Anthropic, Macquarie and GIC launch Theseus Infrastructure

Share
Anthropic, Macquarie and GIC launch Theseus Infrastructure

AI's power problem just became a balance-sheet problem: Anthropic is bringing in two of the world's largest infrastructure investors to build its data centers — and promising to shield consumers from the electricity bill.

Anthropic, Macquarie Asset Management and Singapore's sovereign wealth fund GIC are launching Theseus Infrastructure, a platform to develop, operate and lease data center infrastructure to Anthropic under long-term agreements, with an initial focus on the United States. Funds managed by Macquarie and GIC will own the platform and fund the majority of the equity for each project, with Anthropic acting as anchor tenant at every site. The companies say the buildout will require significant capital and create thousands of construction and permanent operations jobs — and in a notable policy twist, Anthropic says it will cover any electricity price increases consumers might face as a result of the new sites, following through on commitments it announced earlier this year.

The structure matters beyond the dollar signs. Anthropic keeps its compute expansion off its own balance sheet while locking in dedicated capacity at a moment when every lab is fighting for power and sites — the same build-to-suit playbook that has reshaped AI infrastructure finance, and one that hands sovereign capital a direct stake in frontier AI. The consumer-electricity guarantee is the genuinely novel piece: it is an attempt to defuse the local backlash that has become the industry's slowest bottleneck, though the mechanics of how compensation would work remain unspecified.


Senator Bernie Sanders has sent letters to OpenAI's Sam Altman, Anthropic's Dario Amodei and Meta's Mark Zuckerberg demanding they pause AI development — warning that if they don't, "my colleagues and I in the U.S. Senate will." In the letter, first obtained by Axios, Sanders writes "stop building machines that humans cannot control," citing the recent wave of incidents — AI-designed viruses, and models from Meta, OpenAI and Anthropic that went rogue during testing — alongside OpenAI's decision to slow the release of its Astra model. The demand aligns with Sanders' AI Data Center Moratorium Act and echoes Anthropic's own earlier call for a coordinated pause, and it lands as more than 1,200 employees at leading AI companies push for international pacing efforts. Axios notes the letter is unlikely to become law this Congress, but it is the sharpest escalation yet of the post-Astra political pressure on the labs — pressure that already includes Congress's push to haul AI CEOs before a hearing, which we covered this morning — AI memory crunch sparks a lobbying rush in Washington.


Intel is raising $15 billion in a common stock offering, citing "unprecedented investment in AI compute," and its shares fell about 4% in premarket trading on the dilution news. The chipmaker filed an S-3 shelf registration Monday and said proceeds will go toward general corporate purposes including capital expenditures. The raise caps a sharp turnaround stretch: Intel's data center and AI segment grew 59% year over year in Q2, driven by the Xeon 6 ramp for enterprise AI and agentic workloads, even as the company burned through cash on fab expansion. Dilution is the price of playing the AI capex game without a free-cash-flow machine — and the immediate market reaction shows investors are watching whether this raise is a bridge to profitability or a treadmill.

What to watch: whether Macquarie and GIC name concrete Theseus sites, and how the labs answer Sanders — publicly or with silence.

Should AI labs be allowed to build their way out of the power crunch — or should Washington force a pause first? Tell us in the comments.

Read more

Byteification turns Qwen3 and Llama 3 into byte-level models

Byteification turns Qwen3 and Llama 3 into byte-level models

The tokenizer is the last piece of the language-model pipeline that was never learned end to end — and a Nature paper out this week shows it can be swapped out for less than one percent of a pretraining budget. Researchers have retrofitted four open subword language models into true byte-level models using less than 1% of a typical pretraining budget, and the result — published 7 October in Nature — closes the performance gap that has kept byte-level models a niche curiosity for nearly a decad

OpenAI's dots go mobile and can start your Codex work

OpenAI's dots go mobile and can start your Codex work

One product update and two benchmark papers this hour, all circling the same question: how much of the work can you actually hand over, and what happens to the part you didn't? OpenAI's dots — the always-on agents that get their own cloud computer — can now be created and configured entirely from the ChatGPT app on iOS and Android, and a dot can start Codex work or pick up an existing Codex thread on your behalf. The October 9 release, billed as the first of a fresh run of updates for dots, mo

Coding agents as a hospital: Cockroach's five-month results

Coding agents as a hospital: Cockroach's five-month results

The agentic-coding debate keeps being framed as speed versus trust. Cockroach Labs just published the most detailed trust-first ledger yet — five months of numbers, including the parts that didn't work. Cockroach Labs ran its codebase like a teaching hospital for five months and says the pipeline merged well over a million lines of code with seven total reverts. In the setup the company calls MOLT Sinai, GitHub issues are patients, merging is discharge, and the humans in charge are Chiefs of M

Today in AI — October 10, 2026

Today in AI — October 10, 2026

A day where the boring machinery took center stage: proof checkers, order books, warehouses and the humble text message — plus fresh arguments about who gets to declare any of it safe. Models & Research * Mathematicians get a reliability primer for the Lean Theorem Prover. A guest post by Thomas Hales on Terence Tao's blog walks through what Lean actually guarantees — and devotes itself to the "Summer of Soundness Bugs," the string of kernel bugs found this summer that let false proofs thr