Research

Build it and they will come?

---

ON September 3, 1821, 37 years after James Watt invented the concept of horsepower and started charging £5/horsepower/year for the use of his steam engine, the first demonstration of the production of a mechanical force by the interaction of an electric current and a magnetic field was given by Michael Faraday in the basement of the Royal Institution in London.

1820 is, in many chronicles1, noted as beginning of the industrial revolution. It marked the end of 50 years of war2, whose conclusion unleashed pent-up demand. That demand was served by relatively free trade, guaranteed by British naval power, coupled with property rights and financial instruments which allowed inventions of the previous century to finally be harnessed for commercial gain. The steam engine is the canonical invention of the era and its implementation as a source of power for factories, shipping, rail and machinery is legend.

Not usually noted as coincidental to this period is the electric motor. Yet not long after Faraday's 1821 demonstration in London, the first DC electric motor capable of turning machinery was invented by William Sturgeon. In 1832 he built on Faraday's electromagnet by incorporating a commutator which oscillated the direction of current to ensure continuous rotation.

However, for 30 years after its invention, no practical uses could be found for the electric motor. It wasn't because there was no demand. The steam engine was in use and, as Watt's £5/horsepower/year business model demonstrated, very lucratively so. The electric motor was not in use because there was no electricity distribution system. There were no power plants, there was no grid. The motor could not stand alone.

In contrast, the equivalent distribution system for steam was the coal industry. Coal as fuel pre-existed the steam engine. Used not just for heating, it was the fuel of foundries that made iron and other metals. More energy dense than wood, coal had been mined for thousands of years. From mining concessions to barges for transport on newly dug canals to bunkering for storage, sourcing coal was a well-developed, stable and profitable industry.

Finding coal to power a steam engine in the 19th century was not a problem. Finding electricity to power your electric motor was. You could not store it, you could not ship it, you could not meter or price it. It required a completely new value network and infrastructure.

For this reason, even though the electric motor was available through most of the 19th century, the period was still known as the age of steam. The electric grid—an infrastructure—was only possible by the late 1800s as generators, capacitors and AC power were perfected. It wasn't until the early 20th century that the age of electric power began.

This illustrates perhaps better than any other example that the device precedes the infrastructure. That is, that the instrument of utility precedes the network that enables it. the boat precedes the port, the locomotive precedes the railroad, the automobile precedes the motorway, the telegraph precedes the telegraph network, the radio precedes the radio network, the phone precedes the phone network, the cell phone precedes the wireless network and the smartphone precedes the social media. The device induces the supply of its own power.

The implications are profound. It puts to the test the idea that we need to build infrastructures before we know how they will be used. The inventors of electric motors or lights relied on batteries. Investors had to imagine a grid to power more than one. The inventors of the radio had to demonstrate a single transmitter/receiver pair. Investors had to imagine broadcasting. The inventors of the car had to find fuel in pharmacies. Investors had to imagine a gas station. Steve Jobs demonstrated the iPhone without an App store. Investors had to imagine a Services business.

As AI has emerged as a technology, we are in a strange position where the interface, born as a chatbot, suggested an infrastructure that is massive in scale. The magic sauce of GenAI isn't so much the algorithm, which is quite a simple idea. It's the scale of the data and computation that permits the illusion of intelligence.

In developing the commercial idea, investors are imagining the device while investing in the infrastructure. In other words, the investors are poised to put trillions of dollars to work on what amounts to "the grid" without knowing exactly what the motor will be.

The cash flows associated with infrastructures are very different from those of devices. Devices are iterated rapidly and evolve to changing tastes and discoveries of behavior. Infrastructures are built once, maintained and perhaps re-built but only after decades. Therefore financing for infrastructures is more likely through bonds (debt) rather than shares (equity). Returns are also much more subdued as the risk for infrastructures is lower. Regulation and relations with labor are also dramatically different.

Moreover, few businesses implement both the device and the infrastructure. Indeed, the capital structures, financing, longevity and valuations of the two types of businesses are dramatically different. In cases where they are merged, regulators typically intervene to separate them.

Watching from afar, it would seem that if AI is indeed infrastructure then it will not provide a brilliant business model. Like all network business models, you either iterate constantly at great expense to deliver the next generation or you harvest the built environment to pay back the investment, hoping that a parallel network does not steal users. But the cruelty of the model is that if you have a monopoly of any kind, your margins are beaten to zero by regulation and cost of parasitic maintenance.3

We can see this with all current "grids". From electric to hydrocarbon distribution, to communications and even food distribution. The struggles to avoid commoditization, corruption and corrosion of trust and a lack of innovation are endemic.

The device however, is another story.

1. I recommend "The Birth of Plenty: How the Prosperity of the Modern World Was Created" by William J. Bernstein. Bernstein argues that in order to prosper, a country must possess four main attributes: property rights, the scientific rationalism, capital markets and an effective means of transportation and communications.
2. Reverse chronology: French Revolution, 1803-1815, (including the War of 1812 in US & Canada), which followed a decade of the French Revolution 1789-1799 which followed the American Revolution 1765-1783. This global war, involving most world powers, lasted 50 years.
3. In a research note published April 7, Goldman Sachs strategists led by Chief Global Equity Strategist Peter Oppenheimer acknowledged the historical basis for those concerns. "The history of technology breakthroughs, from the steam engine to railways, PCs and the internet, is littered with examples of new technologies that attracted large sums of capital to build out underlying infrastructure which have led, ultimately, to low returns," he said. "The gains are then enjoyed by other companies, many of which piggyback off the original investment," he added.

Redundant or Post-PC?

Today marks the 16th anniversary of when Apple released the first-generation iPad. After Steve Jobs announced the ‌iPad‌ on January 27, 2010, it launched a few months later on April 3, 2010.

The ‌iPad‌ was the first device that Apple released with a custom-designed chip, the Apple A4.

The concept of a tablet device like the ‌iPad‌ actually predated the iPhone, and Apple had a prototype project in the works in 2004. Then-Apple CEO Steve Jobs wanted to prioritize the iPhone, so the ‌iPad‌ was put on hold until after the iPhone was established.

Apple priced the ‌iPad‌ at $499, and Jobs called it a "magical and revolutionary device at an unbelievable price." Like the iPhone, the ‌iPad‌ dominated the tablet space, and it still continues to do so today. Apple sold a million iPads in the first month, and while competitors like Microsoft and Samsung quickly came out with competing tablets of their own, the ‌iPad‌ has no match.

Apple has led the global tablet market for over a decade and no competitor comes close to matching its tablet market share. As of the fourth quarter of 2025, Apple's market share was 44.9 percent, with Samsung the next closest at 14.7 percent.

Apple has expanded its ‌iPad‌ lineup over the years, and we now have multiple sizes at multiple price points.

