Is the Market Headed for Doomsday? Why We See AI as a Supercycle, Not a Bubble

Headlines are asking whether markets are heading for disaster. With company profits growing fast & record levels of business already signed, we explain why we see AI as a long-term supercycle rather than a bubble, and what investors should keep an eye on.

Michele Carby·2 November 2026
ai supercycle

$1.7tn

signed cloud business at Amazon, Microsoft and Alphabet

$720bn+

planned Big Tech AI spending in 2026 ($720–745bn)

19x

price paid for each $1 of expected profit, vs 25x in 2000

+103%

Google Cloud sales growth in the latest quarter

Key takeaways

  • A supercycle, not a bubble. AI is a long-lasting change in how businesses work, and it is still early.

  • Today is not 2000. Share prices are backed by real profits and signed contracts, not promises.

  • The risks are real but contained. Excess sits mainly in smaller companies that rely on borrowing.

  • We invest in what AI runs on: chips, data centres, power and the grid.

Is AI a Bubble? Our View

October has tested investors' nerves, and headlines asking whether markets are heading for disaster have multiplied. On 22 October the Nasdaq had its worst day since June. China's STAR 50 index of AI and chip companies has fallen 40% in under four months. Alphabet, Google's parent company, saw its shares suffer their worst day in more than 20 years, even though it reported record growth in its cloud business, because it also announced heavier spending (CNBC, 29 October 2026). Some well-known market strategists have warned that AI shares could fall further than at any time since 2008.

So it is natural that investors are asking two questions: are we in an AI bubble, and are markets heading for a bigger fall? The Redstone Capital Investment Desk does not believe either is the case. In our view, AI is not a bubble but a supercycle: a deep, long-lasting change in how businesses work, make decisions and grow, which is still in its early stages.

That view is based on evidence, not mood. In this article, we compare today's market with the major crashes of the past 25 years, ask whether company profits support today's share prices, follow the money behind the AI build-out and set out the warning signs that would make us change our mind.

“Volatility should not be mistaken as fragility. What we are observing is a market recalibrating its expectations, not one losing its foundations. Our role is to distinguish between the two.”

Michele Carby, Chief Executive Officer and Founder, Redstone Capital

Why This Time Is Different: Today's Market Against Past Crashes

Each of the four big market falls of the past 25 years had its own cause. Comparing them with today is a useful way to check whether history is repeating itself.

Episode

How far markets fell

What caused it

Does it apply today?

2000 dot-com bubble

Nasdaq −78% (Mar 2000–Oct 2002)

Share prices ran far ahead of company sales; the Nasdaq-100 cost about 60 times expected profits

Not much. Today's leaders make record profits, and their shares cost far less relative to profits than in 2000.

2008 global financial crisis

S&P 500 −57% (Oct 2007–Mar 2009)

Too much borrowing in housing and the banks

Not much. The AI build-out is led by the world's most cash-rich companies, not heavily indebted banks.

2020 pandemic

S&P 500 −34% in 33 days (Feb–Mar 2020)

A sudden outside shock that brought the economy to a halt

Low. Demand for AI computing power is still greater than supply.

2022 inflation shock

S&P 500 −25%, Nasdaq −33% (Jan–Oct 2022)

Inflation at 40-year highs and the fastest US interest-rate rises in decades

Partly. Interest rates still matter, but they are no longer rising at that pace.

Source: S&P Dow Jones Indices and Nasdaq historical index data; J.P. Morgan. Declines are approximate, peak to trough. Past performance is not a reliable indicator of future results.

Looking across these episodes, markets did not crash simply because people got too excited about a new technology. The big falls came when share prices lost touch with the cash companies were actually making, when too much borrowing built up in weak parts of the financial system, or when a sudden shock stopped the economy. In our view, none of these conditions is clearly present today.

In short

The big crashes came from prices losing touch with profits, too much borrowing or a sudden shock. We do not see any of these clearly today.

Are AI Share Prices Justified? Profits, Not Hype

Share prices are high, and it would be wrong to pretend otherwise. One long-term measure, the CAPE ratio, compares share prices with the average profits companies made over the past ten years, adjusted for inflation. It stands at about 41, against a long-run average of 17.5 and a 1999 record of 44.2 (Robert Shiller data; MAS Economics, October 2026). A second measure compares share prices with the profits analysts expect over the next 12 months. On that basis, the S&P 500 cost 19 times expected profits on 30 September 2026. That is above its 30-year average of 17.2 times, but well below the 25.2 times reached at the peak in March 2000 (J.P. Morgan, Guide to the Markets).

