Good morning.

Before we get into it, if you’re out in LA right now, I hope you and your families are staying safe. You can find some fire department resources here.

Today, we’re getting into the money behind AI. The numbers for 2024 are astronomical, whether you’re looking at capital expenditures, stock surges or VC investments.

And all the trends are seemingly angled up for 2025.

— Ian Krietzberg, Editor-in-Chief, The Deep View

Overcoming Challenges in AI Adoption for IT and Security Leaders

AI adoption in the enterprise can feel like an uphill battle, as conflicting priorities, unclear ROI, evolving regulations, and security risks add layers of complexity.

That’s why Tines wrote Securing AI in the Enterprise: A step-by-step guide for IT and security leaders. This guide is designed to help IT and security leaders take a proactive, security-first approach to AI adoption, with practical strategies to address some of the biggest challenges.

From AI skeptics to believers to enthusiasts, you’ll get tactical advice on what safe and successful AI adoption looks like – and how your security team can accomplish it.

Get the Securing AI in the Enterprise guide today to start making real progress toward achieving your AI adoption goals.

The 2024 tech surge

A consistent undercurrent to the AI story — riding somewhere beneath regular announcements of new models and new capabilities and sci-fi visions of a fully autonomous future — involves a vast investment landscape. 

These models are costly to the point of being cost-prohibitive, for almost anyone who lacks pockets the size of those wielded by Big Tech and venture capitalists around the world. 

And surrounding that landscape is a combination of blind excitement, tempered excitement, caution and blistering criticism. 

There are two sides to this story. The first involves the public market and Wall Street — the ways in which AI-exposed stocks have performed, and the impression of the analysts that study them. 

And despite rising costs, surging capital expenditures and little revenue to show for it, the AI names have spent the past year riding a wave of investor excitement to truly record-breaking heights. 

The AI Boom 

In 2024, as with 2023, powerful, ceaseless stock surges from the tech sector continued to lift markets ever higher. The tech-heavy Nasdaq finished the year up 29%; the S&P 500 was up 23% (after a 24% surge the year before) and the Dow Jones Industrial Average was up around 13%. This push was led by a handful of tech stocks, most of which have clear exposure to artificial intelligence. 

Here, I’m talking about Nvidia, Microsoft, Meta, Google, Amazon, TSMC and Tesla.

(Tesla is a bit of a weird case — its late-year surge seems tied to President-elect Donald Trump’s win in November, more than the company’s AI or electric vehicle business, but still, the stock spiked more than 60% in ‘24, following a year in which it spiked more than 100%). 

A look back at related stories we’ve done in the past year: 

• Interview: The mismatch of the AI bubble

• The math behind the AI bubble

• Goldman Sachs publishes blistering report on the AI bubble

How high did they fly?

Nvidia, the maker of those (insanely) expensive “picks and shovels” for the AI race — the GPU chips vital to the training and operation of AI models — surged some 171% in 2024 (following a 240% spike the year before). On Dec. 30, 2023, Nvidia’s market cap was $1.22 trillion; a year later, the company was valued at around $3.5 trillion. 

• Shares of Microsoft, following a late-December retreat, spiked some 13% as its valuation soared from $2.79 trillion last year to $3.1 trillion this year. Shares of Amazon spiked 50%; its market cap jumped from $1.5 trillion at the end of 2023 to $2.4 trillion by the end of 2024. Shares of Google spiked roughly 37%; its market cap jumped from around $1.6 trillion to $2.3 trillion. 

• Meta, meanwhile, jumped around 70% in 2024; at the end of 2023, it was valued at around $900 billion. It is now valued at $1.5 trillion. And in the span of a year, TSMC went from a $500 billion stock to a trillion-dollar stock, surging nearly 100%. 

Though this isn’t comprehensive, just this small handful of companies added hundreds of billions of dollars to their market caps in the past 12 months. 

But the stock surge is almost nothing compared to the capital expenditures this handful of companies has notched. 

But at what cost?