Source: MacRumors

The graph above shows the iPad revenues compared to the Mac and the iPod. I included the iPod since the iPad was originally dismissed as a "big screen iPod touch". On the other hand, it was also seen as a PC or Mac substitute. To many, including Alan Kay, the tablet form factor makes much more sense than the PC form factor. It's more approachable, more intuitive and certainly smaller and more portable. As the graph shows, quickly after launch, the iPad overtook sales of the Mac.

The problem arose that whereas the iPad was smaller, the iPhone what smaller still--so much so that users would take phones with them everywhere. At the same time, the "PC" shrank as well so that many notebook (laptop) computers became thin, light and had day-long battery life. The Mac at least became no bigger a burden to carry than an iPad.

So the iPad was stuck in the middle between an iPhone, that did more and more, and a Mac that was very portable. At this point many said that the iPad was a dead end form factor.

However, nor was the iPad a PC killer nor was the PC and iPad killer.

As the graph shows, the iPad continues to do rather well. It's lagging a bit behind the Mac but sales have been, for at least a decade, highly correlated to the Mac. iPads are therefore less seasonal and more product cyclical. The category has seen resonance for the very young, the very old and also for many professionals.

 

What iPad has done is carved a niche that is resilient. After 16 years, I reckon about 910 million have been sold with about 70 million more sold each year. The product has a particularly long lifespan in use and may continue to be used for as long as 10 years. Total lifetime revenuel has been over $403 billion with about $29 billion last year. The total user base is between 400 and 450 million users. In the U.S., the iPad is highly prevalent, with over 40% of Americans owning an iPad in 2024.

Neither killer nor killed, the iPad is just a great "in between" product with plenty of value. I have no doubt that it will still be popular 16 years from now.

A tale of two Apples

What are the odds of it surviving another year, ten years, 50 years?

Companies don't last long. A recent study that analyzed thousands of European firms yielded a corporate life expectancy figure of 12.5 years. The Boston Consulting Group analyzed the entry, growth, and exit dates of 35,000 publicly listed companies and determined that one in ten of all public companies fail each year, a fourfold increase since 1965.

As I illustrated many years ago in a talk titled "The Unfortunate 500" most companies die not long after adolescence. Of the Fortune 500 1955-1960 cohort, the 81 companies which joined the index in those years, only 7 remain there today. One third of the corporations on the 1970 Fortune 500 list were no longer in existence thirteen years later. Broadly speaking, as a population, large companies that make it to the Fortune 500 listing have life expectancy within that group of 17 years.

To illustrate further, here is the list of the companies that joined the Fortune 500 In 1976, the year of Apple's birth:

Entry	Company	Exit
1976	SuCrest	1976
1976	International Systems & Controls	1977
1976	Kewanee Industries	1977
1976	Hunt International Resources	1979
1976	Peabody Holding	1979
1976	Consolidated Aluminum	1981
1976	HP Hood	1981
1976	Envirotech	1982
1976	Peabody International	1984
1976	Hughes Tool	1987
1976	Cameron Iron Works	1990
1976	Foxboro	1990
1976	Tektronix	1994
1976	Mapco	1998
1976	Molson Coors Brewing	2009
1976	Baker Hughes	2014
1976	Masco	2014
1976	Ball	2014
1976	Ecolab	2014
1976	CF Industries	2014

Not one of those in the 1976 cohort are on the list today. The longest to survive managed to stay on for 38 years.

Apple joined the Fortune 500 in 1983. Here is a list of all the other companies that joined the index in 1983.

Entry	Company	Exit
1983	North American Coal	1983
1983	Datapoint	1986
1983	MEI	1986
1983	Mobay	1986
1983	Coleco Industries	1988
1983	Liquid Air	1988
1983	AH Robins	1989
1983	Sipco	1989
1983	Wilson Foods	1989
1983	Oxford Industries	1991
1983	DR Holdings Inc of Delaware	1992
1983	Media General	1992
1983	Wisconsin Dairies Cooperative	1992
1983	Ametek	1994
1983	Hanson Industries NA	1994
1983	Meredith	1994
1983	Snap On	1994
1983	Sun Diamond Growers	1994
1983	Universal Foods	1994
1983	Worthington Industries	1994
1983	AMDAHL	1994
1983	Dr Pepper Seven Up	1994
1983	Diebold	1994
1983	Echlin	1998
1983	Stone Container	1998
1983	Tyco International	1998
1983	Flowers Industries	2001
1983	Cooper Tire & Rubber	2005
1983	Tribune	2008
1983	Holly	2011
1983	Apple	--

Apple is the only survivor from that cohort.

Out of about 2500 companies that existed in the Fortune 500 index since 1955, 50 are still in it. That's a survival rate of only 2%. Approximately 87% of all people born globally between 1955 and 2026 are still alive today. This is why I titled my talk The Unfortunate 500. If this was a human population it would be considered chronically unhealthy.

However, unlike people, large companies that make it past age 50 tend to stay alive (or independent, as most who exit the index do so because they are merged or acquired--another distinction from biological entities).

But the way they manage to be resilient is by adapting to a changing world in some fundamental ways. The environment a business lives in (technological, market, economical) changes more rapidly than the biological environment (genetic, radiological, reproduction).

Companies have to essentially re-define themselves and become something that has little to do with their initial conception.

In the case of Apple there are clearly two separate entities. The original and the one that was reborn in 1998 when Steve Jobs returned. When you look at the company closely, it's pretty clear that the second Apple was fundamentally different even though the original founder returned.

In some ways one could argue that Apple 1.0 eventually strayed from its original culture and the reset in 1998 returned it to a core that was resilient but it's more complicated than that.

To measure "distance" in terms of corporate culture you need to think of culture as a three dimensional vector. it's a composition of resources, processes and priorities (RPP). The combination of these forms what we call the "DNA" or the "culture" or the unwritten constitution of the firm.

The vector of RPP for Apple changed dramatically in Apple 2.0 in terms of resources (capital, software, silicon), processes (engineering, operations and marketing), and priorities (focus, integration, product/service cohesion)

I would summarize it as there being a tangible "theory of the firm" for version 2. The first version was very much a theory of the team: a group of great people without a binding operating model.

Stay tuned as we explore the Theory of Apple.

Cash flow is a fact, earnings are an opinion and share price is a dream.

A company's total worth is the number of outstanding shares multiplied by their value. This product is called the market capitalization or Market Cap.

That's a simplification. If you want to be more accurate, you should take the Market Cap and net out cash and debt--these are not part of the future flows of cash that can benefit shareholders.

More precisely, there is another measure called the Enterprise Value (EV) which is the total price tag of a company, representing what it would cost to buy the entire business and pay off its debts. EV = Market Cap + Total Debt + Preferred Stock + Minority Interest - Cash and Cash Equivalents.

