China’s Broad Economic Slowdown Raises Stimulus Expectations
(…) Production at Chinese factories and mines expanded 5.2% last month from a year earlier, according to data released by the National Bureau of Statistics on Monday, the smallest gain since August 2024.
Retail sales grew 3.4% on year in August, down from 3.7% in the previous month. Expansion in fixed-asset investment in the first eight months of the year decelerated sharply to 0.5%, the worst reading for the period on record except for the pandemic year of 2020. (…)
“The high base from the fourth quarter of 2024 suggests that we probably will see fourth-quarter growth slowing more significantly, jeopardizing the government’s 5% growth target if no major stimulus measures rolled out.”
With a boom in exports cooling off, many analysts and investors expected a downshift in China’s economy during the final months of 2025 after it clocked growth of 5.3% in the first half. The extent of the deceleration in China, set to be the top contributor to global growth over the next five years, will matter to a vulnerable world economy that’s slowing under pressure from Donald Trump’s tariffs.
Sales of subsidized consumer goods such as home appliances, furniture and communication devices all decelerated from the previous month, partly due to the statistical effect of a high base from 2024 as authorities began to ramp up their trade-in program around the same time last year. (…)
A broad measure of credit slowed last month for the first time this year, while export growth fell short of forecasts and dropped to 4.4% in August. The labor market also likely weakened in recent months, based on purchasing managers’ index surveys and private polls.
Another source of pressure for the economy is the government’s “anti-involution” campaign that aims to ease overcapacity and excessive competition among companies. The effort escalated in early July and may have contributed to a fall in output that month for products ranging from steel to copper. (…)
- August exports missed expectations but remained solid (Goldman Sachs)
Chinese export growth slowed from 7.2% yoy in July to 4.4% yoy in August, missing expectations. But given the 30% US tariffs added this year, this is a solid print.
Exports to the US did decline by around 30% from the 2024 levels. But increases in exports to ex-US more than offset the decline. During the first eight months of this year, Chinese exports grew 5.9% yoy in nominal USD terms and trade surpluses totaled US$785bn. 2025 is on track to be the first year of Chinese goods trade surplus exceeding US$1tn.
Housing Slump Ripples Through Manufacturing Economy Air-conditioner manufacturer Carrier and plumbing product distributor Core & Main warned of a steep whipsaw in residential demand.
Air conditioner and heating manufacturer Carrier Global Corp. this week warned that volumes in its residential business would be 40% below last year’s levels, a much weaker rate of demand than the company was forecasting only about six weeks earlier. Data from the Air-Conditioning, Heating and Refrigeration Institute showed a similar 33% plunge in US air-conditioner shipments in July and an 18% drop in deliveries of heat pumps, suggesting the trend isn’t just confined to Carrier. (…)
New single-family home sales could decline by low- to mid-single digits this year, a much bigger decrease than initially expected that compares with a 3% increase in 2024, Bloomberg Intelligence analyst Drew Reading estimates. That’s leading to a slowdown in lot development activity and curtailing purchases of goods like air conditioners and new plumbing that are typically associated with housing turnover.
