Intro
The Art of Investing's three presenters work through Nvidia's results, the bond market's fight with the Treasury, and the case for and against gold, then take three viewer questions on stock picking, ETFs and what happens when companies start issuing shares again. Rich McDonald chairs, with Chris Fellingham on bonds and Mark Holden on equities, known on air as CJ and Spice.
Host: Rich McDonald
Also on: Chris Fellingham (CJ) and Mark Holden (Spice)
Published: 28 August 2026 on The Art of Investing
Episode page | 1 hr
Key Takeaways
Nvidia guided to 70% revenue growth and the market barely moved
The old consensus was 44%; the guidance covers the fiscal year ending January 2028
Growth is limited by supply, not demand, which is why Nvidia is funding its own customers' infrastructure
The multiple is getting cheaper while the numbers get better
24 times today, falling to 14 times a year later if the guidance is delivered
Druckenmiller called the Treasury's bond buying price management, not liquidity management
He was Scott Bessent's colleague at Soros, where Bessent ran the office
Only 13% of actively managed US large-cap funds beat their benchmark over ten years
27% managed it over the single year to June 2026
Stock dispersion is at a decade high and stock pickers still are not beating the index
There are now more ETFs quoted in the US than there are US-listed companies
Gold's problem is real yields, not the inflation argument
Gold has historically done badly when real yields are high, and they are back around plus two
The supply of shares is turning from a tailwind into a headwind
Shrinking share supply added roughly 0.7% a year to US returns between 2015 and 2025
Exceptional earnings growth is the offset to that dilution, not a reason to ignore it
The exit is coming but is not here yet
A 50% tariff war between Trump and Carney pulled international money out of US markets
The Producer Is Away and Spice Is at the Fringe
Rich McDonald opens by noting that the show's producer, Sophie, is on holiday and there is nobody supervising the three of them
"So again, the lunatics have taken over the asylum." — Chris "CJ" Fellingham
Rich's version: with Sophie gone there is nobody to force CJ to laugh at his jokes. CJ insists that has never been a problem
Mark "Spice" Holden is recording from Scotland and spending the next couple of days at the Edinburgh Fringe
His son is in a play called Nesting, running at Summerhall, which Spice says has great reviews
He is also seeing an Australian comedian, Garry Starr, who "does the whole show apparently naked apart from a pair of flippers"
Asked by Rich whether he will be in the front row, Spice says he is going nowhere near it
Druckenmiller Breaks With His Old Soros Colleague at the Treasury
CJ picks up the previous week's bond story, when Treasury Secretary Scott Bessent "tried to manipulate bond prices because he didn't like those naughty boys in the bond market, doubting his fiscal credibility"
A new player has entered the field: Stanley Druckenmiller, and CJ's point is that he is anything but a neutral observer
Druckenmiller ran Duquesne Capital and charged three and 30, which CJ explains for the audience as 3% a year in fees plus 30% of any money made
By comparison, CJ says the industry standard now is a 1% fee with a 10% share of the profits, down from much larger numbers ten to twenty years ago
Before Duquesne, Druckenmiller ran the Quantum Fund with George Soros. Rumor had it he was the real brain behind the investment themes, while Soros had the sixth sense for which ideas to press and the willingness to take enormous positions
Bessent was the administrator at Soros, making sure the office ran smoothly
Bessent himself told the FT last year that "in macro, there's Stan and there's everybody else" — which is why CJ frames this as a former mentor criticizing his own protege
Druckenmiller called the Treasury's bond buying plan a mistake, and said it was "not liquidity management, as Bessent was saying, but price management"
His position, as CJ relays it, is that if the bond market wants yields higher because fiscal policy is out of control, it should be allowed to do that
The effect on bonds has been very little. A falling oil price is good for bonds because it lowers inflation, so yields came down a touch
Against that, nobody knows what Fed chairman Kevin Warsh will say at Jackson Hole on Friday, and CJ's read is that bonds have not decided which way to break yet
Rich points out that CJ made the same criticism of the Treasury a week earlier, and jokes about the Soros alumni lining up together
The Week in Markets
Oil is down about 7% on the week, partly because Oman and Iran have come to an agreement on revenue share
Ships look likely to be charged to pass through the Strait of Hormuz, though the amount is not yet known, and the oil price has fallen in anticipation
Another poor week for chips and semiconductors, right up until Nvidia's results
Bitcoin is up another 12% and back above $80,000, a couple of weeks after the show suggested the crypto winter might be over