This is the precise and complete measure of the firm's value but we use the shortcut of Market Cap since In most cases the Preferred stock and minority interest are negligible and if cash equals debt then the result is EV =~ Market Cap. it's easier to calculate anyways as the shares outstanding and share price are easily available. Not so for EV.

To get a sense of how much you're paying for the company it makes sense to divide the Market Cap by the company's Earnings so you get an idea of how many years it will take to get back your investment. When dividing both Market Cap and Earnings by the number of shares it becomes the famous P/E ratio or Price per share / Earnings per share.

But if you want to be fully precise, instead of using Market Cap you should use Enterprise Value. In that case the denominator should be Free Cash Flow or FCF1. The ratio becomes EV/FCF. Free Cash Flow is a better measure than earnings because it takes into consideration depreciation, change in working capital and capital expenditures. It measures the actual cash a company generates after paying for its operating and capital expenditures.

EV/FCF takes into consideration unusual situations such as huge amounts of cash or debt or capital expenditures which removes shareholder access to what the company "earns" on paper. This why the adage of "earnings are an opinion but cash flow is a fact" exists.

I used EV/FCF years ago to value Apple because at a time when it held huge amounts of cash. This was before it began aggressively reducing its cash through its capital return program. The P/E ratio at the time was very low but excluding the cash, it was even lower. It was a huge red flag indicating that the company was being priced for liquidation, or as some analyst put it, priced like a steel mill going out of business. If at that time you did not believe that its demise was imminent you would have surely been richly rewarded.

Today the anomaly that throws the P/E off is not Cash but CapEx. And it's not for Apple but for the other mega caps or AI "hyper scalers". Without worrying who is or isn't in the "Magnificent 7" I compiled the largest 10 market capitalization companies and looked at their P/E and EV/FCF ratios. At time of writing:

Company	Market Cap	P/E Ratio	EV/FCF
NVDA	$4,014,456,963,072	33.6	44.4
AAPL	$3,624,949,514,240	31.3	30.3
GOOG	$3,304,174,911,488	25.2	44.3
MSFT	$2,667,926,257,664	22.5	35.0
AMZN	$2,157,182,320,640	28.0	297.0
TSM	$1,641,528,688,640	30.6	50.0
TSLA	$1,333,163,982,848	332.0	225.0
META	$1,356,802,949,120	22.8	32.6
BRK-B	$1,023,780,323,328	15.3	44.1
LLY	$793,550,454,784	38.6	142.3

As recently as six months ago, there was a lot of talk about an AI bubble. Back then I looked at the data then and by looking at the P/E ratios convinced myself that the valuations were not "bubble" level. They were in the mid-thirties. These were not historical high especially when considering the growth these companies were experiencing (over 30%). Since then prices have come down significantly.

This is as the CapEx elephant in the room has been making itself known and there is renewed alarm at the scale of capital spending. This is reflected in the EV/FCF number but not in the P/E number.

The table above shows how much more elevated the EV/FCF ratio is relative to P/E for many of the companies. This "truer" measure of what you are paying as a share buyer is now reflecting in reduced share price and hence reduced P/E.

Of course, this is not true for Apple. Apple's P/E ratio and EV/FCF are nearly identical. To see this more clearly I put together these graphs.

The graph shows the P/E vs. EV/FCF for the companies in the table. The graphs are the same data but the one on the right is zoomed in on the lower left corner (note the red squares).

The reason for the zoom is that there are three outliers. Tesla, Amazon and Lilly. Tesla has absurd valuation on both P/E (300 to 400 range in last 6 months) and EV/FCF while Amazon has an outrageous EV/FCF value due to abnormal capital spending. Basically, Tesla has negligible earnings (and negative growth) while holding astronomical share price and Amazon is spending far above its median on CapEx. Amazon's 10-year EV/FCF is much lower, at approximately 55.

But let's ignore the outliers and concentrate on the other seven in the cluster. These are below 50 P/E and 50 EV/FCF thresholds.

Note that I drew a diagonal line representing the equation of P/E = EV/FCF. This should, under normal circumstances, hold for most stocks. Just like Market Cap should roughly equal Enterprise Value so Earnings should equal Free Cash Flow.

Most companies should not stray too far from this line. But what we are seeing, with the notable exception of Apple, is that most tech companies have cash flow far below earnings. This is again due to the CapEx elephant.

The argument goes that this is a temporary situation with that capital being wisely deployed for returns from AI. Presumably, these companies will swing to the other side of the line, where cash flows will exceed earnings, compensating for the capex that was denied to the shareholders.

This is where the old adage comes into play: Earnings are an opinion but cash flow is a fact. I added to that "and share price is a dream". Take the free cash flow and multiply it by a "dream" and you get the price the market sets. This dream is the multiple.

What makes the AI "bubble" interesting is how it's not fueled by debt2 but by cash flow. Companies are turning over large portions (almost all and sometimes more) of their cash flows to investing in the dream of AI profits. All but one, of course.

There is something obviously contrarian in Apple's strategy. Perhaps it's not Apple that is odd but rather there is groupthink on behalf of the others. The question of who's right will rest on whether it turns out that AI is going to commoditize quickly or perhaps never, whether AI turns out to be modular or integrated, whether AI exhibits disruptive or sustaining competitive responses, whether it's an infrastructure subject to regulatory corrosion or whether it will migrate to the edge and iterate freely, whether it's a new experience or simply better existing experiences, whether it creates new markets or just makes existing ones more efficient.

These are not new questions. They have been applied to all software and hardware innovations of the last 40 years. We have a lot of history to inform us.

Join me as we learn.

1. To calculate FCF, look at the Income statement, start with earnings before interest and taxes (EBIT), add depreciation, subtract taxes and CapEx.
2. There is some debt of course, but the scale is so immense and the risks so high that most lenders would not touch this type of investment.

On April 4, 1975, Bill Gates and Paul Allen founded Microsoft, a software company destined to become a tech behemoth (and key Apple frenemy).

Perhaps the most interesting thing about Microsoft and Apple is just how similar the companies were in several ways. Gates and Steve Jobs, the CEOs who came to represent the two companies, were both born in 1955. Both later dropped out of college.

Gates and Allen created Microsoft when they saw a gap in the market following the launch of the Altair 8800 kit computer. In Silicon Valley, the Altair inspired the creation of the Homebrew Computer Club, where Jobs and Steve Wozniak first showed off the Apple-1 computer.

Source: Cult of Mac

Much more was said of Jobs' departure signaling the end of Apple than of Gates' departure signaling the end of Microsoft. And much more was celebrated of Apple 50 than Microsoft 50. Why is that?

If you were to structure an experiment to determine the influence of processes and personalities on the outcomes of a company, you would be hard pressed to find a better pair to study. Apple and Microsoft were founded almost at the same time, due to the same enabling technologies (microprocessors) and led by two iconoclastic founders of the same age acting out of similar impulses.

