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.

NVIDIA WITHOUT JENSEN HUANG

NVDA Corporate & Technical Tenure Stack, Offside Capital

While everyone else is focused on the Jensen Huang affair scandal, buying the dip on impulse ahead of the company’s most important earnings announcement…

I’m looking at the part of the story that could actually determine what this stock is worth next.

Because Jensen leaving NVIDIA is a problem.

But,

It’s not the biggest one.

The company still has a handful of highly qualified engineers, product developers, and technicians to keep developing industry-leading chips moving forward.

What it really loses from Huang is much harder to replace:

  • Customer relationships with OpenAI, hyperscalers, TSMC, and governments like the U.S. and China

  • A capital allocator behind some of the largest and complex AI deals ever constructed

  • The “switchboard” between all of the company’s moving parts, as there is no stated modus operandi or Chief Operations Officer (COO) in place

It would be like Apple losing Steve Jobs…

The people capable of building great products don’t disappear overnight.

But the person deciding which products matter, where the company goes next, and how it all pieces together does.

What makes this even more interesting is the timing…

NVIDIA’s board dismissed Huang within four days, saying it “has confidence in Jensen” just days before voting him out.

And,

He did not simply leave the CEO post, he also lost his board seat.

So yes, the entire “brain” of the operation is out of the equation, and that makes it significantly harder for the company to move forward.

Even with a recent shareholder petition to bring him back, and Elon Musk’s backing, Delaware law directly prohibits this election to take place especially when Huang only owns 3.5% of the company.

But, like I said…

Huang’s departure is the least of NVIDIA’s problems, especially as more than $500 billion in off-balance sheet debt could already be massively impaired.

That is why Tuesday matters so much.

Let’s get into it,

NVIDIA Lost Jensen Before Tuesday’s Earnings (What Now?)

CHART OF THE DAY

Roughly 27% of all pending data center projects in the U.S. are supposed to go to Texas.

But, there’s a problem…

The recent midterm election results create a new headwind for these projects, as the proposals point to curbing further construction.

What this means is a ~65% haircut to the expected computing power capacity for the next 12-18 months.

And the market immediately took this to heart.

ARE YOU COVERED? —>

Companies with contracted compute obligations dependent on Texas were the most punished around the election dates.

Those with already completed or non-Texas compute contracts didn’t sell off at all.

That should be the least of your concerns after the election though.

NVIDIA’S PRICE STRUCTURE

NVDA Stock,Thinkorswim

Before we cover what the market is pricing in for NVIDIA’s first earnings announcement without Huang at the helm,

Let’s begin by reviewing the current price structure in the stock:

  • $194 is the 20% drawdown from the highs, the official “bear market” cross

  • $147 is where the distribution changes for the year, a “P” shaped distribution

  • $180 is the deciding cutoff on whether we lose that “P” or continue into new highs

Where we are relative to those three areas matters more than anything that’s being talked about.

Losing $194 meant the underlying thesis for NVIDIA is not just being challenged, but rather it is being abandoned altogether.

Be it Texas-related data center cancellations, Anthropic’s $800 billion valuation haircut, or recent chipmaking competition from China…

It really doesn’t matter, the fact is the market no longer believes NVIDIA can achieve what it once promised.

And yes, this is exactly why the stock is so cheap on a P/E multiple.

That leaves us with the “P” shape distribution, and we’re now sitting at the lower half between $147-$180.

A post-scandal rebound failed to close the week at or above $180, and that means the odds of the market finding “fair” values for NVIDIA sit lower and closer to $147 for now, and the earnings print could cement that view.

What Markets Expect

NVDA Options Implied Earnings Moves, Offside Capital

Options traders now expect an ~8.6% move on Tuesday’s earnings announcement.

Much quieter than previous swings, and I’m guessing this is because of how certain participants feel about that “P” distribution and bear market implication from the chart.