Earlier this week, Core & Main Inc., a distributor of pipes, valves and other water products, lowered its sales and profit outlook for 2025 because of sluggish residential demand. (…) Core & Main had expected its residential business to be flat overall this year; it’s now calling for a low double-digit sales decline. (…)
Consumers that might normally have replaced their [HVAC] units in the fall “may just wait until the spring,” he said, giving the high rate of economic anxiety. (…)
Carrier is “aggressively” reducing its own production output and “our manufacturing facilities are much less active than what we expected them to be,” Chief Financial Officer Patrick Goris said. (…)
The housing markets in Florida and the Southeast had stayed more robust even as building in other areas of the country slowed, helping to prop up overall residential demand, he said. But that’s now changing with those regions seeing a material drop-off in the most recent quarter that’s continued into the current period, Witkowski said. (…)
Power-tool manufacturer Stanley Black & Decker Inc. has also lowered its outlook since the beginning of the year amid a weaker-than-expected consumer spending backdrop and the squeeze from tariffs. The company initially forecast a low-single digit sales gain for its outdoor equipment and tools business in a recovery from several years of weakness after the pandemic boom in demand. As of its last earnings update in July, Stanley now expects revenue for that unit to be down about 1% this year. “It does seem like the consumer and especially the DIY consumer or anybody buying higher price point items, they kind of ebb and flow with the political mood,” CFO Patrick Hallinan said at the Morgan Stanley conference this week. (…)
In addition to the warnings coming out of the housing market, McDonald’s Corp. said last month that traffic by lower-income customers at quick-service restaurants dropped by double digits industry-wide in the most recent quarter versus the year-earlier period. Middle-income Americans are also feeling economic pressure and pulling back on spending, CEO Chris Kempczinski told CNBC earlier this month.
Colgate-Palmolive Co., meanwhile, is seeing customers dial back their toothpaste and toothbrush orders, keeping just one tube lying around instead of an inventory of two or three and using their toothbrushes beyond the recommended three months, CEO Noel Wallace said at a Barclays Plc conference this week. (…)
(…) Consumers expect prices to rise at an annual rate of 4.8% over the next year, unchanged from the previous month, data released Friday showed. They saw costs rising at an annual rate of 3.9% over the next five to 10 years, a jump from the 3.5% rate seen last month.
“Consumers’ expected probability of personal job loss grew sharply this year and ticked up in September as well,” Joanne Hsu, director of the survey, said in a statement, “suggesting that consumers are indeed concerned that they may be personally affected by any negative developments in labor markets.”
“Moreover, consumers also feel squeezed by the persistence of high prices,” she added. (…)
Gauges of sentiment among Republicans and independents slipped to four-month lows, while the outlook among Democrats improved slightly.
Household net worth jumped $7.1 trillion, or 4.2%, from the prior quarter to $176.3 trillion, a Federal Reserve report showed Thursday. The value of Americans’ equity holdings rose $5.5 trillion, and the value of real estate also climbed.
The value of real estate holdings increased $1.2 trillion, the first increase in a year.
The Fed’s report showed that consumer borrowing rose at a 3.8% annualized pace, the fastest pace in almost three years. Growth in consumer non-mortgage credit advanced by the most since 2023, while mortgage debt also picked up.
California Trucking Firms Go Under, Fueling Wider Industry Fears No recovery from yearslong slump in sight as imports, factory activity and other drivers of demand sag
(…) America’s trucking companies are battling one of the longest freight slowdowns executives can remember. After three years of rock-bottom freight rates and rising costs that pushed weaker carriers out of business, analysts and executives are watching to see which well-established carriers may be next. (…)
Trucking is a cyclical industry in which rates rise and fall based on the supply of drivers and demand for freight. The lows usually last no longer than about 18 months. Cargo rates tanked in 2022, yet there is still no recovery in sight as demand drivers such as imports and domestic factory activity sag. (…)
Vise said that although tens of thousands of small trucking firms have gone out of business in recent years, the sector still has more trucks on the road than before the pandemic and more than needed to meet current demand. At the same time, carriers are battling rising costs for labor, equipment and insurance. (…)
Trucking companies had been hoping freight rates would recover in the second half of this year. Instead, rates have continued to bounce along at about the same level they were at before the pandemic, said Dean Croke, principal analyst at DAT Freight & Analytics. (…)
Will the Rally Continue?
Hubert Marleau says yes:
(…) While considerations should be given traditional valuation metrics like free cash flow (FCF), price to earnings ratios (P/E), and debt-to-equity ratios (D/E), nothing works as well as the PEG ratio – P/E divided by growth prospects. In this connection, stocks should scale up to new all-time highs over the next 12 months. Why? Because 7 big things are expected to keep the growth trajectory of per share corporate profits intact, thereby likely outweighing the risk of bad news.