Just Microsoft, Amazon, Meta and Google were expected to collectively spend around $240 billion in capex — all largely going to AI — by the end of 2024, a significant increase over last year’s numbers, according to Forbes. All of these firms have spent the past year assuring investors that their capital expenditures will increase in 2025 as they work to build out AI infrastructure. 

And, while some have been quick to suggest that they are starting to see a bit of revenue from all of these massive expenditures, many of the big players here have made clear that the AI game is a long one. 

Microsoft CFO Amy Hood said in August that its current investments in AI infrastructure will support the monetization of AI “over the next 15 years and beyond.” Microsoft plans to spend some $80 billion on AI infrastructure in 2025 alone.

Meta CFO Susan Li similarly expects “returns from generative AI to come in over a longer period of time,” adding: “we don’t expect our GenAI products to be a meaningful driver of revenue in ’24. But we do expect that they’re going to open up new revenue opportunities over time that will enable us to generate a solid return off of our investment.”

And in the midst of this, some sectors of Wall Street have been losing patience.

D.A. Davidson analyst Gil Luria told CNN in August: “if you’re going to invest now and get returns in 10 to 15 years, that’s a venture investment, that’s not a public company investment. For public companies, we expect to get return on investment in much shorter time frames. So that’s causing discomfort, because we’re not seeing the types of applications and revenue from applications that we would need to justify anywhere near these investments right now.”

A look back at related stories we’ve done in the past year: 

• AI companies need a lot of money. Investors are still forking it over

• OpenAI, others dealing with ‘diminishing returns’ in current architecture

• Competition and power consolidation in AI

At around the same time, Goldman Sachs released a report titled: “GenAI: Too much spend, too little benefit?” In the report, Jim Covello, Goldman’s head of Global Equity Research, said that, considering the $1 trillion (+) the industry is expected to spend on AI in the next few years, it remains unclear what trillion-dollar problem AI will solve. 

“Replacing low-wage jobs with tremendously costly technology is basically the polar opposite of the prior technology transitions I’ve witnessed in my thirty years of closely following the tech industry,” he said, adding that, where the internet provided a low-cost solution to a high-cost problem, GenAI offers a high-cost solution to a low-cost problem. 

That high cost has thus far been eaten by the developers as a means of bringing people on board. And that brings us to the startups.

• OpenAI, still the dominant developer in the generative AI race, reportedly lost $5 billion in 2024. And that’s after it raked in $3.7 billion in revenue. And it expects these losses to keep on mounting. 

• The Information reported in October that, according to internal documents, OpenAI expects to spend some $200 billion between 2023 and 2030; more than two-thirds of that amount is expected to go toward the training and deployment of its generative AI models. The startup expects to lose $44 billion between 2023 and 2028, saying that it expects to turn a $14 billion profit in 2029. 

Last year, it raised $6.6 billion in funding at a $157 billion valuation; at the same time, it secured a $4 billion revolving line of credit. It’s not just OpenAI, either; xAI, through a Series B and C funding round both conducted in 2024, raised a total of around $12 billion for the year and Anthropic raked in an additional $4 billion from Amazon. The costs are almost incomprehensible, but a combination of Big Tech backers and venture capitalists are still interested in footing the bill. 

And that brings us to the private markets and the venture capitalists, and it tells a story that is both different and the same from the trends on Wall Street. 

In 2023, according to Crunchbase data, the AI sector raised $55.5 billion.

In 2024, the sector raised $99.6 billion, an 80% year-over-year increase. Around a third of that amount went to foundation model companies (largely, in billion-dollar raises). The remaining amount went to sectors seeking to drive an AI impact.

The data here is pretty clear; VCs don’t care about high expenditures, low revenue and uncertain time horizons. They — even the skeptical ones — believe in artificial intelligence. 

Sequoia’s David Cahn wrote last year that there was — at the time — a $600 billion gap in the AI industry, meaning simply that, considering the magnitude of the cost of development, the sector would need to bring in $600 billion in revenue — annually — to just meet the scope of cost. 

Still, he believes in the potential of AI. “In reality, the road ahead is going to be a long one. It will have ups and downs. But almost certainly it will be worthwhile.”

