mattee2264

Changing of the guard really. Buffett did say he regretted not investing in Google back in the days when they had a wonderful monopoly with Google Ads a tollgate business. But Google are a different animal now they are betting their future on AI. Buffett has enough intellectual honesty and humility to admit he doesn't understand AI. Greg presumably thinks he does. One interesting perspective which might bring Google within Berkshire's circle of competence is that the Big Tech companies are becoming a lot more like heavy industry/utility companies with high capex requirements and making a lot more money from commoditised products such as cloud/compute and LLMs also are becoming commodity like as they are all trained on the same public databases and deliver similar outputs. Although those type of companies generally don't sell for 30x earnings so I think there is some overvaluation risk.

Anthropic "profit" is because they are getting a discount on compute costs from SpaceX as they ramp up.

Space X prospectus does have a definite dot com/South Sea bubble whiff about it. Pivoting their entire business model towards AI. Target valuation of $1.5TR. Over 100 times sales. Claiming a TAM of $28TR. Compelling story of data centers in space and colonizing Mars etc. Of course Elon is Elon and Tesla has had an anomalous valuation for ages while the rest of the market was reasonably enough valued but if someone with his promotional instincts is choosing this summer to IPO then it is probably a warning sign. Then you have Anthropic and OpenAI as well all rushing to IPO. Also hugely money losing companies. And if IPOs go as planned they will represent 5% or so of the S&P 500 and that is a lot of equity issuance for the market to absorb. There is a lot of debt being issued relating to AI and interest rates are trending up and if the market does get into trouble it is not clear how much leeway there will be for the Fed to cut rates. Although I imagine there would be a bailout whether from Mag7 or even government/central banks as tech companies have so much political power and have convinced the government that beating China to AGI is a matter of national security. The broader pitch seems to me to be that AI can displace jobs and therefore deliver huge OPEX saving and the tech companies are no longer limited by corporate IT budgets because they can make inroads into wage budgets. But AI isn't exactly free even if it appears to be because it is heavily subsidized to drive adoption and when something has a very low price then it is inevitable that demand is going to be incredibly high which is often used as an argument as to why the capex isn't irrational and we aren't in a bubble. And if everyone can use AI to reduce their costs and AI reduces the value of having an edge from being able to recruit highly skilled scarce labor then competition will drive down prices so it will be consumers who ultimately benefit. If this plays out like other bubbles then we will end up with overcapacity which will drive down the cost of compute which will be great for society but not so good for the first movers who funded the buildout. Then after the dust clears there will be companies who will take advantage of the cheap cost of commute and ubiquity of LLMs and build successful business models built around applying AI to solve real life problems.

Agree with the above if we were going through a gradual transition. But Big Tech are spending trillions a year on AI capex and the corporate spending required to provide sufficient return on those investments likely does need a significant reallocation of corporate budgets from wages to IT spend. Productivity is often defined as output per worker. For your more benign scenario where the productivity improvements means the technology pays for itself you need it to come primarily from higher output rather than same output but achieved through fewer workers as a result of automation etc. That may be true for some industries but I don't think it is the case across the board. Eventually I agree there will be some kind of reallocation of workers the same way there was during the agricultural and industrial revolution but it takes time and the transition period will be painful and I don't think governments are ready for it.

Hussman Funds did some analysis recently observing that consensus forecasts over the next year or two are projecting a broad based margin expansion presumably because of AI. In other words assuming that the benefits of AI will be significant, widely distributed AND realized quickly. Something that also seems to get lost is that if AI is as successful as earnings estimates seem to be projecting then there will be a rise in unemployment and many of the unemployed will be white collar workers with significant salaries. These workers will therefore need to cut back on consumption and even have to sell their investments to make ends meet. Government debt and deficits are already stretched so it is difficult to imagine much appetite to quickly implement some kind of universal basic income or increase the generosity of unemployment benefits.

Plus the public markets will be tapped only when it is incredibly favourable for these companies to do so. So public are perhaps more likely to end up as bag holders. I mean if you are valuing OpenAI, Anthropic etc at close to $1TR how much future growth are you already discounting? These are companies losing billions of dollars a year and being floated at as much as 65x sales. That is reminiscent of the 2021 SaaS madness. And these are market caps that put them in the same company as mature technology companies such as Google, Amazon, Facebook etc that took decades to grow into their valuations and only achieved them after demonstrating the strength of their moats, overcoming all kinds of competitive, regulatory and other challenges, posting consistently strong earnings growth and returns on capital etc. And as I pointed out in my earlier post these valuations are not the product of careful investors making conservative projections of future growth. The big placements have and will be with Mag7 companies who need the IPOs to be highly successful because the funds raised will be spent by Open AI/Anthropic on their products/services.

