Volatile energy prices are whipping around headline inflation statistics. July’s CPI reading showed a 0.1% increase, following a rather drastic 0.4% decline in June.

On a 12-month basis, inflation cooled slightly to 3.4% from 3.5%. Core inflation, which strips out food and energy, is estimated to have eased from 2.5% in June to 2.4% in July. There is a tendency for higher energy prices to work their way into the core reading with a lag, which could create some upward pressure going forward.

The Federal Reserve’s new chairperson, Kevin Warsh, gave his second Open Market Committee press conference on July 29, announcing the Committee’s decision to leave rates unchanged at the target range of 3.50%-3.75%. Three of the twelve voters dissented in favor of a rate increase. The target rate has remained the same all year, with expectations for rate cuts giving way to expectations for increases. Those expectations are bending back toward unchanged. Like a World Cup match, the result is a lot of back-and-forth drama without much scoring.

Warsh’s press conference caused grumbling among the media. He deflected questions designed to tease out his thoughts about the likely timing and direction of the Fed’s next action. Warsh stated that he would prefer for the Federal Reserve to communicate less while listening to financial markets more. His point about less communication was a subtle rebuke of colleagues using their positions for clout or to stump for their own policy preferences. It’s harder to infer exactly what he meant by listening more to markets.

What financial markets seem to be saying in the wake of the FOMC announcement is that this incarnation of the Fed looks more accommodative than anticipated. The yields on 10- and 20-Year Treasuries rose 0.21% and 0.14% through mid-August. Gold and silver increased about 10% in two weeks before easing somewhat. Rising inflation expectations could explain all of those moves.

Warsh has appointed five task forces to look at the Fed’s balance sheet, data collection, productivity estimates, communication strategy, and its treatment of inflation. Inflation has been above the Federal Reserve’s ostensible 2% target for five years, and perhaps cynically, it seems likely that Warsh’s Fed will devise a way to define inflation down. When the government changes its inflation grading rubric, the outcome is always to reduce reported CPI. The last major methodological adjustments came in 1995 as a result of the Boskin Report under Alan Greenspan. The Boskin Report estimated that inflation was overstated by approximately 1.2% per year. The methodological changes resulting from the report reduced reported inflation statistics by an estimated 0.4%-0.6% per year. Thirty years later the Fed might do something similar.

That’s not to say that consumer prices actually rose any slower following the Boskin report. The prices of first-class postage stamps and McDonald’s Big Macs have risen about 3.2% per year since 1995. Meanwhile, reported CPI inflation has averaged 2.5% during that period. Most “stuff” that you could buy in 1995 is about 50% more expensive today than the CPI implies, more in line with the CPI methodology before the Boskin adjustments. However, a few things have become drastically cheaper and better, especially consumer electronics, and the modern methodology treats the march of progress as deflationary above and beyond the extent to which productivity improvements tamp down supply chain costs. What will the Fed think of next to explain away the rising cost of goods and services?

The S&P 500 and the Dow 30 achieved new highs in the wake of second quarter earnings, with the NASDAQ still modestly below the highs achieved in early June. Second quarter earnings increased an astonishing 50% on 15% combined revenue growth through early August according to FactSet’s Earnings Insight. Investment gains at Amazon, related to its position in Anthropic, and at Alphabet, related to equity holdings including Anthropic and SpaceX, drove a meaningful portion of the earnings increase. Still, excluding these tailwinds, earnings in the second quarter grew more than 30%.

AI infrastructure providers and their hyperscaler customers caused the bulk of the historically anomalous earnings rise, while the energy sector enjoyed the fruits of surging commodity prices. Tariff refunds provided a further boost. Profitability was strong throughout most of the market, with only the healthcare and real estate sectors reporting profit margins below recent averages. The S&P’s P/E multiple of 20 is only slightly above the recent average and could be justified if this level of corporate earnings is sustainable.

The big question is how much of corporate America’s surging profit growth is at risk if AI capital investments aren’t soon justified by paying customers. The revenues that Nvidia, Broadcom, Micron Technology, and other AI vendors are earning at historic profit margins are not being counted as expenses by their customers. They are investments which will be recognized as depreciation slowly over subsequent years. Meanwhile, tax incentives permit companies to depreciate capital investments at an accelerated rate and save the resulting income tax up front. The system appears more profitable due to timing mismatches between revenues recognized now and costs recognized later. The hyperscaler’s customers, corporate America, need to ante up or else mounting depreciation will create an earnings drag in the future.

There is no reason to think the AI investment cycle will suddenly stop, but it could certainly slow. The largest spenders are devoting all of their operating cash flow already and are issuing stock and debt to push the envelope beyond what their operations can sustain organically. Meta tacitly admitted that its investments had overreached its actual compute requirement, and it announced an initiative to resell the excess compute. Investors boosted Meta’s share price, seemingly rewarding the company for its efforts to produce more current period revenue from its recent investments.

Nvidia made headlines by partnering with six institutional money managers to raise $500 billion of investor capital for companies to spend on Nvidia chips and other AI infrastructure. The goal is to find investors willing to treat rapidly depreciating computing hardware as loan collateral. You can finance an iPhone, so why not a rack of servers? The skeptic’s view is that AI equipment prices cannot be justified by end-market demand for AI applications, with prices being propped up by bidding wars between dozens or even hundreds of companies who all dream of becoming one of a handful of big winners. As competitors start to drop out, the price of AI compute should recede to its marginal cost of production. Pity the lenders in that scenario.

Accommodative fiscal and monetary policy and torrid reported earnings growth pave the way for steady stock market gains. Investors may wonder at what level all the good news is priced in, and we agree that certain market sectors may be priced beyond perfection. However, valuations do not look unreasonable where the underlying earnings can be sustained. As always, we preach earnings growth and diversification. Don’t let your portfolio become too reliant on the continued strength of the stocks making all the news.

Looking for high quality stock picks?

In this issue, our analysts recommend a large-cap leader in the utilities sector operating within the independent power producers industry. As a large private-sector power producer and U.S. nuclear energy company, this clean-energy pioneer is positioned to benefit from surging electricity demand driven by artificial intelligence infrastructure. Over the next five years, it is projected to deliver a robust 19.4% compound annual total return, supplemented by a 0.6% dividend yield.

Our second selection is a large-cap player in the technology sector specializing in the electronic components and connectors industry. As a global leader in interconnect systems, this company maintains critical design partnerships for next-generation GPU architectures and is poised for outstanding growth as data centers continue their expansion. Over the next five years, it is estimated to achieve an impressive projected compound annual total return of 22.1% to 23.4%, alongside a 0.6% dividend yield.

Subscribe today to unlock these and other top picks!

The commentary is excerpted from the issue of the Investor Advisory Service newsletter published at the end of August 2026. To receive commentary like this in a timely matter and receive actionable stock ideas each and every month, subscribe today. The Investor Advisory Service stock newsletter was named to the Hulbert Investment Newsletter Honor Roll for the 16th consecutive year for outperforming every up and down market cycle since 2007.

For more information about the Investor Advisory Service, to download a sample issue, or to subscribe to the best investing newsletter in the U.S. for long-term consistent returns, visit Investor Advisory Service.