Anthropic’s $60bn ‘North Star’ Deal Signals Soaring Cost of AI Ambition
Key Takeaways:
- Escalating Capital Demands:Anthropic’s “Project North Star” highlights the extraordinary capital intensity of AI development, with a staggering $60bn debt package sought for Google TPUs, setting a new benchmark for chip financing.
- Rising Cost of Capital:The deal’s offer of a 10%+ yield for junior debt, a notable increase from previous financings, underscores investors’ demand for higher risk premiums amid the rapid, large-scale AI infrastructure build-out and growing lender caution towards tech’s heavy capex.
- Evolving Financing Landscape:Private capital giants like Blackstone are stepping in to provide bespoke, high-yield debt solutions for AI firms, potentially paving the way for these specialized assets to become tradeable in institutional markets, while strategic guarantees from partners like Broadcom mitigate risk.
Unlock the Editor’s Digest for free
Roula Khalaf, Editor of the FT, selects her favourite stories in this weekly newsletter.
In a financial maneuver that underscores the insatiable capital demands of the artificial intelligence revolution, Anthropic is reportedly offering private equity behemoth Blackstone a high yield exceeding 10 per cent to finance its latest, massive chip procurement deal. This move signals a critical juncture for AI start-ups, highlighting their willingness to absorb steep borrowing costs in the race to secure indispensable AI infrastructure.
Dubbed “Project North Star,” this colossal $60bn debt package is earmarked to fund Anthropic’s acquisition of approximately 1.5 gigawatts of custom semiconductors – specifically, Google’s tensor processing units (TPUs). This unprecedented scale positions it as the largest chip financing to date, serving as a crucial litmus test for investor appetite in AI-specific debt. The market is increasingly scrutinizing tech companies’ burgeoning capital expenditure and rising leverage, creating an environment where lenders demand greater compensation for risk.
Sources familiar with the matter reveal that Blackstone funds are poised to inject $9bn into this debt, which is secured by Anthropic’s five-year lease payments on these cutting-edge chips. This investment represents roughly half of the $18bn junior loan tranche. Blackstone was reportedly offered an initial coupon of 10 per cent, with the total return anticipated to climb even higher once new-issue discounts and other structuring fees are factored in. This pricing stands in stark contrast to the public bond market, where a borrower with a comparable credit profile might expect to pay closer to 7 per cent, according to the ICE BofA BB US High Yield Index. The remaining $9bn of junior debt is slated for distribution to a broader syndicate of investors in the coming months, potentially at a slightly lower yield, as anchor investors often command a premium for early, significant commitments.
Neither Blackstone nor Anthropic offered immediate comments on the developing deal, which is consistent with the private nature of such large-scale transactions.
Just four months prior, Anthropic executed a similar financing, “Project Big Sky,” securing approximately $4.4bn in debt at an interest rate of 8.5 per cent, as reported by the Financial Times. The notable surge in borrowing costs between these two deals—from 8.5 per cent to over 10 per cent—is a stark indicator of the market’s evolving perception of risk and reward within the AI sector. Investors are now demanding higher risk premiums to fund the intensive AI build-out, grappling with the twin challenges of tech companies’ monumental funding requirements and the compressed timelines for deployment. This escalating cost of capital is also a function of the sheer size of these mega-financings, where borrowers are paying up to attract crucial anchor investors capable of committing substantial capital early in the process.
This financial dynamic also throws into sharp relief the strategic challenges confronting Anthropic, a five-year-old AI lab, as it navigates its path towards achieving an coveted investment-grade rating. Blackstone, in its discussions with investment partners, reportedly pitched “Project North Star” as a unique opportunity to capture attractive high yields on Anthropic-linked debt, with the long-term expectation that the company would eventually secure an investment-grade rating post-public listing, thereby significantly reducing its future borrowing costs. While Anthropic was assigned a one-notch-below-investment-grade rating by KBRA in June, the latest junior debt tranche is anticipated to receive a similar assessment, underscoring the gap between current credit perception and future aspirations.
The comprehensive debt package also includes a substantial $42bn of senior debt, critically guaranteed by Broadcom. Broadcom, a key player in semiconductor development, collaborates with Google on the creation of these specialized chips, offering a layer of supply chain and technological assurance to lenders. Upon the full drawdown of capital, anticipated by 2027, this senior debt could potentially become tradeable in the institutional 144A bond market. This future marketability suggests a growing trend towards structuring bespoke, illiquid private credit deals with an eye towards eventual secondary market liquidity, appealing to a broader base of sophisticated institutional investors.
Market Impact:
Anthropic’s “Project North Star” is more than just another financing deal; it’s a bellwether for the broader AI market, signaling a new era of extraordinary capital intensity and a re-evaluation of risk by lenders. The escalating cost of capital, demonstrated by the rising yields demanded, will likely put pressure on other AI start-ups and even established tech giants to secure funding for their compute infrastructure. This could lead to a bifurcation in the market, favoring well-capitalized players or those with robust strategic partnerships capable of providing guarantees. For investors, the deal highlights the growing prominence of private credit as a vital funding source for high-growth, high-capex sectors, offering attractive yields in an environment where traditional public markets might be more cautious. It also sets a precedent for how specialized AI infrastructure — from TPUs to data centers — will be financed, potentially paving the way for new classes of securitized or tradeable debt instruments tied to the future performance and strategic value of AI assets. Ultimately, the success and future trading of this debt will influence the investment strategies of institutional funds and the ongoing valuation of AI innovation across the financial ecosystem.

