AI Summary. Hyperscaler cloud revenue is growing rapidly (AWS +37%, Azure +43%, Google Cloud +82%), with combined capital spending and R&D rising from $460bn to $870bn over the last 12 months.

Date Posted:

Since the start of the data-center buildout, S&P earnings breadth has “eclipsed” that which would be historically implied by the ISM manufacturing survey. Cembalest argues earnings growth “appears highly contingent” on continued hyperscaler spending.

Are hyperscalers investing faster than their revenue growth justifies?

Core argument: Hyperscaler capital spending and R&D nearly doubled from $460B in 2024 to $870B over the trailing twelve months, driving S&P 500 earnings breadth well beyond levels that ISM manufacturing surveys alone would predict—evidence of a powerful AI-led trickle-down across sectors.

The Q2 earnings season was positive for hyperscaler annual revenue growth: AWS 37%, Microsoft Azure 43% and Google Cloud 82%. Rising revenue backlogs reinforce evidence of AI monetization: Google Cloud’s backlog increased $52B to $514B, and the AWS backlog rose to $496B, up 36% q/q. Earnings growth is also more broad-based than it was in 2024-2025 when AI was the primary driver. Earnings growth appears highly contingent on continued hyperscaler capital spending and R&D which rose from $460B in 2024 to $870B over the last 12 months. The share of S&P 500 companies with positive expected earnings growth is normally well-predicted by the ISM manufacturing survey. But since the AI capital spending boom began, earnings breadth has eclipsed the improvement in manufacturing surveys; this suggests a powerful trickle-down effect from hyperscalers to other sectors.

Takeaways by Macro Roundup® AI

  1. Hyperscaler capital spending and R&D nearly doubled from $460B in 2024 to $870B over the trailing twelve months, driving S&P 500 earnings breadth well beyond levels that ISM manufacturing surveys alone would predict—evidence of a powerful AI-led trickle-down across sectors.

AI Summary. Big Tech hyperscalers reported over $160bn in paper gains from stakes in other AI companies last quarter, inflating headline profits beyond what underlying operations generated. At Alphabet and Amazon, profit growth was driven by investment gains on holdings such as SpaceX and Anthropic rather than new revenue or cash flow.

Date Posted:

In Q2, Alphabet, Amazon, Nvidia and Microsoft reported $160B in pre-tax profit from valuation gains on AI-related equity stakes alone.

Are Big Tech profits masking weakness in core business operations?

Core argument: Big Tech hyperscalers booked more than $160bn in pre-tax gains last quarter from stakes in AI companies including SpaceX and Anthropic, with Alphabet and Amazon attributing record profit growth primarily to “other income” rather than core business expansion.

Big tech giants booked a more than $160bn windfall last quarter from investments in other AI companies, flattering their earnings and raising concerns that paper gains are overstating the strength of the AI boom. In the most recent round of earnings reports, pre-tax profits reached a record at several of the Big Tech “hyperscalers”. But the main source of these increases at Alphabet and Amazon was ‘other income’ from gains on SpaceX and Anthropic, rather than new business lines or fresh cash generation. “A couple of years ago investors started to ask about the circularity of [Big Tech] revenues,” said Ben Snider, chief US equities strategist at Goldman Sachs. The profit boost from AI investments now “raises the question of whether the growth these companies are reporting is based on underlying demand or if it is misleading in some way,” he added.

Takeaways by Macro Roundup® AI

  1. Big Tech hyperscalers booked more than $160bn in pre-tax gains last quarter from stakes in AI companies including SpaceX and Anthropic, with Alphabet and Amazon attributing record profit growth primarily to “other income” rather than core business expansion.
  2. The concentration of Big Tech earnings gains in unrealized investment income rather than operating cash flow signals that headline AI-era profitability overstates the sector’s underlying commercial monetization.

AI Summary. Rising long-term bond yields reflect higher expected short-term interest rates over the next decade, not concerns about government debt sustainability, as both inflation expectations and the risk premium for holding long-term bonds have remained stable.

Date Posted:

Citing a stable term premium, Miran argues the rise in the 10-year Treasury yield most likely reflects investors “marking up their expectations for long-run economic growth, not becoming concerned over central bank or fiscal credibility.”

Are rising bond yields signaling higher future rates or debt concerns?

Core argument: The rise in ten-year real Treasury yields is driven almost entirely by higher expected long-run overnight rates, not by fiscal credibility concerns or debt sustainability fears.