Gold and silver are up 2 or 3%, and exchange-traded funds specializing in Bitcoin, Ethereum, gold and silver have seen their biggest inflows in ten months
The only all-time high among the major markets the show watches was the FTSE 250
CJ notes that UK prime minister Andy Burnham said he would not upset the applecart on growth or on companies, and does not want to create instability
His caveat: what a politician says and what he does may be two entirely different things — but as mood language for a bullish UK story, it is the right sort
Trump has slapped a 50% tariff on Canadian goods and Canada, led by former Bank of England chairman Mark Carney, has matched it tit for tat
"we've got all out war in North America" — Spice, who blames the trade war for the week's drag on US equities relative to the rest of the world
Nvidia Guides to 70% Revenue Growth
The results hit at about 9:20 UK time and beat all expectations, but market makers immediately marked the shares down 2 or 3%
That reversed when the finance director spoke on the conference call and gave guidance for the fiscal year ending January 2028: 70% revenue growth, against an old consensus of 44%
Rich asks whether investors are missing the best stock in the market because they think it has got too big, or remembering that semiconductors are cyclical — and whether 70% growth on 24 times earnings is simply too cheap
"I mean I hate to correct you, but it was 70% revenue growth next year" — Spice, whose point is that the distinction matters because equity investors prefer revenue growth to earnings growth, which can be flattered by buybacks and other tricks
"the revenue that will be generated by this company will have trebled to nearly $670 billion in one year" — Spice calls the number mind-boggling
Nvidia said openly that growth is being constrained by supply, not demand, which Spice says is why it is spending and lending to infrastructure suppliers
Spice calls it high-quality, repeatable growth and invokes the law of large numbers: a big percentage increase on an already very large number is enormous
The shares moved from down about 3% to up 3 or 4%, and were up about 6% as Spice spoke ahead of the US open, lifting the Nasdaq and the S&P 500 with them
Why the Multiple Is Getting Cheaper as the Numbers Get Better
CJ's observation for listeners wondering how the process works: the numbers keep coming through and the stock price is not going up, so the multiple is getting cheaper
He reads that as caution rather than mania — the market is showing it is suspicious about whether the growth will come through, which gives investors some protection
His first concern is that the finance director said exactly what you would expect him to say about the balance-sheet-as-a-service model, in which Nvidia helps fund the spending that becomes its revenue
Morgan Stanley published a document on the debt ratings of the hyperscalers and whether to buy the debt funding data center construction, and came out neutral
Rich's read: that is from an investment bank that wants all their business
"No investment bank is going to say sell when they want the business. So the worst they can say is neutral." — CJ
The bank's stated reason was that the tail risks are too great, and it mapped the circular funding into flowcharts
CJ's historical analogy is the CDO market of 2006 and 2007 — he stresses he is not suggesting we are there
Rich asks what CDO stands for; CJ explains collateralized debt obligations, where debt is piled into a vehicle and investors buy whichever piece they like
"And they skated off into the sunset as they left the financial system to destroy itself in 2008" — CJ on the bankers who built the product, once the underlying mortgage credit turned out worse than anyone thought
The parallel he draws is to complexity, and to Nvidia funding its own chip purchases through other private debt holders
His second concern is real yields: historically, as bond yields and interest rates rise, price/earnings multiples fall, and 24 times may be too expensive in that environment
CJ is clear that the numbers were great and the market was right to react positively in the short term
The Case for Buying Nvidia on an 18-Month View
Rich's framing: a lot of people running money today have never seen valuations contract on rising real yields, and the idea seems to have been forgotten since the financial crisis
Spice walks through the math: the multiple drops to 14 times a year after this one, and if the AI boom is still running by January 2028 the shares will not be on 14 times but nearer 20 — still lower than today's 24, and still a 50% increase in the share price
Spice's reason for trusting the guidance is legal, not sentimental: the finance director cannot knowingly or recklessly, to use the SEC's words, make numbers up, because that would be misleading the market. There have to be orders behind them
CJ's rebuttal is that this is an argument from authority. You expect a company to say it, and afterwards it can always say that is what it thought would happen and it did not