First, capital expenditures are unusually strong and so are prospects. Spending by hyperscalers on AI data centres is robust, and all the entire outlook remains unscathed. Moreover, big investment incentives in the federal budget will become law in 2026.
Second, the US economy is producing very solid productivity gains, rising at an annual rate of 2.5% in Q2, and will likely trend even higher in Q3 and in 2026 as more and more firms adopt AI, automation and robotics to replace workers, raise efficiencies, and widen their customer base. (…) Morgan Stanley estimates AI embracement could deliver around $920 billion in benefits over time, generating $13-16 trillion in long-term value for companies in the S&P 500, the strategists who put this together claiming that the opportunity amounts to more than 25% of the S&P 500’s estimated 2026 adjusted pretax income.
Third, the US energy bill as a percentage of N-GDP is historically low at 2.2%, which tends to lower input cost, raise the personal savings rate and/or boost other expenditures. (…) There are also growing expectations that the world will soon be oversupplied significantly, because major OPEC oil producers are members emphasising the importance of fighting for market share at a time when a glut already exists (…). Indeed, Saudi Arabia is taking measures to free up more oil for export, having vast amounts of spare capacity, even as growth in oil demand is on a structural downward trajectory, thanks to shifts in China’s economy and a growing adoption of electric vehicles around the world.
In this connection, the best that could be expected from geopolitical tensions following Israel’s attack on Hamas’ leadership in Qatar and prospects of tighter Western sanctions on Russian energy exports, is a cap on oil prices around $63.00 a barrel, the point of this being that most recessions have been associated with high energy bills, which is not the current prospect at all. (…)
Fourth, the Fed is ready to trigger an easier monetary stance, waiting for an appropriate time, excuse and opportunity to raise the money supply with a series of rate rate cuts, which should keep the level of business activity at an affordable level.
Fifth, the deregulation bias of the current administration supports M&A activity and favours the extension of bank and market credit, while lightening up on the enforcement of regulations, lowering taxes, and encouraging foreign investments.
Sixth, stock buybacks, which are expected to reach $1.2 trillion in 2025 and probably as much in 2026, are suppressing volatility and keeping stock prices at record highs. According to Nomura’s Charlie McElligott, these are not only creating constant demand, but also tend to be more active into pullbacks and passive into rallies.
Lastly, in the year to date the DXY- an index of the value of the US dollar relative to a basket of foreign trade partners’ currencies – is down 10% with little hope that a reversal is in the making. This should bring about higher foreign operating earnings and larger translation of foreign assets.
Investors should also take into account that Wall Street is even more bullish than me. Goldman’s David Kostin expects the US benchmark to flirt with 7000 by mid-2026; SocGen’s Manish Kabra sees upside to 7300 by May 2016; Evercore’s Julian Emanuel thinks it could reach 7750 by the end of next year; and Yardeni has a range of 6900-7130 for 2026. Perhaps stocks are poised for a lifetime bull run.
BCA Research is more cautious:
Unfortunately, looking ahead, Berezin reckons that both major supports of the current bull market — a resilient economy and AI euphoria — will wane.
The deteriorating labor market shows the U.S. economy is vulnerable. “The seemingly modest rise in the unemployment rate has been masked by the fact that many people have given up looking for work. If one were to include persons who are not in the labor market but who still currently want a job, the unemployment rate would have increased by 0.76 percentage points since January,” says Berezin.
The housing market is looking “increasingly shaky,” he adds, while outside the U.S., growth has been propped up by demand from tariff front-running.
In addition, though AI is a secular growth story, it will not be immune to a cyclical downturn. BCA notes that companies such as Meta Platforms META and Alphabet GOOGL are highly exposed to advertising, and if consumer spending slows then ad revenue will fall sharply.