Could be a big year for IPOs. OpenAI and SpaceX over $1TR, Anthropic $500B and if those are successful no doubt quite a few others as well. Those are not insignificant amounts of money. Perhaps as much as 10% of the market cap of the NASDAQ. Especially within the context of so much money being passively managed these days. There are rules such as 10% minimum float requirement and 12m seasoning period. But there are ways for index commitees to fast track inclusion into their indices and they'd be under a lot of pressure to do so. That would require them to sell seasoned profitable companies with more moderate valuations to buy these unprofitable startups that have speculative unproven technologies. This would mean a lot of forced buying that would drive valuations even higher. It would also mean that an even greater % of the indices would to a large extent be a big bet on the AI theme. So if AI follows the same path of the internet (short term bust, long term success) does not really bode well for already expensive markets. I also get the feeling these IPO valuations are deliberately inflated because the float is so small and a lot of the companies investing such as Mag7 are forced to invest because OpenAI and Anthropic are customers and if they run out of money then it will have knock on effects for Mag7. Plus you always have SoftBank who will invest in anything shiny and new at any price. Then once the valuation is set and index inclusion is achieved there is a tidal wave of forced buying by index funds. So no real effective price discovery going on. Anyone else getting some 1999/2021 vibes?

Isn't the reason AI is using so much energy because all these startups are being heavily subsidized by Mag7 and a lot of the usage is because AI is being provided to consumers and businesses either for free or at low prices which result in the likes of OpenAI losing lots of money? Obviously historically that has been the playbook to drive adoption and later on you monetize it. But there doesn't seem to be a lot of visibility on what kind of returns can be generated on all this spending and it is clear that scaling is running into diminishing returns and we may never achieve AGI.

I think when he reckons 4-5% interest rates are normal and sustainable he is assuming that inflation returns to target. If the new normal for inflation is 3-4% and GDP growth is 2-3% then you'd think eventually interest rates get up to 6-7% (which is not that high by historical standards) and if that happens then I don't think PE multiples will stay well above 20x. Obviously financial repression is the plan and we've seen in the post WW2 era that central banks and governments are capable of doing so and maybe that will keep interest rates low and support high valuations. But that only works if it is accompanied by fast economic growth so you can grow out of the debt. If it is associated with low economic growth then you just get stagnation.

Top 10% of American wage earners account for nearly 50% of consumer spending. They also own lots of financial assets which have done amazingly well over the last 5, 10, 15 years. So that is a considerable ballast to the economy. Since the pandemic companies have increasingly focused on premiumisation raising prices rather than increasing volumes knowing that the rich will pay up and that will offset any loss of lower income consumers. Niall Ferguson was talking the other day how he feels we've returned to the Gilded Age of the 1890s. For rapid industrialization read AI capex build. For robber barons read Tech CEOs. Lots of corruption and poverty and huge wealth inequality. Of course that period from 1870s to 1890s had quite a few market panics and economic downturns. Although in those days there wasn't the central bank around to bail markets out but there is a lot of instability so it could still be a bumpy ride ahead.

This seems to be the trade de jour: pile into gold, cryptos and other scarce assets. Idea being that governments have issued too much debt and central banks will have to expand money supply and reduce interest rates to help finance it and this will lead to an erosion of the purchasing power of fiat currencies. At the same time for the first time since 1995 foreign central banks hold more gold than US treasuries as a share of their international reserves. There are also the usual claims that real world inflation is far above the 3-4% central banks claim. This has also been used as a justification for historically high stock market valuations on the basis that equities are a better inflation hedge than cash and bonds and cash is trash. Goldman Sachs are predicting gold will hit $5,000 and positioning themselves for more widespread institutional adoption of bitcoin and other cryptos. Other Wall Street banks are following suit. And of course retail investors are all over it. Kuppy was banging on about Project Zimbabwe years ago and while he has shifted his focus to the AI capital cycle on his twitter he liked a post pointing out that the S&P 500 has yet to recover from the dot com bubble crash in gold terms and said we need to care more about real returns from now on and gold is the scorekeeper. It sounds convincing at first sight. But I am generally sceptical when it comes to consensus trades and remember the general maxim that the best time to hedge against a risk is when hedges are cheap and not when hedges are at record highs. Gold has outperformed the S&P 500 over the last 25 years but 25 years ago gold was very out of favour and now it is very popular. Also there has been a mania for speculation since COVID especially amongst retail investors and gold and bitcoin lend themselves particularly well to speculation as there is no fundamental value to anchor to. Also I think if any of the dire consequences came to pass then gold and bitcoin would probably crash a lot harder than stocks. Any thoughts?