As Pitchbook wrote in an August report: “Valuations have ramped up, and despite high prices, VCs have consistently been willing to invest.”

So, Despite Goldman’s skepticism — and despite Wall Street’s feelings toward the high costs and long time horizon of generative AI — the stocks have inched ever higher, enterprise adoption of the technology is certainly growing, the costs of development are spiking, the private valuations are increasing and the VCs remain undeterred, all of which makes for a rather complicated picture.

But, with adoption rates starting to increase, some investors are feeling good about 2025. 

Wedbush’s Dan Ives, for one, is super excited about what 2025 might hold for tech and AI. He expects tech stocks to soar another 25% in 2025 “based on underlying strength from the AI Revolution and the 2nd/3rd derivatives of this tidal wave of tech spending and AI cap-ex.”

• Ives wrote that the $4 trillion market cap club will get “its first three members. We believe the first member of the $4 trillion market cap club will be Apple as the iPhone 16 brings the AI Revolution to Cupertino. In 2025 we also firmly expect both Nvidia and Microsoft to follow Apple in joining the exclusive $4 trillion market cap club as the AI Revolution takes hold into its next gear of growth.”

• “We expect many enterprises and governments to aggressively head down the AI path with our view that AI cap-ex budgets will exceed $1 trillion and catalyze the next wave of use case and software buildouts across the consumer and enterprise landscape.”

JPMorgan wrote in November that it expects 2025 to see a broadening out of the market, rather than a bursting of the bubble. 

• “The valuation gap between the biggest stocks (the megacaps) and the rest is unlikely to persist indefinitely. If the broad AI ecosystem generates sufficient revenues to justify the earnings expectations already assumed for a handful of companies, the ‘rest’ should catch up over time,” according to JPM. “If instead, the broader corporate universe does not see the clear use case of these technologies and are unwilling to pay for them, then a ‘catch down’ scenario is more likely.”

• “The strong fundamentals of the megacaps, both relative to other parts of the S&P 500 today, as well as relative to the 2000 tech bubble, provide some comfort that a major ‘catch down’ is unlikely,” according to JMP.

Still, the firm noted that the gap between revenue expectations, capital expenditures and actual revenue growth will come into sharper focus next year, with investors saying: “show me the money.” 

“If the developers and the integrators can’t generate sufficient profit, this weakness will eventually spread up the value chain,” JPM wrote. “In our view the gap between the valuation of mega-cap tech and the broader S&P 500 is unsustainable. However, unlike 2000, we see a ‘catch up’ scenario as more likely than a ‘catch down’ scenario.”

Sequoia expects the question of investment returns on AI spend to remain a problem in 2025, but also expects Big Tech capital expenditures to begin to stabilize. 

• “Going into 2024, the Big Tech companies were nervous about AI being a threat to their oligopoly in the cloud business. Entering 2025, the picture has changed dramatically,” Sequoia wrote. “Big Tech companies have their arms locked firmly around the AI revolution. Not only do they control the vast majority of the data centers that power AI, but they own significant equity stakes in the big model companies, and they are among the largest backers of new AI startups.”

• “With Big Tech feeling more confident, we think 2025 will be a stabilization year for AI CapEx. If 2024 was a scramble to sign deals for land and power, 2025 will be an execution year. Shovels are in the ground, and these companies will be focused on completing their new projects on-time and on-budget. They will then need to sell this installed capacity to customers and work with enterprises to help them achieve success with their new AI capabilities.”

A look back at related stories we’ve done in the past year: 

• 2 years after ChatGPT, how VCs are approaching investing in

• OpenAI’s long, costly road to profitability

• Health tech VC talks AI: If you buy into the hype, you’ll lose

Deepwater Asset Management believes, as it has said in the past, that we are in the second year of a 3-5 year AI-driven bull market “that will end in the bursting of a bubble.” Deepwater expects the Nasdaq to experience two 10% pullbacks in 2025, though it will still finish the year higher than it started.