“Past experience suggests that investors often get jittery when free cash flows begin to decline,” says BCA. “This happened in late 2021 after the combined free cash flows of the hyperscalers – Amazon AMZN, Google, Meta, Microsoft MSFT, and Oracle – temporarily rolled over.”
BCA thinks hyperscaler capex will not peak until the second half of 2026, but the stock prices of AI-linked companies, which currently represent a third of the S&P 500 market capitalization, “could swoon well before then.” Consumer spending would then fall further as equity wealth declined.
“While it is impossible to know exactly when global equities will peak, there are now enough vulnerabilities to justify keeping one’s finger near the eject button,” Berezin summarizes.
The consensus currently expects hyperscaler capex growth to peak now. As Goldman Sachs points out “Timing this inflection is challenging, as consensus has consistently underestimated hyperscaler capex. Estimates for hyperscaler 2025 capex have increased by $100 billion YTD to $368 billion currently. 3Q and 4Q earnings seasons will represent key tests.”
Goldman tracks stocks involved in each of the 4 phases of the AI build out:
Phase 1 is NVDA, 2 is Infrastructure companies (semiconductors, electrical equipment companies, technology hardware firms, power suppliers), 3 is Enabled revenues and 4 is Productivity beneficiaries.
Source: FactSet, Goldman Sachs Global Investment Research
Returns of AI-exposed equities have continued to be concentrated in firms with tangible near-term earnings benefits. Our AI infrastructure basket (Phase 2) has outperformed the equal-weight S&P 500 by 8 pp YTD, compared with -0.3 pp for our AI-enabled revenues basket (Phase 3) and -5 pp for our AI productivity basket (Phase 4). While we expect the AI trade will eventually broaden to Phases 3 and 4, we believe investors will require evidence of a meaningful near-term impact on earnings.
(…) equity share prices of the hyperscalers ((AMZN, GOOGL, META, MSFT, ORCL) have only modestly outpaced the trajectory of earnings since 2023 and consensus estimates already embed a slowdown in EPS growth from 18% in 2025 to 11% in 2026. In addition, a reduction in capex would alleviate concerns about whether these companies can generate sufficient returns on their investments, and at the same time would increase free cash flow generation and allow for further cash return to shareholders via buybacks and dividends.
The inevitable slowdown in capex growth poses a risk to the valuation of AI infrastructure stocks. Prices of Phase 2 stocks have far exceeded the trajectory of near-term earnings, reflecting optimism about the forward outlook. Within Phase 2, power stocks have become one of the most popular expressions of the AI trade.
Alongside the acceleration in hyperscaler capex, investors have priced the associated power deals and earnings uplift. As a result, the P/E of the median stock in the Power Up America basket has increased from 17x at the start of 2024 to 22x today.
Consensus estimates imply that the hyperscalers will maintain their pace of capex growth in 3Q 2025, likely extending the performance of AI Phase 2 stocks However, analysts currently assume a sharp deceleration in 4Q 2025 and 2026. While many power announcements come with a lag and could lead to positive earnings revisions, the second derivative of capex growth turning negative poses an eventual risk to valuations. (…)
An extreme scenario in which the hyperscalers cut capex back to 2022 levels would pose substantial downside risk to both the AI trade and the broad S&P 500. Hyperscaler capex totaled $158 billion in 2022, $275 billion below the expected level of capex among the group in 2026. If capex were to immediately revert back to 2022 levels, that “lost capex” would represent a reduction of roughly 30% to the consensus estimate of $1 trillion in 2026 S&P 500 sales growth.
As a result, revenue growth would decline from the current consensus estimate of 6% to roughly 4%. While the decline in near-term revenues would be modest, this extreme reduction in capex would likely be accompanied by a deterioration in the outlook for long-term AI-driven earnings growth, weighing on valuations as well. In our macro valuation model, a reversion of long-term growth estimates back to early 2023 levels would imply 15-20% downside to the current valuation multiple of the S&P 500.