The divergences were that one concentrated on software while the other on hardware. But rather than pairing up in a match made in heaven, both tried to outmaneuver the other in order to avoid dependency.

For many decades, Microsoft's approach of modularity for hardware (Processor/components) but integration of software (OS/Core apps) seemed to win vs. the integration of Apple's hardware and OS. The licensing of Microsoft software to commodity hardware builders was far more successful than the Apple's insistence on quality of experiences. At least in terms of PC users, Windows overwhelmed the Mac.

This lasted until mobile computing demonstrated that integration was the only way to deliver the experiences that consumers would use in narrow time slices. Mobility meant that the pendulum swung towards the device and the device was good enough only with a high degree of integration.

But stepping back from the titanic clash of architectures, what strikes me is how much more personified Apple appears to be in the public perception. Apple *was* Jobs far more than Microsoft *was* Gates.

I think this perception is wrong but it indicates something about how we think about causality of business managment. I call it The Auteur theory of business.

Auteurism originated in French film criticism in the late 1940s. In this theory, film is an art form and as such it must have an author, and that would obviously be the director. "The best films are personal ones, made by forceful directors" - James Agee, critic. The film is therefore art only if there is a singular mind behind it and, conversely, if there is no such singularity of mind then it's not art. Unartistic film was merely staging or placement of performers.

But films are complex and multi=causal with many "artists" involved, any one of whom could make or break the film. Critics of auteur theory demonstrated that all the films that were celebrated as "auteur" masterpieces were inherently collaborative.

And so it is in business. The premise that "the best companies are personal ones, made by forceful founders" means that the best companies cannot outlast their founders and surely something that lasts only as long as the duration of a career cannot be considered great.

In the case of both Microsoft and Apple, we have seen successful transitions from founders to successor managers that not only preserved but broadly expanded the companies. Cook, Ballmer and Nadella outdid their predecessors as measured by every metric of financial performance. They both reached the pinnacle of business success as measured by market capitalization.

It's romantic to think that business are embodiment of people but in reality they are the seemingly sterile combination of resources, processes and priorities. These are heavily influenced by the people who assemble or define them but ultimately the processes must outlast biology or we are doomed to have to re-invent everything.

Put another way, if a founder builds something that does not outlast him then what has he really accomplished?

How everyone is wrong about AI

The newspaper was an unexpected development

It was not a product that anyone asked for. Like many innovations, it wasn't an invention. It was not patented or discovered in a laboratory. It emerged as a result of the convergence of three disparate technologies: continuous roll printing presses, the telegraph, and the railroad. This convergence occurred by the 1830s, facilitated by a business model aptly named “the penny press”.

Stepping back, the printing press was initially developed in the 1450s, 40 years before Columbus' voyage. It went through numerous evolutions. The big change was moving from batch (page-at-a-time) to continuous (rotary) processes. In the 1600s, Gutenberg type presses could operate at the rate of 240 impressions per hour. By 1818, The Koenig press could reach 2400 impressions per hour.

The steam-powered rotary printing press, invented in 1843 in the United States by Richard M. Hoe, ultimately allowed millions of copies of a page in a single day. Mass production of printed works flourished after the transition to rolled paper, as continuous feed allowed the presses to run at a much faster pace. Hoe's original design operated at up to 2,000 revolutions per hour where each revolution deposited 4 page images, giving the press a throughput of 8,000 pages per hour. By 1891, The New York World and the Philadelphia Item were operating presses producing 90,000 4-page sheets per hour. -Wikipedia

By harnessing steam engine power, the marginal cost of printing approached zero and typesetting productivity was such that tens of these pages could be laid out in hours. The production problem was solved.

The telegraph gave rise to news services which could relay national or international news to local papers nearly instantly. Small towns did not need to employ correspondents to reside on assignment to report what was happening in any and all national capitals. New syndication business models represented by Reuters and the Associated Press would distribute to thousands of smaller papers and communicate the stories via the Victorian Internet. Local reporters could flesh-in the paper with the latest from the community. The content sourcing problem was solved.

Finally the railroad offered daily distribution from urban to rural populations. This meant the entirety of a country's population would have access to news and commentary and, “all that was fit to be printed” on a daily basis. The distribution problem was solved.

A Bundle and Institution

From the capital point of view, the greatest investment for a newspaper publisher was the printing plant. The bigger the plant, the bigger the circulation. And the bigger the circulation, the bigger the profit. Printing presses kept getting bigger and faster. An old newspaper man told me that a newspaper is nothing more than a mechanism for depreciating a printing plant.

In order to pay off this investment, publishers not only sold papers but also advertising. To appeal to more readers they added a variety of content types beside news. Sections emerged to cover sports, lifestyle, photographs, opinion, comics, serialized novels, financial information, entertainment/gossip (thus forming the Hollywood star system). This included style pages, advertising and serialized novels. Some of the greatest literature of the time by authors like Mark Twain and Charles Dickens was first published in newspapers. Weekend editions were an obvious product extension.

Thus an institution was born. The newspaper became the primary source of information for the world. It became highly influential and a political (and propaganda) instrument. It came to be the historical record. We still seek out old papers in archives to find out what happened day-to-day in history.

The bigger the bundle the better the business

What made it resilient was its continuous bundling of additional information types, catering to diverse audiences. There was something in a newspaper for everyone. For men and women, from youth to elderly and for rich and poor. Since the cost of adding pages to the paper was marginal, bundling ensured a growing audience. This audience attention could be monetized through newsstand sales, subscriptions and advertising.

Obviously nobody would read a newspaper cover-to-cover. There was far more information than could be consumed and also far more than was interesting to any one person. The style reader probably did not care much for sports. The Hollywood gossip reader did not probably check the finance pages. But most readers of newspapers bought the bundle because it offered optionality.

They did not need to read it all but they could if they wanted. Maybe a scandal in Wall Street would attract the gossip reader occasionally. Maybe a pennant race would attract the style reader. If something exciting happened, you could read all about it.

Consumers were happy with the bundle and the newspaper became a tentpole holding up the information a society needed to function. It swayed elections, influenced public opinion and investigated the bad and the ugly in society. By the turn of the century, more than 9000 newspapers were publishing daily or weekly in the US. A proliferation such that almost every one of the 3140 US counties had more than one.

Unbundling the Newspaper

Then the internet arrived. Gradually, it became possible for every single piece of the bundle in a newspaper to be made available as a web page, free with some display ads. If you were interested in style, there were 1,000 style sites. If you were a sports fan, there would be 1,000 sports sites. News junkies could get the latest news faster online. Financial news was now real-time. No longer did you have to wait for the morning to find out what happened in the markets the day before.