Ten-year real yields are equal to the average expected overnight yield over the next ten years, plus a risk premium that investors demand for locking up their money for ten years. This term premium reflects the fact that realised overnight rates might deviate from expected overnight rates. If the market saw fiscal credibility or debt sustainability problems, this would show up in an increased term premium. But the term premium has also been stable and is only a hair lower since the end of last year. With term premium and inflation expectations contained, the recent increase in bond yields is therefore almost entirely due to higher expected overnight rates over the long term.

Takeaways by Macro Roundup® AI

  1. The rise in ten-year real Treasury yields is driven almost entirely by higher expected long-run overnight rates, not by fiscal credibility concerns or debt sustainability fears.
  2. The term premium on ten-year Treasuries remains stable and only marginally lower since year-end, ruling out market pricing of sovereign fiscal stress as a driver of yield moves.
  3. Contained inflation expectations and a flat term premium together confirm that bond market signals reflect monetary policy rate expectations, not structural deficit risk.

AI Summary. The gap between long-term government bond yields and equivalent swap rates has narrowed to its smallest in months, reflecting increased investor willingness to hold long-dated government debt following expanded buybacks of longer-dated bonds.

Date Posted:
Is Database:
Database

Modest compression of the spreads between Treasury yields and synthetic “swap” securities (~5.5bp for the 30 year and ~3bp for the 10 year) suggest Bessent’s Treasury purchase program has had a degree of success at lowering long-term government yields.

Are investors returning to long-term government bonds?

Core argument: The 30-year Treasury-swap spread narrowed to its smallest since February following Bessent’s announcement, as Treasuries outperformed equivalent-maturity swaps and benchmark yields drifted lower.

Since Bessent’s announcement, Treasuries have outperformed equivalent-maturity swaps, narrowing the 30-year spread to the smallest since February. Swaps are popular with some investors as an alternative to owning bonds; the gap between [swap rates] and US government yields [gauges] how willing [investors] are to hold Treasuries instead. The 10-year swap spread has compressed too, with the gap three basis points smaller at around 38 basis points. Still, the recent drop has only dented a years-long rise in long-term US government borrowing costs. The 10-year US yield inched up 3bp to 4.66% after touching 4.75% last week. “While conducting buybacks at the long end of the yield curve may technically decrease yields, higher structural US budget deficits, which [require] a significant supply of Treasuries to finance the US debt, [are] not changing anytime soon,” said Libby Cantrill, head of public policy at Pimco.

Takeaways by Macro Roundup® AI

  1. The 30-year Treasury-swap spread narrowed to its smallest since February following Bessent’s announcement, as Treasuries outperformed equivalent-maturity swaps and benchmark yields drifted lower.
  2. The Treasury’s plan to at least double longer-dated bond buybacks drove the repricing.
  3. the swap-yield gap—a direct gauge of investor preference for Treasuries over derivatives—compressed in response.

AI Summary. Treasury intervention in a functioning bond market suppresses the price signal that transmits collective market information to decision makers, removing the mechanism by which orderly volatility performs its intended economic function.

Date Posted:

Stan Druckenmiller cautions that there is no question that the Treasury’s efforts to cap long-term yields will fail: “Governments defending prices against fundamentals always lose. The only variable is how much they spend before conceding.”

Core argument: The long-term Treasury yield is the world’s most consequential price signal.

Markets aggregate information no committee possesses, and prices are how that information reaches decision makers. The long-term Treasury yield is the most important price in the world. Treasury’s announcement gave the game away. It justified the larger operations as liquidity support in sectors with “consistent strong sponsorship from market participants,” but strong sponsorship is the definition of a healthy, working market. There were no failed auctions, no dealer balance-sheet seizure, no forced unwinds, nothing resembling Treasurys in March 2020 or U.K. gilts in September 2022, the sort of genuine dysfunctional episodes that justify official action. Volatility was contained, and trading was orderly—not a malfunction but the machine doing its job.

Takeaways by Macro Roundup® AI

  1. The long-term Treasury yield is the world’s most consequential price signal.
  2. unwarranted official intervention corrupts the market-aggregated information that no policy committee independently possesses.

AI Summary. The safety premium that investors historically paid for US government bonds over equivalent alternatives has compressed toward zero and, at longer maturities, reversed — with global investors now pricing foreign government bonds as safer than US Treasurys.

Date Posted:
Is Database:
Database

Lustig shows the premium investors pay for Treasurys over substitutes such as AAA corporate debt and G10 sovereign debt has compressed post 2020. “Investors are now indifferent between [Treasurys] and close substitutes.”

