AMD's record $4.75B bond sale funds its AI chip push

Share
AMD's record $4.75B bond sale funds its AI chip push

AMD went to the debt market for the biggest raise in its history this week, and investors lined up — the latest sign that the AI buildout is now financed with bonds as much as with equity.

AMD priced $4.75 billion in investment-grade bonds Thursday — the largest bond offering in the company's history — to fund its accelerating AI infrastructure push. Bloomberg reported AMD planned to raise as much as $5 billion in what could be the chipmaker's biggest-ever investment-grade bond sale as it ramps up spending to meet AI demand, and Reuters confirmed the company was looking at $4 billion to $5 billion. By early Friday the deal was done at $4.75 billion across four tranches maturing in 2029, 2031, 2033, and 2036. Demand was strong enough that the longest-dated tranche priced about 25 basis points tighter than preliminary guidance, at 90 basis points over US Treasuries. Bank of America, JPMorgan, Barclays, and Wells Fargo led the sale, which is expected to settle August 17; proceeds will support general corporate purposes, including refinancing $875 million in bonds maturing next month.

The raise is the financial engine for AMD's most serious challenge to Nvidia yet. The company just locked in deals with Anthropic and Microsoft to scale up deployment of its AI chips, and earlier this month it forecast data-center sales would more than double by 2027 — we covered the shape of that ambition in AMD buys Taalas to hardwire AI models into silicon. Expansion on that scale costs more than AMD generates in cash, so it is borrowing: a notable choice, because Intel funded its own buildout with a $15 billion stock offering, while AMD opted for debt it will have to service regardless of how the AI cycle turns. That distinction is the quiet story here — the bond market is now effectively underwriting the AI capex cycle, and the durability of AI demand has become a credit question as much as a technology one.

Investors have answered for now: the tighter-than-guided pricing on AMD's longest bonds says the market wants in, and the company's data-center outlook — revenue more than doubling by 2027 — is the collateral. But a record bond sale is also a record obligation. If AI infrastructure spending wobbles, the leverage cuts both ways, and AMD, unlike Nvidia, still has to prove its accelerator business can generate the returns its debt assumes.

What to watch: settlement on August 17, and whether AMD's data-center growth actually tracks the doubling forecast that underwrites this debt.

The AI buildout is increasingly a debt-funded bet — are you comfortable with the bond market carrying it? 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