As national and global news could be obtained on web sites, newspapers hoped to continue as sources of local information. That also gradually got peeled away by social media. The result was nothing short of catastrophic. Since 2004, the United States has lost one-fourth (2,100) of its newspapers. Weekday circulation declined 45%, from more than 122 million to 68 million. Consequently, half of the counties only have one newspaper, and two-thirds lack a daily newspaper. The loss of newspapers has resulted in the displacement of 36,000 journalists, including photographers, editors, comic strip artists, and copywriters.

The internet unbundled the newspaper but also the music album, the magazine, the TV channel and almost all forms of entertainment. Unbundling built new empires: iTunes/Apple Music, Netflix, Spotify, YouTube, Facebook.

This unbundling led to a new way of finding opportunities. You could actively look for bundles and once you found one you could ask when and how it would get unbundled. And vice versa. Jim Barksdale summarized it as: “i know of only two ways to make money: bundling and unbundling.”

So much so that if you found a huge bundle that somehow resisted unbundling you would have to investigate it for what's wrong with it. The classic example is the automobile. As a bundle of many types of trips and configurations into a simple but over serving business model clearly it was ripe for an unbundling. That is what ride sharing (with or without a human driver), and new modes represented by micromobility seek to do.

The AI Bundle

The iPhone is an enabler in this game, capable of either bundling or unbundling depending on the apps which are built for it. Wearables do some of the side quests but the phone is still the core of the personal network.

I would contend that AI is exhibiting a lot of bundling characteristics. We assume one model is good enough for all types of tasks even though there are pitfalls in privacy, security or complexity. Being a prompt mechanic feels like more of a DIY experience. But indications are out there that AI will be unbundled. According to The Information,

“Apple has complete access to the Gemini model in its own data center facilities. Apple can use that access to produce smaller models that power specific tasks or are small enough to run directly on Apple devices so they can run the tasks faster.”

This approach relies on a process called model distillation, where a larger AI model trains smaller versions that can run efficiently on-device. These smaller models reduce the need for cloud processing and improve speed, which fits Apple’s focus on privacy and performance.

“Since Apple has full access to Gemini, its student model can also learn to imitate the internal computations that Gemini uses to arrive at its answers.”

But it does not stop there. The implication of AI as a job destroyer is also challenged by bundling theory.

Jobs as Bundles

We have long been warned that AI would eliminate millions of jobs. One of many forecasts predicts that by 2030, approximately 10.4 million US jobs, 6% of the workforce, will be lost to AI. The underlying assumption behind these figures is simple: if AI can perform tasks that are currently performed by humans, the human workforce will be rendered obsolete.

A new paper by Luis Garicano, London School of Economics, Jin Li and Yanhui Wu, University of Hong Kong suggests it's not that simple.

The paper argues that jobs aren't tasks. They're bundles. For instance, radiologists don’t merely read scans; they interpret edge cases, collaborate with clinicians, and approve decisions that individuals make. Replacing the image-reading aspect doesn’t necessarily equate to replacing the Radiologist job.

Authors draw a line between “weak bundles” and “strong bundles” based on their ability to be split apart. Weak bundles can be easily separated, while strong bundles cannot be without losing their value. You can think of bundling strength as also the degree of modularity/integration. A weak bundle is more modular, a strong bundle is more integrated.

"In weak-bundle occupations, AI automates some tasks and narrows the boundary of the job… In strong-bundle occupations… AI improves performance inside the job, but does not remove the human from the bundle."

In jobs with weak bundles, AI doesn’t merely replace a task; it fundamentally reshapes the job. The human is left performing tasks that the AI cannot, often in a more specialized and limited capacity compared to the original role.

Once AI takes over a portion of the work, humans cease to divide their time. They dedicate themselves entirely to the remaining tasks, resulting in increased output per worker, i.e. higher productivity. Thus even weak bundles don't necessarily imply job reduction but increased output.

Artificial intelligence is transforming job roles, rather than eliminating them. While certain tasks may be reallocated and productivity could potentially increase, employment and working hours have remained relatively stable. In many instances, the bundle holds.

In a strongly bundled occupation, AI still improves task-performance, but the human retains both tasks and a larger revenue share. Whether the aggregate labor share rises or falls depends partly on the distribution of coordination costs between tasks within an occupations.

If most employment sits in strongly bundled jobs, the aggregate labor share is more resilient than the task-exposure models predict. The distribution of bundle strength across the occupational structure is therefore a first-order empirical question for understanding AI’s macroeconomic impact.

In other words, the doom/utopia AI outcome dichotomy is false. It depends on the potential for unbundling. If you're in an integrated (strong-bundle) job, blending judgment, context, and responsibility, AI is sustaining: it will make you faster and better paid. If you're in a modular (weak-bundle) job, it will shunt you into an integrated role, whereby productivity rises.

The degree of job displacement will depend on the degree of bundling in most jobs across the economy. It may turn out that integrated jobs will be added as the opportunities created by productivity will be market creating.

History is filled with examples of restructuring of ideas and the re-configuration of businesses due to technological change. AI is an example of both enabling and disabling of existing businesses and occupations. Very much the way the Internet turned out to be.

Displacement happens but innovation creates more opportunity and reallocation of effort to better (or at least more) outcomes. The newspaper and internet created markets and jobs. Will AI do the same? Modularity may hold the answer.

Says Neal Stephenson who first coined the term “metaverse”

This revelation came soon after Meta canceled the Metaverse.

My thought was immediately, will Meta rename itself back to Facebook? We'll get to that soon, but first let's look into Neal Stephenson's mea culpa:

When I was working at Magic Leap, and people asked me why I thought that [Metaverse] was a good idea, I would ask the rhetorical question: “do you really think that twenty years from now everyone is still going to be going around all day staring at little rectangles in their hands?” At the time it seemed obvious to me that the answer was no.

Reader, I have changed my mind. Twenty years from now, everyone is still going to be staring at handheld rectangles. Or at least that is the case if the only alternative is wearing things on their faces. Maybe this should have been obvious to me given the amount of time, effort, and money people put into making their faces look as good as possible.

A possible workaround is to keep refining and miniaturizing the devices to the point where they just look like eyeglasses. This, however, turns out to have the unintended side effect of making these things seem sinister. It happened with Google Glass, which instantaneously spawned the term “glasshole,” and it has happened again with Meta’s product that looks like normal, albeit heavy-framed glasses.

When someone around you is staring at a rectangle in their hand, it might be incredibly annoying, but at least you can tell they’re doing it. When someone’s wearing a head-mounted display, on the other hand, you don’t know whether they are looking at you or not.

Likewise, when someone holds up their phone and aims it at you, it’s obvious that you are on camera. That’s not true in the case of glasses or goggles. So it’s creepy.

Timing is everything and 20 years is always kicked around as some "infinite time horizon". It's notoriously difficult to predict when a transition happens and too many times we give up because we were certain how long it would take. That's rather the story of my life.