Two years ago, when the market was speculating about how much money Nvidia might make, it traded on 70, 80, 90 times and was right to. Now it says 14 times two years out, and CJ says maybe it is right again
"just because Nvidia goes up doesn't mean that the index will go up because their revenues are the costs of Amazon and Meta and Alphabet" — CJ, noting Nvidia is about 8% of the S&P 500, so bullish on the stock is not bullish on the index
Rich, a self-described natural optimist, is happy to sit with what he expects to remain the biggest company in the world for a while yet
The Model Portfolio: Miners Lead, US Indices Lag
The show runs a model portfolio showing what the three of them, with 100 years of investing experience between them, would do with their own money
Performance: up half a percent on the week, 12.9% year to date and 25.4% since inception
The best performers were commodities rather than just gold — the BlackRock World Mining Trust up 4.7%, copper up 1.9% and emerging markets up 1.8%
Rich reads it as the inflation trade returning alongside yield curve control from the Treasury; Spice agrees commodities are the biggest beneficiary
The three biggest losers were all American: the US industrials position down 1.2%, the Russell 2000 down 1.2% and the Nasdaq 100 down 1.1%
Rich attributes the US weakness to the Canada trade war, with international investors taking money home — which also helped emerging markets and the FTSE 100, the portfolio's fourth-best performer
Spice calls it a value bounce after a long stretch of lagging, plus profit-taking in tech ahead of Jackson Hole: every dog has its day
What Kevin Warsh Might Say at Jackson Hole
Spice's position is that all bets are off until the Fed chairman speaks on Friday, and that he hopes Warsh will row in behind the Treasury Secretary
"Remember he doesn't believe in forward guidance." — Spice, on why the speech may say nothing about the direction of rates
His hoped-for outcome is a dull speech that leaves markets steady into a not-too-bad September, with the caveat that September is not a great time for risk markets
CJ's version of the risk runs the other way: Warsh has said he wants a proper functioning bank balance sheet again, and does not want the Fed acting as a buyer of the equity market to keep equities doing well
If Warsh succeeds in restoring a properly functioning bond market, investors have to factor in a higher cost of capital
Viewer Question One: Why Bother With Single Stocks
The first viewer question: if most investors, including the professionals, underperform, why invest in single stocks at all rather than going the ETF route
"And that is don't bother investing in the single stocks." — CJ, who is careful to say it is not because he wants it to be true
Over 40 years of interviewing managers, as a hedge fund allocator and as a chief investment officer, he has found it very difficult to find people who can create alpha
He cites a Wall Street Journal article from the previous week, "Stock-Picking Funds Are Performing as Poorly as Ever"
"only 13% of actively managed US large cap equity funds outperformed the benchmarks" over ten years through June 2026 — about 1 in 10
Over the single year from June 2025 to June 2026 the figure was 27%, or roughly 1 in 4
The paradox is that this should be a stock picker's market. Dispersion — the gap between the best and worst performing stocks, some down 60% and some up 200% — is at its highest level in a decade, and fewer people are outperforming anyway
Rich turns to Spice, his equity expert, and asks whether less efficient markets full of passive money are exactly when you should pick stocks
Spice's answer is speed. News is now absorbed in fractions of a second: Nvidia's shares went from down 3% to up 6% the moment the finance director spoke, a 10% swing with no chance to react
Rich discloses his own trade — he bought shares at 205 the night before
Spice's conclusion is that watching stocks that closely is a full-time job, and that even he no longer has the time. ETFs may be the lazy route but they are also the safer one: fewer big up days, but none of the pain on the big down days
Factor Favoritism and the ETF Alternative
CJ wants to believe stock selection works, and has seen numbers suggesting it does — but when he looks at the periods where it worked, he finds the managers were aligned with a factor
Momentum, value and quality are all factors, and a manager who is really just delivering a factor can be replaced by the equivalent ETF for ten basis points
His example is Terry Smith, a long-standing quality investor whose last three years have been poor because quality has not done particularly well
CJ's bottom line: he is sure some people can do it, but he does not have the time to find them, so he would rather buy an ETF
Rich's follow-up is that with factor ETFs and industry ETFs available, ETF investing now looks a lot more like single-stock investing than it used to
"So there are now more ETFs than there are stocks and shares." — Spice, who counted roughly 4,400 ETFs quoted in the US against about 4,200 quoted companies