Nonetheless,

It doesn’t remove the surprise factor from a few outcomes down the line, outcomes such as:

  • A successful succession plan for the CFO to take over

  • Bringing in an outsider CEO instead

  • Huang returning in a Steve Jobs heroic fashion to save the company

That would be the most positive outcomes, and my guess is we erase that $194 bear mark and return to the higher levels of the “P” distribution instead.

I’m talking a $195-$256 price range on these outcomes.

Now the bearish ones the market may be considering:

  • Unsuccessful succession, leaving unused talent without guidance and failed customer financing relationships

  • National security, export penalties, financing stress upon a deep Huang investigation (to begin on December 1st)

If those are the probable scenarios, I believe the stock goes toward $100-$120 or less.

Will it then be a buying opportunity?

Maybe, but more so than it is today and that’s for sure.

Historical Stock Performance After CEO Departures, Offside Capital

With the exception of Apple, Netflix, and Google…

Most high-profile CEO departures resulted in a 5% to 50% drawdown in the underlying stock relative to the S&P 500 index.

The worst performers, specifically, were due to a conduct-related departure as it is for Huang.

Just keep that in mind.

NVIDIA’s Bigger Problems

NVDA’s Revenue Growth & Gross Margins, Offside Capital

Most NVIDIA investors are of a simpler retail crowd, and that’s become evident through social media coverage.

What these people focus on is revenue and EPS growth, with the P/E multiple being a secondary focus.

Let’s unpack revenue first, because it looks fantastic at first glance:

  • 65x since 2015, 68% growth expected for next year

  • Gross margins sitting at an all-time high of ~75%

However,

The way NVIDIA generates this revenue, and what the quality of those revenues are, matter a whole lot more than just how much growth there is.

NVDA’s Changing Balance Sheet Through Time, Offside Capital

A company’s balance sheet is the ultimate driver of the business model, and therefore, how it makes revenue possible.

All through 2026, NVIDIA went from a cash-heavy position which allowed it to invest heavily into R&D and develop industry-leading chips,

To less than 20% of assets being in cash, used up in acquiring equity stakes in private and public AI companies.

Right now, over 50% of the balance sheet is made up of:

  • Private & Public AI stakes

  • Receivables

Which are one in the same.

NVIDIA buys a stake in Anthropic/OpenAI, then these AI companies commit to buying NVIDIA GPUs, that purchase commitment goes into “Receivables.”

Repeat this process across:

  • IREN, CRWV, INTC, NBIS, GLW

And what you get is an extremely bloated and weakening balance sheet.

What exactly do you own?

Asset & Shareholder Equity Composition of AI Stakes, Offside Capital

For every $1.00 you invest into NVIDIA stock, $0.43 goes to fund stakes in these other AI companies.

In essence,

Buying NVIDIA is like buying MSTR, the only difference is you are taking on leverage to invest into AI companies instead of Bitcoin.

You are no longer buying into a chipmaking company, know that.

The bottom line for revenue quality?

NVDA Day Sales Outstanding & Cash Conversion Cycle, Offside Capital

When this cash was used to transform NVIDIA’s balance sheet, something else changed.

Sales started to take longer to realize, now taking ~60 days on average as customers take longer and longer to pay for chips.

And,

NVIDIA’s top customers now get preferential treatment due to Chinese chip competition, those sales can take up to 180 days to show up!

The result is a slower cash conversion cycle, and a steady burn of working capital trickling through to weaker free cash flow and book value.

In other words,

Revenue growth is not enough if it destroys stock value.

Earnings “Must Watch” Checklist

NVDA EPS Bridge, Offside Capital

2Q27 earnings were not all that great, and I’m willing to bet this week’s results will be even worse…

$1.24 of the $2.46 reported earnings (50.4%) came from revenue growth, but no cash actually came into the business.

Another $0.19 (7.7%) came from higher values on these AI company stakes, again no actual cash coming into the business.

If 80% of all sales take 180 days or more to show up, then we can assume $0.99 of the $1.24 boost from revenue is fiction.

