Speech models ace the lab. Canto's built for your commute.
Speech recognition is great on clean studio audio. Real dictation happens at your desk, on a walk, or in a loud cafe. That's why Wispr Flow built Canto, its own speech model, trained on how people really talk.
On 10 hours of real dictations from 2,300+ speakers, it had the lowest word error rate of every model tested, ahead of models from Google, OpenAI, AssemblyAI and Deepgram. Background music, traffic, someone talking next to you, a whisper: on a challenge set full of exactly that, Canto had the lowest error rate among real-time models tested.
Talk naturally and Wispr Flow types clean, formatted text in any app. 4x faster than typing.
PEAK CAPITAL CYCLE

The chipmaking industry usually goes through two cycles:
An R&D intensive cycle with low ROIC and margins, all to create the next industry-leading product
A sales cycle where competition roars along with margins as the fruits of that R&D investment hit the tape
All through 2023 and late 2025, AI created one of the most intense R&D cycles in chipmaking history.
Which, by nature, led to the most competitive environment as well given the record margins and ROIC levels printed by these participants.
It is undeniable that the industry is now headed into the sales cycle, and whoever tells you chips aren’t cyclical anymore should be fired.
Korea, Taiwan, and Vietnam have flooded the market with new tech exports (mostly chips and manufacturing equipment) in the tune of ~35% more than last year.
China? It added another ~22%.
Why this matters to you should be pretty simple…
If you own any exposure to the AI trade, specifically any of these chipmakers, then you may want to consider taking some risk off the table.
When cyclicals stocks see their forward P/E multiples fall (as they are for chips), it usually means loss of confidence in earnings growth and quality.
Now you know why the market is betting in that direction.
And then there’s the cherry on top,
~27% of all data center projects, and 65% of all expected computing power, are going to be wiped out of the equation after the recent midterm election results.
Which of course could hit NVIDIA’s balance sheet the most now that Jensen Huang has left the company altogether…
Now let’s get an update on what the market likes and dislikes:

A sudden shift took place this week, being led by Value and Momentum.
Let’s recap some of the market structure behind this pivoting action back into the AI momentum/growth complex.
Two weeks ago, the AI trade sold off in a big way, kickstarting what I call a three-phase structure to be fulfilled in the coming weeks and months.
It goes like this:
Phase I: Initial correction/bear market selloff, triggering margin calls and some forced liquidations as the market’s leverage and concentration creates more of these tail risks
Phase II: Brokers who are forced to buy these liquidated assets create the false belief of a rebound, where more and more participants come and chase what they think is an organic move
Phase III: These brokers, like Citadel after buying Leopold Aschenbrenner’s fund in July, will take advantage of the short-term rebound and dump the assets they didn’t want in the first place
Then the cycle begins again.
Margin calls, forced liquidations, and a second test of whether there is any bailout interest or available assets to meet these margin calls.
You can backtest this all you want…
Two to three times is the usual amount of times these 10-20% drawdowns from highs get tested before the market makes a decision.
So no, I am not buying the pivot in momentum, especially as value was the second leader this week.
It’s good to see our recent buy didn’t fall into the AI trap rebound too, confirming its previous AI premium has been mostly erased after a 50% selloff from the highs.
CHART OF THE DAY
Some people will take a chart like the one in this section and consider its face value.
The belief that S&P 500 earnings, and its P/E, are pointing to a bullish run for the market and the economy isn’t as clear as you might think.
Excuse me for sounding bearish here…
But I just cannot justify joining the party when the S&P offers me a -2% to +2% earnings yield and bond yields being close to 5%.
The reason the P/E should be 40-50% higher, and corporate earnings 40-50% lower is simple.
ARE YOU COVERED? —>

Roughly 10 companies make up ~40% of the entire S&P right now, and 60% of all stocks have something to do with the AI trade in one way or another.
And,
Most of the AI economics are based on future revenues, non-cash items, and the infamous circular financing scheme.
So on paper things are great, in actual cash coming in, not so much.
IMPORTANT GAUGES

In this relative risk model, I aimed to measure the breakout points of volatility in the S&P 500.
By taking the daily market volatility vs the daily VIX volatility, you can land at the spread above, and see when the good times to buy/sell are.
Right now, the S&P daily volatility has been above the implied daily VIX volatility for over a week now.
Historically, these crossovers tend to last roughly 20-30 days to represent volatility clusters in the market amid some unforeseen event or repricing.
The crossover these limits have been erased sooner than expected, as the past two weeks’ breakout should have triggered that 20–30-day VIX seasonality.
Because it didn’t, and it all reversed during a week where momentum came back into the AI trade on that Phase II move…
I’m willing to bet this is nothing less and nothing more than another “dispersion” trade episode, as traders begin to hedge index volatility by buying individual AI stock volatility ahead of their earnings announcements.
From what we saw in July and October, it won’t end up much different than a 10-20% drawdown.
Now let’s cover some items for this coming week:
New Earnings:

The mother of all earnings are going to hit the tape this week…
Tuesday: NVIDIA and Home Depot, the ultimate tug-of-war results to give us a gauge into the differentiated AI economy and real consumer economy.
You already know I will be covering NVIDIA in as much depth as possible this week, so make sure you grab a seat (if not already) in Premium to see what I plan to do about this company.
Meanwhile, here’s my take on where NVDA stock may head next.
Thursday: Among the big China tech names like BIDU on Monday and PDD on Thursday, I am mostly interested in Alibaba.
Because it is already a part of our portfolio, and now trading at a significantly attractive valuation, I will be going back to the well and cover their earnings for you alongside a renewed and updated thesis take for you to action.
Retail Sales - Tuesday:

An update on how the consumer is choosing to spend their money is coming.
Over the past quarter, the trend has definitely been headed toward defensive spending rather than more discretionary items.
Can’t buy a home due to rising mortgages and home prices? Invest in renovating your current place or your new rental - Rising furniture sales.
Travel and other such experiences are getting too out of budget? Go out to eat instead as you wait for better times to invest in your leisure time - Rising food services sales.
All of which will probably have an impact on the rest of the earnings season, especially for restaurant stocks like CAVA and Chipotle.
A Final Note
COMING UP NEXT (Reminder for This Week)
The recent NVIDIA leadership debacles and changes have created a storm of uncertainty for current (and prospective) shareholders, if that’s you, next week will bring all the answers you need.
As Chinese data continues to deliver on our original thesis, a proper update on what makes that the most attractive emerging market for me is due to hit your Premium inbox.
More company deep dives are being done from the PMI picks this month, and I’m looking forward to bringing you another double-digit upside set of opportunities.
Meanwhile, here is the latest from Goldman Sachs covering the recent CPI print and how that may affect the Fed’s thinking going into December’s rate decision.
I would take this with a grain of salt, and if anything use it as a gauge as to where these institutions see the potential rate paths heading to:
Until next time,
OFFSIDE RESEARCH
Against the Tape, Ahead of the Curve.

