alternatives. Gemini Spark is entering beta for AI Ultra subscribers first.
Real estate angle: If your business is centered around Google Workspace, a persistent AI agent that connects to your email, calendar, documents, and client communications gives you a major upgrade. But before you connect your client database or transaction files to a product like this, you need to know: Where does the data go? Who can access it? Does connecting your client contact list to a third-party agent create any disclosure obligations under your representation agreements? Make sure you know before you integrate this into your workflow, not after.
Source: CNBC
3. A Supply Chain Attack Exposed 3,800 GitHub Repositories, Including AI Tool Credentials
On May 18, a fraudulent version of Nx Console, a popular VS Code extension with 2.2 million installs, was published to the Visual Studio Code Marketplace. The fake extension was live for only 18 minutes, but that was long enough for a credential-stealing payload to harvest GitHub tokens, AWS keys, npm credentials, and Anthropic Claude Code configuration files. The hacker, identified as TeamPCP, illegally downloaded approximately 3,800 internal GitHub repositories. OpenAI, Mistral AI, and Grafana Labs were also affected through a related npm supply chain compromise.
Real estate angle: You may not use VS Code directly. But it's a popular platform behind a lot of apps you may use for transaction management, e-signatures, or AI tools. A breach that harvests cloud credentials can filter through SaaS products built on compromised infrastructure. This is a reminder to check with any software vendors about whether they were affected by any recent hacking incidents.
Source: Bleeping Computer
4. Anthropic Is Projecting Its First Operating Profit on $10.9 Billion in Q2 Revenue
Anthropic, the creator of Claude, announced it is on track for a $559 million operating profit in Q2 2026 on $10.9 billion in revenue, up from $4.8 billion in Q1. The growth is attributed to enterprise adoption of Claude for software development and Mythos, a security-focused model. A detail buried in SpaceX's IPO filing revealed that Anthropic is paying SpaceX $1.25 billion per month for access to GPU computing power through May 2029, a $45 billion total commitment.
Real estate angle: The $15 billion annual compute bill tells you something about the economics of AI products. Companies spending at that scale are not building sustainable businesses around permanent free tiers or deeply discounted enterprise contracts. Anthropic's first quarterly profit is a sign the industry is maturing. It also means enterprise AI pricing will start reflecting real costs. If your operation has built workflows around a free or low-cost AI tier, budget for pricing changes within the next 12 to 18 months. Lock in annual contracts where you can.
Source: CNBC
5. Pope Leo XIV Released an Encyclical Framing AI as the Industrial Revolution of Our Time
On May 25, Pope Leo XIV published his first papal encyclical, Magnifica Humanitas. The document focuses on protecting the human person in the age of artificial intelligence and compares AI with the Industrial Revolution. It calls for safeguards around dignity, labor, and autonomous decision-making. The Vatican's position follows years of Pope Francis raising AI ethics concerns at international forums, and it arrives as state and federal regulators are also moving to define oversight frameworks for high-stakes AI applications, including those used in housing and lending decisions.
Real estate angle: AI has cleared the early adopter phase when an institution with 1.4 billion followers issues a formal teaching document about it. Your clients are forming opinions about AI from sources beyond pure tech outlets. That has practical implications for how you communicate when AI touches a transaction, whether through automated valuations, document review, or fraud screening. A plain, one-paragraph explanation of what AI does and does not do in your process is not optional anymore. It is part of informed consent.
Source: Build Fast With AI
The common thread this week is scaling. OpenAI is pursuing a trillion-dollar public listing. Anthropic is projecting $10.9 billion in quarterly revenue on a $15 billion annual compute contract. Google launched a 24/7 AI agent for consumers at a $100 monthly price point. AI is no longer in its experimental era as Swifties would say. For real estate professionals, the practical question is not whether AI is here. It is which AI commitments your operation has made, and whether you understand the implications of using them.
- Jason