business · Echo Wang/Reuters
Anthropic IPO filing shows rapid growth alongside heavy losses and compute commitments
Confirmed: Reuters reports that Anthropic’s draft IPO prospectus shows 2025 revenue of about $4.6 billion, an operating loss of about $8.06 billion and a net loss of about $42 billion, including roughly $34 billion of accounting charges. The filing also discloses about $518 billion of future cloud and infrastructure spending commitments, while roughly a quarter of revenue came from two customers.
Why it matters: The filing shows both rapid revenue growth and the capital intensity of frontier AI. Operating losses, non-cash accounting charges, long-term infrastructure commitments and customer concentration are separate risks and should not be collapsed into a single headline loss figure.
Our read: The figures provide evidence of a business scaling quickly while carrying substantial operating costs, accounting effects and future infrastructure obligations. They do not establish how investors will value those risks or whether the spending commitments will produce sustainable returns.
Open the source for story 1
technology · worldlabs.ai
AMD agrees to acquire World Labs for $8.2 billion
Confirmed: AMD has agreed to acquire spatial-intelligence startup World Labs in an all-stock deal valued at about $8.2 billion. World Labs co-founder Dr. Fei-Fei Li is set to join AMD as Executive Vice President and Chief Scientist.
Why it matters: The acquisition would add World Labs’ spatial and world-model research to AMD as the chipmaker expands beyond hardware into a broader AI technology portfolio.
Our read: The deal is a significant move into spatial AI, but the acquisition alone does not establish that AMD will own an end-to-end software-and-hardware stack or how quickly World Labs’ research will translate into commercial products.
Open the source for story 2
ai · anthropic.com
Anthropic releases Claude Sonnet 5.5 with migration changes for some integrations
Confirmed: Anthropic has released Claude Sonnet 5.5 and says it generates tokens more than 30% faster than Sonnet 5 while reducing cost per task by up to 30% in its measurements. Anthropic also documents breaking migration changes affecting specific API behaviours, including some tool-call and thinking-block handling.
Why it matters: Teams using the affected API behaviours may need code changes when migrating. Developers who do not rely on those behaviours should not assume that every Sonnet integration requires immediate modification.
Our read: Anthropic is emphasising faster, lower-cost agentic workloads, but the performance and cost figures are company claims and migration effort will depend on each integration.
Open the source for story 3
business · Garrett Downs/CNBC
Khanna plans Human Control Over AI Act as Congress debates frontier-AI rules
Confirmed: CNBC reports that Rep. Ro Khanna plans to introduce the Human Control Over AI Act, a proposal that would impose strict liability on frontier-model developers and restrict recursive self-improving AI until specified government safeguards are in place. CNBC reports that floor votes are not expected before the midterm elections.
Why it matters: The proposal illustrates one approach being discussed for federal AI oversight, but it is not enacted law and its final bill text, committee path and prospects will depend on the congressional process.
Our read: The proposal adds to the policy debate over developer liability and human control of frontier systems. It should be treated as a legislative proposal rather than evidence that a new federal regulatory regime has already been established.
Open the source for story 4
Practical takeaway
If you use Anthropic’s API, check the Sonnet 5.5 migration notes against the specific tool-call and thinking-block behaviours your integration uses before switching models. When assessing frontier-AI providers, separate operating performance from non-cash accounting charges and consider infrastructure commitments and customer concentration alongside revenue growth.
What to watch next
Watch for Anthropic’s IPO filing to move through SEC review, for further detail on its infrastructure commitments and customer concentration, for real-world Sonnet 5.5 migration reports, and for Khanna to formally file the Human Control Over AI Act and publish its final legislative text.