Take out that $0.99 plus the paper gains on AI investments of $0.19…

And you get a $1.18 haircut to true company earnings, less than half what they reported.

Ouch.

Even after a 50% haircut to reported earnings, whatever is left is not even that good in quality either:

NVDA Earnings vs Cash Flow, Offside Capital

The Sloan accrual ratio is set to measure what percentage of EPS are made up of actual cash versus fictional items.

Anything over 10% is considered junk, and highly probable to lead to a share price crash.

NVIDIA now sits at 25%…

Their Net income is closer to $50 billion while operating cash flow sits at ~$25 billion.

That’s the real problem here, and the biggest risk to a halving in EPS this year.

Ready for the big one?

NVDA Off-Balance-Sheet Commitments & Debts, Offside Capital

Out of the entire $3 trillion market cap in NVIDIA, nearly 18% is exposed to massive off-balance-sheet commitments.

Why are they off the books?

Because they are directly tied to the commitments and circular financing schemes assigned to private companies and other “unofficial” agreements with other public AI companies as well.

Most of these off-book deals are made up of supply and capacity, which have just come under direct threat from the Texas-related data center hurdles.

The second majority is in guarantees, which have more exposure to Anthropic and OpenAI’s ability to make good on their obligations.

If you read my Anthropic S-1 financial breakdown, you know the likelihood of that happening is diminishing almost by the day.

Going into the quarter, buyers want to see:

  • An improvement in the Day Sales Outstanding (DSO) ratio for better working capital allowances

  • Organic revenue/EPS growth from the core business rather than these “accrual” items to regain some of the earnings quality and a more optimistic P/E ratio

  • A closer gap between reported net income and operating cash flow to help boost the quality of the financials

If the above items get worse, and the stock pops on a mere revenue boost, I would be all for looking into a short position.

Otherwise,

Maybe the $210-$256 targets can be attained (perhaps higher) in due time.

Even though these risks are currently unacceptable, it’s nothing a controlled risk position (like options) can’t take.

You now have the ranges, the odds, and what to look out for in the coming earnings report…

Premium members will get the earnings coverage and the ultimate decision on what we plan to do about the stock moving forward.

WHAT’S THE TRADE?

There isn’t one trade in specific to take into account with all of these changes.

What’s more important is to understand the reason why these changes are happening in the first place.

Just like every such episode before this one…

When the markets get this concentrated and uncorrelated to what the underlying economic data suggests,

Combined with valuation extremes,

A scenario that made Buffett stay out of the market in 1999 comes into play.

A “lost decade” as they call it.

From 2000-2010, those who relied on the S&P 500 to get them through wealth creation and/or retirement failed to see any gains other than their monthly deposit rates.

I believe we are about to see another decade like that coming up.

In fact, my belief is beginning to turn into a reality.

And after studying 36 countries over 50 years, I’ve found the two-way formula that leads most investment capital into two specific areas.

So rather than give you one trade here for free…

I want to give you a new way of doing things moving forward.

A Final Note

COMING UP NEXT

  • NVIDIA’s earnings is on deck for this week, and so are Alibaba’s, so you can expect me to deliver the most in-depth analysis on the outcomes and future scenarios as possible.

  • A thorough review of the China opportunity is also on the list for you to review and consider in this potential “lost decade” scenario, leading hundreds of billions of capital into the most attractive emerging markets.

  • Speaking of which, China’s AI agents continue to take on further market share (45% of global AI usage last time we covered them) so that will most likely make up a large part of our following coverage of AI & China’s opportunity set.

Meanwhile, here is a video from one of my favorite non-BS YouTube channels covering the latest Michael Burry take on the big AI IPOs.

Take it with a grain of salt, the guy is probably perma depressed and always bearish, but it doesn’t mean you shouldn’t listen to hard evidence when it is found and relevant to all the tail risks I just outlined here today:

Until next time,

OFFSIDE RESEARCH

Against the Tape, Ahead of the Curve.