To remedy this, first, I never thought Magic Leap was interesting. Nor did I bother even skimming stories about Google Glass. I went nowhere near anything made by Facebook/Meta. Did not even try to use any one product, did not read specs, reviews or any other commentary.

I knew the products would not be very good being not much more than demos or experiments. To understand the market, you first need to make a marketable product and these were not good enough.

The goodness here was defined in the absence of creepiness. I've always argued that wearables are easily categorized: just look for places on the body where jewelry is placed and you'll know where wearables will be worn.

Jewelry has been developer for thousands, perhaps hundreds of thousands of years and it is inextricably connected to our psyche. Glasses do pass for jewelry but only if they conform to high standards of style. Functional objects rarely are granted status as ornament. Finding that balance is the key. It happened with watches and maybe ladies' hats.

Stephenson is right in saying that face ornaments are difficult "given the amount of time, effort, and money people put into making their faces look as good as possible."

That does not disqualify face wearables entirely. It simply needs design thinking and that's what was always what I expected from Apple. The Apple Vision Pro was not meant to be used socially. It does not even work outdoors.

It's a productivity product and one aimed at professionals. Apple will release a face wearable only when it conforms to its principles of user-centric design.

That was not the case for Meta because it is not a company concerned with customer satisfaction. Its business model is to sell users to advertisers and was motivated to enter device development only so as to disintermediate the smartphone. Doing so would remove any impediments to harvesting users data. The priority of user affinity and affection was nice to have but it requires a maniacal attention to even have a chance of success at being loved.

Rebranding around a hypothesis like the metaverse was impulsive but Facebook was a dying brand anyway and having diversified well beyond it, deprecating it was no great loss. Slapping on the Meta moniker was a bark up the wrong tree but not much was lost. I doubt there will be another change unless Zuckerberg falls in love with another way to get rid of its distributors.

NB: Apple never chose to use Metaverse branding. What it developed was spatial computing. It approached the problem not as a content but as a platform and user interface. Will we move to spatial computing en masse to the detriment of "rectangles in our pockets". I don't think that the right question. In reflecting on the past of computing, a new platform did not just replace the previous but created new markets altogether.

Each new interface was an order of magnitude more popular than its predecessor. Keyboard, mouse, touch, each added a few zeroes to the number of users. We are now at nearly 1:1 machine to human. I don't think that this is the limit.

We are entering the n:1 machines to human and wearables are the only way we get there. And as I said, where there is jewelry there will be a machine. Spatial interfaces are more natural as far as our perceived reality is concerned. The rest is engineering.

Make no small plans

50 already? It seems only a blink of an eye since I wrote Apple at 40. Let's try copy & paste.

In Apple’s first 40 50 years it shipped 1,591,092,250 5,126,211,000 computers1.

This shipment total is higher than any other computer company in its first 40 50 years. Actually there are no other PC makers that are 40 50 years old. One computer maker (IBM) is older but they only sold PCs for 24 years and what they still sell they don’t sell in high numbers.

That does not make Apple the top seller in a given year. Looking at only the Mac, Apple’s traditional form factor personal computer, Apple is fifth fourth in market share. If including the iPad, it was the top computer vendor in 2011 and, including the iPhone, it was first in 2009.

The following graph shows the top computer makers by units sold since 1975, a year before Apple was founded. (Click on the graph to see a higher resolution.)

Note that the graph is log scale and in multiples of 1000. Following the yellow Apple line we can see how it started with fewer than 1000 units shipped in its first year to today when it is shipping more than 1 million computers every day2. It's also quite remarkable to count how many other companies have come and gone in this top five list and imagine how many came and went in the "other" category.

Who remembers Atari, Commodore, Packard Bell, Compaq? Few said that they were doomed while they were in the running. Going back to Apple at 40,

After having a 40 50 year run and after selling more computers than all American and Japanese computer companies put together, how should we think about the next 40 50 years?

First, clearly Apple shifted from being a “computer company”. It has already changed its name to exclude the word “Computer.” That [came to mean that] it sells devices (which happen to also include computers.) We don't use computers to compute anyway. We use them to make decisions, keep track of things, speed things up and to communicate and to entertain.

After convincing the world that it's a not just a computer company, the next challenge has been to convince the world that it's also not just a product company. It did so by launching Services as a group (formerly known as iTunes Software and Services) and then separating the financial reporting so that Services can be measured with its own gross margin. Year after year management emphasized Services growth in its commentary. Services has grown into the second largest segment, 26% of total revenues ($112.8 billion out of 435.6 billion in calendar 2025.)

Services, if it stood alone, would be number 40 in the Fortune 500 list of largest companies by sales. It has been impossible to ignore it. Considering how much customers actually spend on Apps, Content and Services, the total is even bigger at $240 billion last year3. As a result, some, perhaps most, recognize Apple as a Products & Services company.

The App store alone attracts $169 billion in spending per year (2025) and that is now 5x what theatrical movie sales bring in world-wide.4

So if Apple is today at least a bundle of Products and Services then what will Apple be 50 years from now?

Again from Apple 40:

My simple proposal is to think of Apple as a customer creator. It creates and maintains customers. The more it creates, the more it prospers. The more customers it preserves the more it’s likely to persevere. This measure of performance for a company is not easy to obtain. It’s not a line item in any financial report.

The closest figure we have is that today Apple has one 2.5 billion active devices in use. It has around 1.1 billion users 5. Using the devices as a proxy, we can say that this audience has more than doubled in the last decade.

So we know that at least Apple has created a large number of customers. Knowing total devices sold and number in use means that we can be sure that 63% of all the computers Apple sold are still in use. This means that most customers are recent customers. Are they going to come back?

If we then factor in the size and persistence of Apple’s audience (customer base) we might conclude that there are factors which will allow it to grow rather than shrink. My analysis based on public data shows that the customer base has been increasing steadily.

And here is a graph showing the size of device base and implications for subscriptions. services and comparisons with other consumer aggregators. Note the forward looking nature of a linear extrapolation of the historical data.

10 years ago, I continued:

Combining the history of customer creation and customer preservation with the value obtained from each customer implies that the next 40 years will be about creating another large tranche of customers whose willingness to spend on whatever Apple creates will be relatively unchanged.

We don’t know the limit. One billion was hard to imagine even one year ago. We might see two billion devices in short order. Perhaps not. Perhaps as we have a multitude of devices about us all day the number will become less meaningful. But if the number of middle class customers grows and as Apple keeps its products within their reach, there is no reason to think that there will be a reversal of the last 40 years in terms of customer acquisition.

This analysis contradicts the current perception of the company as reflected by its stock price. That price assumes the audience will wither ((With a P/E ratio of 11, without a change in revenue per user, the only justification would be the audience shrinking by more than 90% over the next 20 years.)) That is not insanity. It is what we could expect if Apple were unexceptional.