There is an awful lot of circularity which adds to the risks. Cloud companies are getting a big boost in revenues from the AI start ups (with as yet massively unprofitable business models) and a lot of it is investments in exchange for cloud credits. Similarly you are getting investments in exchange for purchase contracts/commitments. You don't have to expense the investment but you get to book the revenue. And NVIDIA gets most of its revenues from a handful of big Tech companies spending most of their operating cashflows on AI capex not because they have any great confidence in the returns on those investments but because they can afford to do so and feel the risk of underinvestment is far greater i.e. the investment is defensive to a large extent. And the circularity extends to the economy and the stock market. Investment spending adds to GDP both directly but also via the wealth effects and increased consumption. So if everyone were to cut back considerably on investment and a lot of this projected revenue doesn't materialize and growth slows or even declines then all of this goes into reverse and even rate cuts and extra government spending might not be enough to prevent a very nasty bear market and recession.

Saw something on Fin Twit today that got me thinking..... What happens when OpenAI decides to introduce advertising for free accounts? It has over a hundred million monthly active users. That is a lot of eyeballs. People also spend a lot of time on Chat GPT. Meta and Google have enjoyed a near duopoly over online advertising and it accounts for a sizeable chunk of their revenues. So what happens to these revenues and associated profits when they get competition for advertising dollars from Chat GPT? Especially if Chat GPT decide to price aggressively to attract advertising clients? Can't be good for profits.

I think you're very wise to recognize that your outperformance represents a mix of luck as well as skill and there is a risk of future underperformance that you are the luxury of insuring against by being more diversified as well as increasing your fixed income exposure. But if you are 70-80% invested in index funds and short term bonds (and I'd also add equity income funds to that mix) then there is plenty of room to add a little to returns by investing in 3-5 high conviction ideas. It may even reduce risk if those ideas are very uncorrelated with market returns.

Most industries aren't that competitive in this day and age. If they were then margins wouldn't have expanded so much as companies pocketed tax cuts, interest cost savings from cheap debt, and retained pandemic price hikes and other things you'd expect would be competed away. Also AI isn't free. So to some extent it will represent IT budget replacing staff budget with perhaps limited benefit to profits in the short term and adoption driven more by the worry that if you wait for AI costs to come down you will be way behind the learning curve that early adopters benefited from. But if AI does become cheap enough fast enough for near term mass adoption (and mass layoffs) then at that point Big Tech margins will probably be a lot lower and they won't be recouping a large chunk of their initial investments made when NVIDIA chips were super expensive and data centre rents similarly so.

https://foundationcapital.com/the-ai-hype-600b-question-or-4-6t-opportunity/ This is a pretty good article. A lot of bears are citing record valuations as a reason to be bearish. But if a few years down the lines companies will be able to reduce their headcounts by 10-20% and a chunk of the money saved goes to Big Tech with the rest increasing profits then both S&P 500 earnings are going to be an awful lot higher. There is a little uncertainty as to whether Big Tech will retain their market leadership or will be eclipsed by nimbler later entrants who aren't burdened by the huge initial capex investments and benefit from lower future chip costs and cost of computes etc. But given Big Tech are scooping up all the promising start ups and have a huge user base they can test their AI offerings on as well as synergies with their existing core products (e.g. Google/Meta can use AI to improve the return on their clients advertising investments, Microsoft can offer an AI office assistant and other productivity enhancing features, Nvidia may be the next Intel able to retain their technological leadership and capture most of the market share for each new generation of AI chips etc) then I think they will get more than their fair share of the AI revenue opportunity and if it is big enough then there is room for new entrants to come in and quickly reach large valuations that will be additive to S&P 500 market cap and earnings. But at 30x forward earnings Mag7 aren't exactly priced for perfection and their core businesses alone are still growing and you could almost argue that you are getting any AI upside for free and the upside could be considerable.