In Phases 1 and 2, there were mainly winners from AI. During the coming Phase 3, investors will have to navigate among winners and losers from AI applications as GS explains:
Our discussions with investors and recent equity market performance reveal limited appetite for Phase 3 companies with AI-enabled revenues. (…)
We expect that AI will create wide return dispersion within Software. (…) investor questions around Software center on whether AI represents an opportunity for eventual sales growth or an existential threat.
As our equity analysts recently highlighted, some investors are concerned that AI will disrupt pricing models, lower the barrier to entry for new competitors, and compress the profit pool available to leading SaaS incumbents. This dynamic could lead to outright price declines for some firms, underscoring the importance of bottom-up stock-picking. The recent earnings season demonstrates the wide dispersion of returns, with an 80 pp differential between the best-performing and worst-performing stock in our Phase 3 basket after reporting results.
We expect the AI trade will eventually broaden out to some Phase 3 companies, but investors will likely require evidence of a tangible near-term impact on earnings.
(…) for AI-native companies to take share from SaaS companies, the AI product has to be meaningfully better and meaningfully cheaper than the incumbent, and SaaS companies continue to progress with their own AI-enabled products. In addition, DIY Software has its own costs, and there are still advantages to outsourcing critical pieces of the technology stack.
Our equity analysts acknowledge that there has been limited value creation in enterprise software applications thus far. However, Palantir is a notable exception and demonstrates the importance of a tangible near-term earnings story for outperformance.
On productivity:
From a productivity perspective, the US economy remains in the early innings of AI adoption. Surveys show that many companies are experimenting with AI. A 2024 McKinsey survey showed that nearly 80% of companies had used AI in at least one business function. But based on the Census Bureau survey, only 9% of US companies currently use AI to produce goods and services (vs. 7% in 1Q). AI adoption is higher among large firms (15% for firms with 250+ employees) and in select industries (e.g., Information, Finance).
Company commentary shows that many S&P 500 firms are starting to describe specific AI use cases in their businesses. In the 2Q earnings season, 58% of S&P 500 companies mentioned AI on their earnings calls. Of the companies mentioning AI, roughly 60% discussed using AI in the context of productivity and efficiency. 24% of firms mentioned specific use cases related to call centers and customer support, 24% related to coding and engineering, and 23% related to marketing.
However, many companies provided company- or industry-specific use cases as well. For instance, BAC highlighted the use of AI agents to help reconcile trades and IQV discussed the use of AI agents to reduce delivery times.
However, the share of companies quantifying the impact of AI on earnings today remains limited. Few companies that mentioned AI during 2Q linked the technology to near-term margins or earnings. The low share mirrors survey-based measures, such as the McKinsey survey that showed more than 80% of companies stated generative AI was having no tangible impact on EBIT today.
Some companies provided quantitative metrics related to AI use cases, however. Examples include CHRW (“One recently announced example is a new AI agent that helps shippers automate the process of classifying LTL freight under a new national system. In the AI agent’s first few months, it has been determining the freight class and code for about 2,000 orders a day, and it has reduced the processing time from 10 minutes or more per shipment to 10 seconds or less.”) and CTSH (“A few data points of our progress include: first, we currently have more than 2,500 early GenAI client engagements, compared to 1,400 in quarter one; nearly 30% of our code was AI-generated in quarter two, significantly improving the productivity of our developers.”).
Some investors have also questioned the impact of AI on employment. Our economists showed that aggregate labor market impacts from generative AI remain limited, but there are some signs of labor demand hits in AI-exposed industries (e.g., marketing consulting, call centers, graphic design, web search and software development).
During 2Q earnings season, only a handful of companies explicitly linked use of AI and reduced headcount. For instance, NOW stated “So, really thrilled with the productivity efficiency gains that we’re seeing from AI. We talked about at Knowledge, $100 million in savings and head count alone in 2025, and we’re seeing that come to fruition as planned.” and ADP highlighted “…client unit counts continue to grow at a very healthy and consistent clip with what they have been growing in previous years, yet we are able to actually see some operational head count reduction in those businesses as a result of generative AI and some of the other like-minded tools and things that we’re deploying in those spaces.” (…)
From an industry perspective, Software & Services, Commercial & Professional Services, and Banks have the highest combination of labor-intensive businesses with high exposure to AI.