Are global investors losing confidence in US government debt safety?

Core argument: The U.S. Treasury safety premium has compressed to near zero post-2022 and at points reversed, signaling that markets no longer treat Treasurys as unambiguously superior to credit-risk-adjusted corporate equivalents.

The top panel uses the credit risk-adjusted AAA-Treasury spread. We use the CDS to strip the default-risk compensation out of the corporate-bond yield. What remains is a clean estimate of the safety premium that investors pay for Treasurys over otherwise-equivalent corporate exposure. Post-2022, it has compressed toward zero, and at points, has reversed. The bottom panel uses the Treasury Premium, defined as the difference between the synthetic-dollar foreign sovereign yield and the US Treasury yield at the same maturity. The synthetic-dollar foreign yield is constructed by swapping the coupon payments on foreign G10 sovereign bonds into dollars using the foreign-exchange forward market. This eliminates currency risk over the life of the bond, so the resulting dollar cash-flow stream is directly comparable to a US Treasury yield of the same maturity. At longer maturities, global investors now seem to prefer the safety of foreign G10 bonds.

Takeaways by Macro Roundup® AI

  1. The U.S. Treasury safety premium has compressed to near zero post-2022 and at points reversed, signaling that markets no longer treat Treasurys as unambiguously superior to credit-risk-adjusted corporate equivalents.
  2. At longer maturities, global investors now price dollar-hedged G10 sovereign bonds above U.S. Treasurys, marking a structural erosion of the safe-haven premium that has historically anchored U.S. borrowing costs.

AI Summary. AI model pricing is falling as competition intensifies, with leading models cutting token costs by up to 80%. Higher-priced models can deliver lower total costs by completing tasks in fewer tokens or attempts.

Date Posted:

Token prices for the leading US AI labs are falling. OpenAI cut the price of its mid-tier GPT-5.6 Luna by 80%, while Anthropic reduced the price of its Opus 5 by 50% relative to its frontier Fable 5 model.

Does cheaper AI pricing actually reduce total costs for users?

Core argument: Headline token prices are an unreliable cost proxy, as higher-capability models completing tasks in fewer tokens or attempts can deliver lower total cost despite carrying a higher per-token list price.

OpenAI cut the price of GPT-5.6 Luna from $1 to $0.20 per mn input tokens and from $6 to $1.20 per mn output tokens. Anthropic launched Opus 5 at $5 per mn input tokens and $25 per mn output tokens — half the price of its Fable 5 model. This week, the company called off a planned rise in prices for its Sonnet 5 model, which had been due to take effect from September. Headline token prices do not provide a straightforward comparison between AI models, however. More capable models can sometimes complete a task using fewer tokens or with fewer attempts, meaning a model that appears more expensive based on the headline price of tokens can ultimately cost less.

Takeaways by Macro Roundup® AI

  1. Headline token prices are an unreliable cost proxy, as higher-capability models completing tasks in fewer tokens or attempts can deliver lower total cost despite carrying a higher per-token list price.

AI Summary. The US government sold $25bn in 30-year bonds at a 5.216% yield, the highest rate since 2001, as investors demanded greater compensation to finance a growing national deficit.

Date Posted:

Though a $25B auction of 30-Year Treasurys drew “decent” demand, the yield rose to 5.216% — the highest 30-year yield at auction since August 2001.

Are rising bond yields signaling investor alarm about fiscal sustainability?

Core argument: The U.S. Treasury’s $25 billion 30-year bond sale cleared at a 5.216% yield—the highest since 2001—as investors demanded elevated compensation to finance the nation’s expanding fiscal deficit.

The US government sold 30-year bonds at the highest interest rate in a quarter of a century, a testament to investors’ demand for compensation to finance the nation’s growing deficit. The yield at the $25 billion sale Thursday came in at 5.216%, the most since 2001, even as a drop in oil prices supported US debt in secondary-market trading. The sale was met with decent demand. The yield at Thursday’s 30-year sale was a little above the prevailing level seen in the market before the 1 p.m. bidding deadline in New York—a sign that demand slightly lagged expectations.

Takeaways by Macro Roundup® AI

  1. The U.S. Treasury’s $25 billion 30-year bond sale cleared at a 5.216% yield—the highest since 2001—as investors demanded elevated compensation to finance the nation’s expanding fiscal deficit.
  2. Demand at the 30-year auction slightly undershot expectations, with the clearing yield pricing above pre-deadline secondary-market levels, signaling investor reluctance at current deficit trajectories.