An ETF does exactly what you want it to do, he says: exact replication of the S&P 500 or the US industrials, for a relatively small fee of 20 to 30 basis points, against a manager who may hit the target one year and miss by a mile the next while charging over 1%
Spice started his career as a stock picker and calls himself a convert: unless this is your full-time job, ETFs are undoubtedly the way
Viewer Question Two: Gold's Real-Yield Problem
The portfolio has not held gold for a long time. Spice bought some for himself when gold dropped under 4000, and the show ignored the idea
CJ says he loves gold and hates gold, with no lukewarm territory in between, and judges it on macro conditions
They made a lot of money in gold in the original portfolio, sold it around the $4,000 range and kept exposure through the World Mining Trust
It then went above 5000, he stayed out, and on his account it has come back to 4000 and done nothing for about nine months
Spice's case is the classic one: for hundreds of years gold has been treated as an inflation hedge because it is a real hard asset in very limited supply
The bond market wants central bankers to raise rates against rising inflation, and central bankers, particularly in America, are reluctant, which he calls a good backdrop for gold
CJ names three drivers — whether inflation is here and getting worse, real yields, and the dollar — then takes a shot at the inconsistency: "Spice doesn't believe there is any inflation. So why is he buying gold? It doesn't make sense to me."
CJ's real-yield history is the heart of his skepticism. On his account real yields were about 4% in the old days, moved down to 2%, went negative, and are now back at plus two. Gold has done poorly in the past when real yields were high
Gold looks cheap against its highs, but the underlying macro picture of rising real yields does not help it
Rich's explainer for the audience: a bar of gold in the safe at home pays no interest, no coupon and no borrow, while money in the bank or in bonds pays more as real yields rise
Spice's counter is that when governments or central bankers are no longer trusted to do their job, people hide in assets outside the system — which he also thinks explains Bitcoin's return to form as the Canada-US trade fight escalates
If Bessent keeps doing what he is doing and Warsh does not raise rates, the dollar should weaken over time
His more specific point is about who has been buying. The Gulf War started almost exactly six months ago, and the surplus cash that emerging markets and China had been putting into gold, silver and copper was diverted into paying a much higher oil price
Central banks were the biggest buyer of gold for a long period and have been out of the market since
With oil peaking and falling back, and possibly falling further on an Oman-Iran deal, that surplus should return
"We're at about 4600 today. The all-time high was 5600, give or take 5595" — Spice, whose short-term view is a return to the 5000 level as a big psychological marker, and then a run at the high
He is positive but not exclusive about it: there are plenty of sweeties in the shop
CJ will not disagree, and notes with satisfaction that Spice has quietly abandoned the inflation argument he started with
He agrees that what Trump has done may undermine the dollar, and that undermining the dollar builds interest in gold
If Jackson Hole calms the bond market, gold may push higher still and they may find themselves buying next week
If the authorities lose control of the bond market, he thinks that hurts gold too, so the two arguments may offset
He is not bearish — the World Mining Trust holds gold, silver and copper — he just cannot get excited
Rich's own view is that this comes down to the US government and whether the debasement trade is back. He does not believe spending will be brought under control, and thinks the hunt for somewhere to store wealth outside the dollar is real
Viewer Question Three: The Shrinking Supply of Shares Goes Into Reverse
CJ sets up the third question: over the past 15 to 20 years equity has been taken out of the public markets by company buybacks and by private equity taking listed companies private, while few new shares arrived
Rich glosses IPO as initial public offering for anyone who needs it
The result was scarcity value in equities, led by the US, where buybacks have been heaviest and the market has performed fantastically
"the shrinking supply of shares added roughly 0.7% a year to US returns" — CJ, citing a paper from an asset manager covering 2015 to 2025
The same paper argues the process is reversing: more IPOs, more issuance, fewer buybacks and fewer takeovers
In the internet bubble, the volume of stock that came to market dragged returns down by as much as 10%
US issuance is picking up sharply, with SpaceX and Anthropic among the names in the pipeline
New floats sell only a small slice of the company, and the insiders locked up for nine months to a year and a half all start selling when they can, so far more stock arrives than expected