We did see 2 billion. We actually saw 2.5 billion. The assumption of a company in distress as reflected by the P/E ratio of 11 is gone. It's safe to say that Apple is exceptional.

But note that we are only 10 years into that 40 year forecast from a decade ago. Do we still feel confident that the company can last 30 more years given what we've seen so far? If one company could get through the next 30 I would say it is Apple. And if it is so then I think it will get through the next 50.

One more Thing

A journalist from The Guardian recently called to ask my opinion on what were Apple's greatest hits and what were its flops. I think it was for a piece on the 50th anniversary of the company.

I thought about it very little and said that the obvious big hits are iPhone, Mac, iPod, iPad, Watch and Vision Pro. The failures are products like wifi routers, PDA, camera, iPod speaker dock and anniversary products.

I went on to say that the pattern for successful products is that they re-defined human-computer interfaces thus inducing a new platform that sustained the company for at least a decade. Specifically, they introduced a new input method: iPhone with multi-touch, Mac with the mouse, iPod with the scroll wheel, Watch with the passive health sensors and Vision Pro with hand and eye tracking. The platforms that matched these innovations are: iOS, MacOS, iTunes (and Apple TV & Services thereafter), HealthKit and finally VisionOS and immersive content. Each platform recruiting an ecosystem of developers, content creators and users resonating in mutual benefit.

The pattern for failed products is that they were modules or components in a bigger network of devices. They may have been neat but they did not induce a new experience profoundly enough to attract their own ecosystems. They did not have gravity fields. Furthermore, as modules, they did not integrate software, hardware and services. They were one or the other. Something else had to integrate them and because of that they were easily substituted.

Now this might be a nice footnote to the story arc of Apple at 50 but I'd say that it's more than that. The pattern to be recognized here is that the success and longevity of Apple are entirely because it chose the right targets for its capabilities and these targets were sufficiently ambitious. In other words, it aimed for big things, it made big plans. In the words of Tim Cook, it chose something where it could make a significant contribution.

It was Machiavelli who said "Make no small plans for they have no power to stir the soul." It would seem perhaps fitting that Steve Jobs thought alike when he said he wanted to make a dent in the universe. That it has done.

So, the plan is to make more of the same. It's more than enough.

1. Including Apple II, Mac, iPhone, iPad and iPod touch; excluding Apple Vision Pro, Apple TV and other iPods. Includes Q1 shipments estimated
2. 2025 shipments were above 364.9 million excluding Apple Vision Pro, Apple TV, Apple Watch.
3. A Fortune 15 ranking. This would be the case if payments to developers were counted as sales but that is not the case for accounting reasons. As a result Services margins are an extraordinary 76%, on the order of a software only business like Microsoft Office.
4. As an aside, when I first started tracking this data it was early 2015. The App store had just overtaken US Box Office revenues and the outrage at my comparison and the skepticism that this sustainable was impressive. Much of the outrage was due to my choice of the headline "Bigger Than Hollywood". You can still read the 72 comments. App store billings are now nineteen times bigger than US Box office billings. I still stand by the idea that the App Store is Bigger than Hollywood. It is in fact bigger than many other forms of entertainment combined--more about this later
5. Stay tuned for a detailed exposition of this audience

Executive exodus explicated

In December, the company lost AI chief John Giannandrea, its top lawyer and a key design executive in a single week. Other recent headlines: "Apple faces unprecedented executive exodus", "Apple rocked by executive departures, with chip chief at risk of leaving next." "Apple home devices executive leaves for Finland's Oura" "Lisa Jackson is set to retire late January 2026, while Kate Adams will retire in late 2026".

That's just the names we hear about. They represent senior executives. But the harsh truth is that Apple lost tens of thousands of talented developers, marketers and mangers last year and this year the numbers will be even higher. Is this indicative of an existential crisis in Cupertino which Tim Cook can't contain? Tim's head alone should be the first to roll to reverse this crisis in headcount.

Alas, math rears its ugly head.

Apple has over 166,000 employees including retail staff. According to one measure, Apple sees 58% of its staff turn over each year, having an average tenure of only 1.7 years, the worst in the tech world. That would imply 96,000 employees leave each year.

That seemed excessive to me and I assumed retail skews the total significantly. Retail, according to comments on Reddit, sees a few months of work per employee. So what is the turnover for its more "permanent" talent?

Let's assume that Apple can hold on to people as well as Alphabet/Google. We don't hear about the crisis of retention at Google. Their average tenure is 3.7 years, implying 27% would quit each year and indicate 49,000 Googlers leaving each year.

That rate applying to Apple would imply 45,000 Apple employees leaving each year. Having twice the tenure of Google would still mean 20,000 Apple employees quit each year.

A headline that screamed that between 20,000 and 30,000 Apple employees [quit, flee, abandon, escape] Apple each year would be perfectly reasonable and suggest a strong degree of loyalty. Moreover, next year the number will be higher because next year there will be more employees.

Of those tens of thousands, surely a few are senior managers. In fact, surveys show that 60% of executives plan to change roles in 3 years. The fact that we only hear of a handful of departures each year implies that Apple is far better at retention than average.

Nevertheless, the exodus continues.

Read more:

https://asymco.com/2026/03/17/tim-cook-squashes-retirement-rumors-says-he-cant-imagine-life-without-apple/

https://asymco.com/2026/03/12/how-many-employees-does-apple-have-what-do-they-do/

Remembering the PC market

Almost 10 years ago I said "Managing the Mac product line must be one of the most challenging problems at Apple. That may not be obvious given the product’s success." I pointed out that although the company grew the Mac into a redeemed and prosperous platform, "The big change in computing has not been a growing Mac vs. declining PC. It has been a huge surge in mobile device use vs. a decline in PC use overall."

It's time to revisit this observation.

First, a status update:

• The Mac is in its 32 42nd year of market presence. A longevity that is (still) unmatched by any other PC maker.
• Apple reached a top five four position in the ranking of PC vendors. This was achieved far along in the evolution of the market.
• With about $23 $33 billion in revenues per year, Apple places among the top four three PC vendors in terms of revenue.
• With an estimated $5.5 $13.8 billion in operating margin Apple is the most profitable PC vendor, capturing over 60% of the available PC hardware profits.
• The product has retained an average selling price of over $1200 ($1400 last year) for at least a decade. At the same time the average pricing of Personal Computers has more than halved.
• Volumes increased by about 20% over a decade.

The overall picture of the market is shown the following graphs. First, the overall quarterly shipments.

Note the Covid-era surge during 2020 and 2021 but overall decline from peak in 2012. Yearly shipments were 262 million in 2025 vs. 371 million in 2012. The install base of users is largely unchanged. Microsoft has cited "over 1 billion" to 1.4 billion monthly active devices, with figures varying between official statements. My estimate has been 1.2 billion PC users with an additional 250 million Mac users, more about this later.

