Nvidia earnings, July PCE and more ?3

08:56 27-Aug-26

Nvidia delivered another beat and raise in its earnings last night, which garnered a mildly positive initial market response that gradually gathered steam as investors digested a growth outlook that would put the company ahead of Google and Apple in sales.  That may have also been helped by reports that that the company will buy Hugging Face for $12.9b which is an eye watering 86 P/S multiple for a company that is somewhere around breakeven profitability.  Pre-market trading is already fading the gains, which topped out just over $225, below prior highs.  The takeaway is that the AI build out push continues, but it faces an ever more cautious market.

For the record, yesterday's 5-year treasury auction had similar activity that continues to confirm a preference for shorter term American debt.  I will only update on today's 7 year auction if it changes that trend, but I will certainly revisit the longer term auctions in mid September, just after Bessent's bogus buybacks, covered by CrowdWisers with thoughts on the longer term outlook below.  

PCE numbers for July make those for June look like an aberration in terms of inflation, again, as predicted.  However, they do indicated further consumer weakness.  The headline and Spending were both a tenth higher than most estimates at +0.2%, whereas Income was two tenths higher than forecast at +0.4%.  Core was in line at +0.2%.  Again, none of this is consistent with the Fed's 2.0% inflation target, but we'll see what Warsh has to say on the matter at Jackson Hole on Friday morning.  I think the mildly negative market reaction in the meantime has more to do with consumer aspect.  If the Fed chair sticks to jawboning increased productivity from AI over the long term we'll likely see no real change in market trends.  However, if he acknowledges increasing concern about AI's impact on the job market, stocks may actually accelerate again as expectations for a hike, which have rebounded above 40%, are dashed once again.  

In my view, the main argument for a hike remains decreased foreign ownership of treasury debt, though you won't hear that from any administration official.  Those types are probably satisfied with yesterday's two-year auction results, even though both the bid to cover and pricing remain poor by historical standards.  I may update with a comparison to today's 5 and 7 year results when available.

PCE figures on Wednesday morning makes that day look like the macro crux for this week, though both have reduced importance compared to pre-Warsh and AI exuberant market conditions.  Bessent's bogus treasury move means I'll also return to giving currency movements more attention as well as the auctions.  Although I may comment on the short term sales this week, I will be much more interested in the T-note and bond results after September 9th and 10th.

In that context, the most remarkable thing about yesterday's Fed minutes was the lack of market reaction, which I think shows just how little credibility the committee has left.  Though a fair bit has been written on it only holding six meetings next year, the only reason that is surprising is that the Fed had already published a schedule.  More novel was how many words were devoted to AI and hyperscaler build out and its effect on economic and market conditions, except unlike that analysis, the "some participants" think it will take poor megacap earnings or major downward revisions to the outlook to produce an effect that would be of concern.  This new found focus reflects earlier Warsh statements that "ALL investment will eventually be AI investment" but it shows how much influence the Chair has over the Fed's theater, and I may take to calling him Max Headroom in light of the complete disconnect with reality.

Topics:

• none