Totally agree that tariffs has become the new fixation in much the same way as people used to obsess over inflation prints and Fed dot plots and forward guidance. It is like COVID. As soon as people figured that the worst case outcomes were off the table the market recovered albeit with some pullbacks along the way but never got close to revisiting the lows.

There's been a trade deal with the UK. But they accepted the 10% baseline tariff which seems for now to be the floor. But UK are pretty desperate. Japan probably will also strike a deal sooner rather than later. EU strategy (which I think is smart) is to delay any deal until the 90 day deferral period expires at which point USA will be under a lot more pressure from the market to deliver deals. China I think are also desperate as even before tariffs their economy is struggling.

https://www.visualcapitalist.com/the-average-u-s-tariff-rate-since-1890/ This has quite a nice chart. Even if we do land on 10% universal tariffs for everyone including China (and Bessent suggested this is the floor and the UK trade deal seems to confirm this -for now at least) that would take us to levels not seen since WW2. And even with the increase in tariffs since Trump's first term end of 2024 the average US tariff rate was under 3%. If we look at a broad sweep of market history the last 45 years have had a lovely disinflationary tailwind that has allowed interest rates to come down and given central banks a lot more leeway to periodically bail out stock markets and given the US government a lot more leeway to increase government spending and cut taxation. To some extent that reflects technological advances and other productivity improvements. But it also reflects the opening up of the global economy and the availability of cheap inputs and other goods from the rest of the world. And the resulting large trade deficits USA ran with the rest of the world were compensated for by ROW purchasing USA financial assets especially US Treasuries. Trying to rebalance this even with much more modest 10% universal tariffs as well as trying to force US goods and FDI on countries (e.g. UK and Saudi Arabia trade deals) probably is going to backfire. Similarly trying to force USA companies to build factories in the USA. Central planning doesn't work. And this is what Trump sometimes looks on the verge of trying to do.

Interesting what Trump did in the Middle East. Looks as though NVIDIA chips will be used as bargaining chips in the trade deals intended to be struck over the next year or so.

For a while now there has been a view that there is a magic number on the 10Y or 30Y that will make something break. But they were saying that when yields were below 3%. And aside from the Silicon Valley thing there's been very little impact from Powell taking rates from 0 % to 5%. And 4-5% bond yields are never going to be competitive with a stock market that has almost tripled in the Roaring 20s.

A rapid V shaped recovery and a continuation to new highs is exactly what you'd expect of an event-driven bear market. Bull market is back on and while there may be a correction if trade deals don't materialize quickly after the 90 day pauses expire we will probably need to look for another catalyst for the next downturn as it is unlikely to be tariffs. Nothing imminently on the horizon. It is possible that Q2 and even Q3 economic data will be shit but markets will probably look through that because they'll attribute it to tariffs which have since been diluted. Even before tariffs there were a few concerns about Big Tech capex on AI. But so long as their core businesses keep growing strongly and they show no signs of slowing their investments in AI markets will probably give them the benefit of the doubt and wait patiently for the investments to bear fruit. I do think the US economy is very reliant on large government deficits and if they ever get cut there will be a nasty hangover but Trump doesn't seem serious about that and is already proposing to increase the debt ceiling by $4TR. Moderate inflation is bullish for markets especially when most companies in S&P 500 have pricing power and can pass them on to consumers and grow nominal earnings as a result. And the V shaped recovery will just embolden more investors to lever up and buy every dip and momentum will then take markets a lot higher

The new "Big Beautiful Bill" proposes an increase in the debt ceiling of $4TR. So I guess that is how we are going to fund the tax cuts.

Looks as though tariffs will be 10% for pretty much everyone when the dust clears. I don't know how much that will raise but probably not enough to fund major tax cuts. Especially not if interest rates head higher as recession fears are off the table which makes financing the existing debt load more expensive. Also it is clear that there is no rebalancing of world trade in Trump's favour. China got the same 10% as everyone else even though they retaliated. And all they had to do to get that was to remove their retaliatory measures and make some vague promises to open up trade and consume more. If anything things are more negative compared to before Liberation Day because it seems the 10% universal tariffs are a floor (for now) and that is still enough to cause some inflation and slow the economy.

It is still market manipulation. Especially as it is quite likely he told his buddies before making the public announcement.