The shipments data provided by Gartner over the last 17 years is shown below:

Notable are the declines in "other" vendors and the gains from Lenovo and Apple. Broadly speaking, "other" are assumed to be mostly consumer markets and that has either declined or been captured by Mac with consolidation around Lenovo. Lenovo, HP, Dell are likely to maintain enterprise accounts serving large customers and that has been a steady but low margin business for decades.

Mac has grown to fourth place but still a way to go to third place Dell. Apple may overtake "other" combined in a year or two.

The market share data shows how "other" has been squeezed between Lenovo and Apple while the other players have largely held steady. Acer and Asus somehow managed to survive.

So what does this mean for the future? What is Apple's strategy with the MacBook Neo. Or for that matter, for Studio and iMac and Mac Mini and MacBook Air and MacBook Pro?

The Neo was such a surprise for many given its incredible price/performance ratio and the overall quality and affordability (and distribution potential). I've read many reviews suggesting it will "disrupt" the market and that Windows PCs are just no longer competitive.

Will the MacBook Neo move the needle? Will it take share from the Windows world (or for that matter from the Chromebook and even iPad?)

I have no doubt the Neo will do very well indeed. It's an extraordinary product with immense appeal--I want one even if I am well served with Macs already. It will also be profitable. Even at the lower price point, the use of an A-series processor and memory and packaging which are very economical and iPhone manufacturing knowhow and scale means that it will probably sell in the 10s of millions.

But it's not likely to be hundreds of millions. The problem ultimately is that the MacBook New is still a traditional computer and phones outsell them 10 to 1.

What is more important is that the MacBook Neo will increase the audience or the user base for MacOS and ultimately resonate inside the Apple ecosystem, building momentum for other devices, wearables, home and services.

Let's go back to that total user number. The PC market, with 262 million units per year serves about 1.4 to 1.5 billion users. Of these Apple has about 250 million. Now consider this: Tim Cook has repeatedly said when he talks about the Mac on the earnings calls that "the Mac install base reached another all-time high, with nearly half of the customers who purchased a Mac being new to the product".

If we go back in the last 17 years, if half of the customers each quarter were "new to the product" then 167 million users were "new to the product". Let's be conservative and say 160 million net new users. That would have to be added to an existing install base of about 100 million before 2008 (which is before the iPhone era).

That implies that Apple has about 260 million Mac users and Windows has a base of about 1.2 billion. Total is 1.46 billion and Apple's share is about 18%.

The real target for the MacBook Neo would therefore be to increase this base and share. I doubt that it will grow the overall user base, but moving users to the Mac is very plausible.

So here is a target for the Neo: Grow Mac user base to 300 million and 20% global PC user share.

Not a stretch perhaps but not a "disruption" either. The real disruption already happened. It's in the fact that there are more iPhone users than all PC users, with Windows and Mac combined.

One more thing: Apple's sales of about 24 million Macs for a base of 260 million users implies a life span for Macs of nearly 11 years. However, all the other PCs sold add up to 238 million for a user base of 1.2 billion. That means the life span for the non-Mac PC is only 5 years.

This is, of course, reflected in the price difference but with the MacBook Neo, the potential for a 10 year old life for a $500 laptop puts the Neo price at $50/yr or 14c/day. Based on data for 2025, the average selling price of a new Windows PC is estimated to be $600 to $900 for consumer models. At $700 for a life span of 5 years makes the PC cost $140/yr or 38c/day.

This means that the average PC is more than twice the cost of the MacBook Neo. Now that is a surprise.

An allergy to centralized computing

Based on Tweets by MilkRoadAI

The 3rd? most valuable company on Earth watched as its rivals lit $650 billion on fire and did nothing. The biggest cash bonfire in history, by far, eagerly fed by all the usual suspects but one.

It's either the dumbest or the most brilliant move in corporate history.

Apple used to be the biggest capex spender, mainly because it paid for most of the property plant and equipment in the factories that made its phones and computers. The so-called tooling or equipment spending plus its leased store renovations and some data centers were costs no other tech giants had and so Apple was an outlier. It was, as I wrote at the time, equivalent to buying a few aircraft carriers every year.

But that all changed with AI. Amazon is spending $200 billion this year on AI data centers. Google, $185 billion. Microsoft, $114 billion. Meta, $135 billion. Combined: $650 billion. [Not including OpenAI, Anthropic and SpaceX/XAI.] That is like buying the US Navy every year. And yet Apple's capital budget is still a modest $14 billion, oscillating with new hardware tooling cycles.

Apple is refusing to transfer its cash flow to Nvidia. Curiously, it believes that its cash flow belongs to its shareholders, not to Nvidia's.

The hyperscalers are now spending 94% of their operating cash flows on AI infrastructure. Amazon is projected to go negative free cash flow this year with as much as $28 billion in the red. Alphabet's free cash flow is expected to collapse 90% from $73 billion to $8 billion. These companies used to be the greatest cash machines ever built. Now they're borrowing money to keep the data center lights on.

The Big Five raised $121 billion in bonds in 2025 alone. Morgan Stanley projects $1.5 trillion in tech debt over the coming years. For the first time in history, hyperscalers hold more debt than cash. Perhaps this is why their P/E ratios slumped from mid thirties to mid twenties.

And what are they getting for that $650 billion? AI services generate roughly $35 billion in total revenue or 5% of what's being spent on infrastructure. There are dreams of more of course, but the business models of AI have yet to resonate, especially for consumers.

Now here is where Apple's bet becomes genius. AI models are commoditizing faster than anyone predicted. Software and hardware both have tendencies to commodify. Protections exist but they have to do with integration and distribution. DeepSeek built a model for $6 million that matches systems costing $100 million. Open source models now power 80% of startups seeking VC funding. The moat these companies are spending hundreds of billions to build is evaporating.

Apple understood this before anyone else. It didn't build its own AI model, it licensed Google's Gemini for about $1 billion a year. Why spend $100 billion building a factory when outsourcing costs a billion? And if a better model appears next year, Apple just switches vendors.

But Apple is not sitting still. It just dropped the M5 chip with a 16 core Neural Engine and Neural Accelerators built into every GPU core. It runs 70 billion parameter AI models locally, eventually even on your phone. The M5 delivers 4x the AI performance of the M4 and Apple doesn't need $200 billion in data centers.

Because Apple turned 2 billion devices into the data center.​ Every iPhone, Mac, iPad gets distributed AI at a scale no server farm can match. While its rivals burn cash, Apple is doing the opposite. $90.7 billion in stock buybacks last fiscal year.​ Its competitors? Combined buybacks collapsed 74% from their peak.​

Apple didn't miss the AI revolution. It just bet that the winners won't be the ones who build the infrastructure. They'll be the ones who own the customer and no one else on Earth owns the best customers.