In my estimation, the consensus is underestimating the fast developing inference demand (queries) as well as the coming rapid development of AI uses in robotics, transportation (self driving) and healthcare, to name a few.
On the other hand, technology is also at play within AI capex. For example, Nvidia keeps developing ever more powerful systems within a smaller footprint.
The NVDA chip most associated with reducing the required size of data centers is the 2025 Blackwell platform which delivers far higher performance per watt and per rack, allowing the same workloads to be run with fewer servers and less rack space. NVDA has stated that Blackwell delivers up to 30x higher LLM throughput with about 25x better energy efficiency vs prior generations, meaning fewer GPUs and servers are needed to complete the same training/inference jobs.
NVDA’s Vera Rubin platform targets even higher per-rack consolidation starting in 2026. Rubin CPX specifically targets million-token and video inference with claims of 7.5x more AI performance than GB300 NVL72 (Blackwell) in a single rack configuration.
The same seems true for energy. While absolute energy demand and bills will rise with AI scale, energy per unit of compute can fall with liquid-cooled Blackwell racks and workload-specific silicon like Rubin CPX, improving cost per token or per training step.
The main risk is whether the 165% currently estimated increase in data center electricity demand by 2030 can be met, physically by the grid, and at reasonable costs.
A May 2025 MIT analysis is more specific:
Given the direction AI is headed—more personalized, able to reason and solve complex problems on our behalf, and everywhere we look—it’s likely that our AI footprint today is the smallest it will ever be. According to new projections published by Lawrence Berkeley National Laboratory in December, by 2028 more than half of the electricity going to data centers will be used for AI. (…)
By 2028, the researchers estimate, the power going to AI-specific purposes will rise to between 165 and 326 terawatt-hours per year. That’s more than all electricity currently used by US data centers for all purposes; it’s enough to power 22% of US households each year. (…)
Between 2024 and 2028, the share of US electricity going to data centers may triple, from its current 4.4% to 12%. (…)
The Lawrence Berkeley researchers offered a blunt critique of where things stand, saying that the information disclosed by tech companies, data center operators, utility companies, and hardware manufacturers is simply not enough to make reasonable projections about the unprecedented energy demands of this future or estimate the emissions it will create. (…)
Individuals may end up footing some of the bill for this AI revolution, according to new research published in March. The researchers, from Harvard’s Electricity Law Initiative, analyzed agreements between utility companies and tech giants like Meta that govern how much those companies will pay for power in massive new data centers. They found that discounts utility companies give to Big Tech can raise the electricity rates paid by consumers.
In some cases, if certain data centers fail to attract the promised AI business or need less power than expected, ratepayers could still be on the hook for subsidizing them. A 2024 report from the Virginia legislature estimated that average residential ratepayers in the state could pay an additional $37.50 every month in data center energy costs. (…)
It all comes down to energy cost and supply. This is from the Center for Strategic & International Studies last March:
This paper focuses on electricity on the basic premise that electricity supply is the most acutely binding constraint on expanded U.S. computational capacity and, therefore, U.S. AI dominance. (…)
This story extends beyond AI. Electricity demand is also growing from other electricity-intensive industries like semiconductor fabrication and battery manufacturing. The broad political consensus to reindustrialize the U.S. economy will drive growth in energy-intensive industries like mining, minerals processing, metallurgy, and beyond. A deep technological trend toward electrification means industry, along with the transport and heating sectors, is growing more electricity-intensive each year. (…)
The total effective capacity of the U.S. generation base has stagnated since 2010, and it may have even declined. Coal-fired generation with high ELCC ratings (84 percent) has been replaced by low ELCC resources like onshore wind (34 percent) and solar (13 percent). Even dispatchable gas-fired generation (78 percent) has a lower rating than coal and nuclear (95 percent) because of fuel supply and gas-electric coordination issues during winter storms. (…)
On a national level, there is effectively no “spare capacity.” (…) Today, every new gigawatt of data center demand must be met with matching new gigawatts of effective capacity sited within the borders of the same reliability planning region. The past failure to grow effective capacity explains why a focus on speed-to-power necessarily follows from the data center boom and AI technology race. (…)
Utilities and independent power producers (IPPs) are turning to gas generation to serve new demand because there is no other technology that brings as much effective capacity online, in as fast a timeline, with as much siting flexibility, under such a manageable financial profile. (…)
With natural gas production booming and prices at or near all-time lows, access to fuel volumes at reasonable prices is a nonissue.
