Weekly Market Outlook 2026/08/09
WARsh and Peace
Hello traders,
Hope you all had a great weekend and enjoyed the sunny weather on offer stateside.
After all the volatility on offer through July, the S&P ended up smashing all time highs last week. This is something that would make Trump administration elated on the cusp of midterm elections.
Realtime Discord access with equity positions tracker(new website rolling out to select subscribers), intraday orderflow & option dealer summaries included with the paid newsletter membership . Click here to subscribe.
Nasdaq on the other hand, is a completely different story.
After having a 3500pt pullback from all time highs right after the fed minutes, Nasdaq came roaring back and is now about 1000pt off the all time highs.
This comes in the aftermath of Leopold Aschenbrenner’s fund almost blowing up, Citadel buying the said book causing the bottom on the AI trade and then coming back again with a fresh $400m equities bet.
TECH IS BACK? KINDA
Speaking of equities, earlier this year we had posited the idea of being long TEAM 0.00%↑ (Atlassian)
The core pillar of our thesis has worked exceptionally well and TEAM 0.00%↑ is now up about a 11000bps from our post.
It is wise to offload some of the stock in this area to “give thanks” to the market by taking some chips off the table.
Again, there are multiple ways to trade the market and one of them is to keep doubling and tripling down on your winning positions and either become a Billionaire or be someone who gets margin called on the day of their wedding. Neither approach is wrong, I just prefer one where my resting heart rate does not collect social security cheques.
ART OF THE DEAL
Last week the US again relented and decided to not attack Iran as the 10Y crept up into the “Peace Deal Sounds Amazing Actually” zone and that posturing will continue until we get back into the “We Will Attack Iran” zone.
The administration is giving Warsh a lifeline with this strategy and now the only thing that can ruin this party is a hotter than anticipated CPI print this week.
Zooming out a little here, as AI keeps advancing we will see some sort of a disinflationary impact on the world around us (oh no Warsh has taken over this post), eventually pushing us to new highs on Nasdaq.
However, there are people who do not believe in the disinflationary nature of AI and they will keep pushing their chips into precious metals like Gold and Silver.
If I had to construct a lazy portfolio for my aunt who doesn’t follow the markets it would be a split between local inference, precious metals (overweight Gold) , emerging markets, drones & robotics and then check back in after a decade or so.
THE WEEK AHEAD
This week we have the CPI, PPI and retail sales, the importance of these cannot be understated lets look at each of these in context of the broader market.
CPI expectations are at 3.4% YoY and for core it is 2.5%. For MoM core you would want it under 0.2% if you are bullish equities(especially tech). A surprise core print above 0.4% will push the Fed into an uncomfortable corner with regards to rates and cause a landslide in high beta tech.
We got a decently soft Jobs report on Friday, if it is combined with a cooler CPI and PPI we would be in goldilocks territory.
However, if the script is flipped we would be at
soft jobs Friday → hot CPI Wednesday → hot PPI Thursday.
This looks fairly like the S-word (Stagflation).
Retail sales seem to be doing fine for the most part. However as equity bulls you want it to be normalized against inflation as retail sales are nominal and not inflation adjusted. So a 0.2% print with a cooler inflation is likely a better sign of booming economy+consumption than a 0.4% print with a hotter CPI print.
TOO CHEAP TO METER
Some interesting ARC-AGI results from this past week on both ends of the spectrum
Prime Agent, A self-improving RLM harness for coding and long-running autonomous tasks. Designed to be both token-efficient and expressive through programmatic tool calling, context as a variable, multi-agent messaging, and a self-modifiable harness state.
Prime Agent is a general-purpose coding harness
On ARC-AGI-3, it scores 95.5%, surpassing the human-expert baseline, but the gain is not benchmark-specific.
We see major improvements across models when compared to their proprietary harnesses:
Back of the napkin math says about $250-290 in output token cost, with total cost not above $400 or so.
That is impressive but still expensive.
Okay, here is DeepSeek V4 Flash.
“DeepSeek V4 Flash from
on ARC-AGI (Verified):
- ARC-AGI-2: 61.4%, $0.04/task
- ARC-AGI-1: 89.0%, $0.02/task
DeepSeek V4 Flash sets the new standard on the cost-to-performance Pareto frontier.”
I don’t want to get all too technical here but a layman explanation is that you can get near-frontier level intelligence for roughly 96 to 98 percent cheaper than Opus.
This makes us all question, is AGI here? and is it really that cheap?