"drag down the index performance by — wait for it — 4.5%" — CJ, on what the paper's authors expect new share supply to do to index performance over the next decade, given how much is coming relative to historic supply and demand
Rich puts it to Spice: if that is where we are, does it not mean lower valuations from here
Whether Earnings Growth Can Absorb the Dilution
Spice concedes the argument cuts against his usual bullishness. Companies raise money three ways: a bank loan, the corporate or private bond market — pay you back in ten years at maybe 2% over what the US government pays — or, least often, issuing equity
Companies issue equity when they think their shares are expensive, and use those shares to buy companies whose shares they think are cheap. So a wave of issuance is itself a signal
His offset is the earnings cycle. The quarter has just ended and the US is going to get 30% earnings growth in the second quarter, which he calls almost unprecedented outside a recovery from recession
Both sides can be right, he argues: performance could end up meaningfully lower than it would otherwise have been and the returns still be good, because the fundamentals are improving. A really good year that is merely a good year
CJ's response is that the ground rules are changing and the jury is out until the market sees how the new supply gets placed
Spice disputes the ten-year horizon specifically. Companies are looking for a three-to-four-year payback on their capital expenditure
If the returns do not come, the stock market is a lot lower anyway, because investors will conclude the money was wasted
If they do come, the companies will be making enormous amounts of cash and will start buying back shares again
"You can't be bullish on Nvidia, which is receiving all the revenue for the capex and be bullish on those that are spending all the money on the capex" — CJ's trap: if the hyperscalers keep spending, that is Nvidia's revenue, and if they stop, Nvidia halves in value
Spice's answer is that the demand underneath is real. He has been paying for iCloud and Microsoft storage for a decade and the price keeps going up
His reminder for listeners is that the cloud is just a very large data center, and the race is to build enough of them to keep up with demand for storage and space
Who Is Left to Buy
Rich's worry, aimed at listeners who have not been investing for more than ten years: AI spending is filling the debt markets, enormous IPOs are soaking up cash, and now companies will start issuing stock again, removing the invisible hand of buybacks that has supported the market
CJ agrees it takes away a number of supports but not all of them
More and more people have got involved in equities, and there are places that have not discovered the attraction of stocks at all — the UK, where a lot of cash sits idle, being one
The case is undermined bit by bit, and nobody knows where the tipping point is
He adds real yields to Rich's list: most people in markets are used to very low interest rates and easy borrowing, and that is not how it has always been
CJ's example of belief overriding arithmetic comes from a Lex column on the price war in footwear, whose last paragraph turned to AI
Anthropic investors told the FT the company's IPO could be worth 2 trillion, even as the price of a lightweight model was cut by 80% and Anthropic cut its own prices by the same amount
When people want to believe, they believe
"The time to hit the exit is coming, but I don't still believe we're there yet." — CJ, who returns to his image from the week before: the kids in the back of the car asking whether we are there yet, and the answer being not yet, but be careful
Rich asks Spice, with producer Sophie presumably watching the clock, who the marginal buyer is in those circumstances
Spice's answer is savings rates and where they go
The UK saves about 4% outside pensions, up from about 3% three years ago, but does not put it into stocks
America saves about 3% and everything else goes into the stock market, on top of very large pension assets
"The Europeans do not want to invest in equities. They are traditionally bond investors." — Europe's savings rate is still 14%
His marker: when Europe's savings rate falls to 4%, that is the time to sell
CJ recalls Spice saying in many meetings that when the retail investor takes over and starts buying, that is the time to worry
The exchange ends on the observation that retail money in Europe is going into the new defense spending instead, and CJ's one-word verdict: defense bonds
No changes to the portfolio this week. Rich says Jackson Hole and the market's reaction to it will decide whether they add to the precious metals complex, increase equity risk, or run for the hills
The show left the portfolio untouched and split three ways on what comes next, with the equity man looking for gold back at 5000 and 70% revenue growth to carry Nvidia, the bond man saying the exit is coming but not yet, and the chair waiting on Jackson Hole to settle it.