Gené Teare, a senior data editor at Crunchbase, expects “funding proportions to AI to increase in 2025. While foundation model companies might not raise as much as they did in 2024, AI infrastructure and applications will continue to see funding increase across multiple sectors." 

OK, let’s talk about reversion to the mean. 

It’s a financial theory that the price of an asset will eventually revert to its long-term average value; nothing rises forever, nothing falls forever. The hard part here is obviously figuring out the timing of such reversions. 

If you look at historical sector weightings of the S&P 500, tech has been on the rise in recent years. In 2024 it came in at 32.9% and in 2023 it was roughly 28.1%; things have been high and climbing since 2020 or so. 

Before the 2020 cycle, the last time the weighting of the tech sector exceeded 28% was in 1999, during the peak of the dot-com bubble. 

The average weighting of the tech sector of the S&P 500 — according to data from Bespoke Investing Group — between 1990 and 2023 is 17.5%. According to a slightly different dataset, between 2002 and 2023, the average weighting is 16.68%. 

So, let’s split the difference and say that the average weighting of the S&P’s tech sector is 17%. It is currently nearly double that average. I see no reason why mean reversion will not kick in at some point here — this is what JPMorgan was talking about with their catch-up or catch-down scenarios. One of two things will happen; either the rest of the stocks will catch up and things will level out, or the tech stocks will fall to meet their non-tech peers. I think we’ll see a bit of both, and I think the catch-down might well begin in 2025 as investor impatience persists.

At the same time, AI has been consuming an enormous amount of venture funding, a reality that just won’t continue forever. And, according to the VCs I’ve connected with, investing in AI has grown more challenging. Excitement might be far from dampened, but the landscape is becoming more expensive and more complex to invest in; I think the VC trends will start to falter next year, as well. 

We are pretty clearly in a bubble. Bubbles do two things; they expand, and they expand, and they expand … and then they burst. Again, nothing goes up forever. If the bubble swells another 20% next year, it would be approaching a danger point. 

I don’t know what the burst will look like — the dot-com burst didn’t kill the internet, it just killed off a bunch of internet companies — but I am sure that it is coming, and probably coming soon. 

To make this trackable, I’ll say we’ll see the beginnings of a burst by the end of 2025, marking 2026 as the year the AI bubble comes apart. A lot of companies will fall, leaving the Big Tech giants in a clear field. When this happens, Nvidia will have a very tough go of things.

🤔 Your thought process:

Selected Image 1 (Left):

• “The fake one doesn't have supports/wires for the lifts and has a building with a tree growing on top of it.”

• Sam Altman’s younger sister Annie has filed a lawsuit against her tech magnate brother, claiming that he sexually abused her for years. Sam denied the allegations in a statement shared on social media.

• Apple Intelligence, Bloomberg’s Dave Lee argues, isn’t that intelligent. This comes in the wake of mounting instances of hallucinations combined with excessive storage requirements.

• What’s next for our privacy? (MIT Tech Review).

• It is now illegal to use AI to deny a medically necessary health insurance claim in CA (Latin Times).

• Microsoft confirms performance-based job cuts across departments (CNBC).

• Los Angeles wildfires come on heels of hottest year on record: Images reveal global devastation (Semafor).

• Microsoft is reverting its Bing AI image generator because of quality complaints (The Verge).

If you want to get in front of an audience of 200,000+ developers, business leaders and tech enthusiasts, get in touch with us here.

💭 A poll before you go

Thanks for reading today’s edition of The Deep View!

We’ll see you in the next one.

Here’s your view on AI progress:

45% are in the middle; 16% think robotaxis will become normal in 2025 and 18% think OpenAI will achieve a (meaningless) AGI in 2025.

Middle:

• “If you follow or look into cognition-based science, you can see there is an extreme gulf between human and machine. LLMs and the present range of AI are nowhere near approaching the capabilities of the human mind. If you remember back to the AI shown in 2001: A Space Odyssey, you would have seen/heard the good doctor teaching the computer. Until we are at a point we can teach the AI, like we teach children, I suspect we have little to fear.”