The gas generation boom is creating upstream supply chain constraints. Orders for new gas turbines are rapidly piling up at major manufacturers like GE, Mitsubishi, and Siemens, with these firms reporting order books with delivery now stretching out past 2028. Though construction of a new gas plant can take as little as a year, with these backlogs, a project placing an equipment order today is unlikely to come online until 2030 or beyond. This order backlog inevitably includes a huge number of U.S. projects at later stages of planning and development, so gas deployment will continue the coming years, but scaling growth rates will be a challenge. (…)
Meanwhile, solar paired with storage, directly on-site or in portfolio, has been shown to greatly improve the project value to electricity buyers, which has made such projects more attractive to developers and financiers relative to stand-alone wind development.
(…) even in a best-case construction scenario, a new AP1000 [nuclear] project will take six years or more, resulting in the earliest possible contribution to the resource mix starting in the early 2030s. A steady scaling of nuclear supply chains, workforce, and technology maturation is crucial for nuclear to play a role in smoothing coal (and existing nuclear) retirements in the 2030s and beyond. Nuclear will play a limited role in the near-term speed-to-power era but could deliver enormous economic and strategic value to the nation over the medium and long term. (…)
Investment decisions during the next several years will determine whether the U.S. grid in the 2030s and beyond allows for unconstrained electricity demand growth, at globally competitive prices, with a world-leading reliability profile—or if dramatic load growth leads to instability and internal conflict over a scarce resource. (…)
Lastly, a key area for attention is minimizing cost inflation for existing ratepayers. Electricity prices are rising rapidly, recently outpacing inflation. State legislatures and public utility commissions (PUCs) are facing a wave of utility investment requirements that translate into increasing rates. Wherever possible, federal policy should enable and encourage policy that lowers costs for generation and grid investment and reduces ratepayer exposure to investment directly tied to data centers. In cases where projects deliver clear national strategic value in the AI race, federal funding should buy down project costs to reduce ratepayer cost inflation.
Electricity prices in the US are up 24% since 2022, against total CPI and wages up 14%. States will soon face increasing pressures to control consumer costs. Who will foot the bill?
My September 12 post included a section “IT’S THE ENERGY, STUPID!” discussing China’s growing lead in energy storage and how Trump’s policies against renewable energy will aggravate the US declining competitiveness in energy costs. In 2024, solar and wind energy costs were between 25% and 60% cheaper in China vs in the US. Increasingly able to store cheaper renewable energy, China will access more lower cost electricity than most other countries.
A lack of accessible, affordable, and sustainable energy can become a major bottleneck, potentially limiting the scope and speed of AI adoption and development.
In essence, energy is the fuel that enables AI, and the rapidly growing demand for it presents both a significant challenge and an opportunity to innovate and manage energy systems more efficiently.
China has a comprehensive plan, unlike the US.
Speaking of future AI uses, Constantin, a long time and generous reader sent me this amazing video on clone robotics: https://www.youtube.com/watch?v=E1theCfcFsA
Charts FYI (courtesy of Isabelnet)