Products, Companies & Tools Mentioned
Nvidia (Guided to 70% revenue growth for the fiscal year ending January 2028 against a 44% consensus, on revenue Spice says will have trebled in a year; the argument that ran through the whole episode, both as a stock and as the counterparty to everyone else's capital expenditure)
Gold and silver (Up 2 or 3% on the week; the subject of the second viewer question, with Spice looking for 5000 and CJ arguing rising real yields work against it)
BlackRock World Mining Trust (The portfolio's best performer on the week at 4.7%, and the way the show keeps gold, silver and copper exposure without holding bullion)
Bitcoin and Ethereum (Bitcoin up another 12% on the week and back above $80,000; Spice reads the strength as distrust of governments rather than a technology story, and crypto ETFs saw their biggest inflows in ten months)
Duquesne Capital, Quantum Fund and Soros Fund Management (The lineage behind the week's bond story — Druckenmiller ran the first two, Bessent administered the third)
US Treasury and Scott Bessent (The bond buying plan Druckenmiller called price management rather than liquidity management)
Federal Reserve and Kevin Warsh (The Jackson Hole speech the whole week is waiting on; Warsh does not believe in forward guidance and wants a properly functioning bank balance sheet)
Morgan Stanley (Published a note on hyperscaler debt ratings and landed on neutral, citing tail risks — which CJ reads as the most negative call a bank that wants the business is able to make)
Amazon, Meta and Alphabet (The hyperscalers whose capital expenditure is Nvidia's revenue, and whose costs rise as Nvidia's sales do)
Russell 2000, Nasdaq 100, S&P 500, FTSE 250 and FTSE 100 (The week's scoreboard — three US indices led the portfolio's losers, the FTSE 250 posted the only all-time high, and the FTSE 100 was the fourth-best performer)
Copper (Up 1.9% on the week and, in Spice's telling, one of the physical inputs the data center build-out has to buy)
Terry Smith (CJ's example of a manager with a factor preference — a long-standing quality investor whose last three years have suffered because quality has not worked)
SpaceX and Anthropic (Named as the issuance pipeline that could bring a wave of new stock to market; Anthropic also appears in the Lex column CJ read out, with an IPO its investors put at 2 trillion)
Apple iCloud and Microsoft (Spice's decade of rising cloud storage bills, offered as evidence that data center demand is real and pricing holds)
Books & Resources Mentioned
"Stock-Picking Funds Are Performing as Poorly as Ever" – The Wall Street Journal (CJ's source for the 13% ten-year and 27% one-year hit rates on active US large-cap funds)
Morgan Stanley's note on hyperscaler debt ratings (Neutral, with the circular funding mapped into flowcharts and tail risks given as the reason)
An asset manager's paper on shrinking and returning share supply (Puts the buyback era's contribution at roughly 0.7% a year from 2015 to 2025, and models the reversal over the coming decade)
Lex, on the price war in footwear (CJ's example of a column that starts on shoe pricing and ends on AI valuations)
Get the latest market chatter and takes as they happen:
X | Threads | Instagram | YouTube | TikTok | Facebook