The gap between the two measures tells its own story. The first looks back over ten years, mostly before AI was earning real money. The second looks at profits that are arriving now. The latest set of company results shows how quickly that is happening at the companies leading the build-out.

Company

Period

Highlights

Microsoft

Q1 FY27 (Jul–Sep 2026)

Azure, its cloud business, up 48% to $33.2bn; total sales up 19% to $92.3bn; signed future business up 73% to $677bn

Alphabet

Q3 2026 (Jul–Sep)

Google Cloud up 103% to $30.7bn, with a 36% profit margin; $561bn of signed future cloud business

Amazon

Q3 2026 (Jul–Sep)

AWS, its cloud business, up 39% to $46bn; $578bn of signed future business; record 13.9% profit margin

Nvidia

Q2 FY27 (May–Jul 2026)

Sales up 106% to $96.2bn; profit of $59.7bn; expects sales of $108bn next quarter

Source: company results and earnings releases, 26 August to 29 October 2026. Periods are company fiscal quarters.

The figure we watch most closely is the backlog: business that customers have already signed up and committed to pay for, but not yet used. Alphabet's cloud backlog grew from $514 billion to $561 billion in a single quarter, and the company expects just over half of it to turn into sales within two years. That points to roughly $300 billion of signed revenue over that period. Across Amazon, Microsoft and Alphabet together, the backlog was close to $1.7 trillion before this week's results (Reuters, 27 October 2026). In 2000, telecoms companies laid cables for customers they hoped would come. Today, capacity is being built for customers who have already signed contracts.

The dot-com era also holds a more encouraging lesson than many people remember. Plenty of companies did not survive the 2000 crash, yet the internet went on to transform shopping, media and banking, and to create some of the most valuable businesses in history. The crash weeded out weak, speculative companies without stopping the technology itself. We see clear similarities today. As in the late 1990s, companies are spending enormous sums, and not all of that money will be well spent. But the business foundations underneath are much stronger, and we believe AI is only at the start of being widely used: a turning point, not a peak.

The financial benefits of AI are also arriving faster than they did for the early internet. AI helps people get more done, cuts running costs and saves time by speeding up decisions and product development. As AI "agents" (software that can plan and complete multi-step tasks with little human help) become part of everyday work in purchasing, finance and customer service, AI moves from being a tool people use to being the foundation that businesses run on.

In short

Shares are not cheap, but this time they are backed by fast-growing profits and hundreds of billions of dollars of signed contracts.

“In 2000, investors were paying for promises. Today, they are paying for profits and signed contracts. That difference is at the heart of our conviction.”

Payal Trehan, Chief Investment Officer and Founder, Redstone Capital

Following the Money: Who Is Paying for the AI Build-Out

The amount being invested has never been seen before. The four largest cloud and technology companies expect to spend about $720–745 billion in 2026 on data centres, chips and equipment, up from around $410 billion in 2025 (company guidance, July 2026). Amazon alone spent $77.3 billion in the third quarter. For investors, though, the more useful question is not how much is being spent, but where the money is coming from and what it is buying.

Company

Planned 2026 spending

Spending as a share of the cash the business brings in (Q2 2026)

Signed future cloud business

Amazon

c.$220bn

119%

$578bn

Alphabet

$195–205bn

115%

$561bn

Microsoft

c.$175bn

65%

$677bn

Meta

$135–145bn

95%

No cloud business selling to other companies

Sources: company guidance and Q3 2026 results; NextMSC, using Q2 2026 company data. Microsoft's guidance reflects a change in lease accounting. Backlog figures follow each company's own definition and are not directly comparable.

Three points stand out:

  • First, not every company is paying for this in the same way. Microsoft is comfortably covering its spending from the cash its business brings in, while Amazon and Alphabet are currently spending more than that.

  • Second, the companies spending the most also have the biggest backlogs of signed business, so much of today's investment is tied to real customer contracts rather than guesswork.

  • Third, Meta, which has no cloud business to sell its computing power to other companies, has faced the toughest questions and on 26 October launched a $22.5 billion sale of new shares and convertible bonds to raise money. Investors are telling apart spending that is clearly backed by customers from spending that is not. That is what we would expect in a healthy investment cycle, not a mania.

In short

The money is real and mostly self-funded. Investors are rewarding spending backed by customers and questioning spending that is not.

The Case for Caution: The Main Risks

There are fair concerns about how fast AI investment is growing and how it is being paid for, and we take them seriously. Independent research firm IoT Analytics breaks the debate into four simple questions: Is AI good enough to do work people will pay for? Are businesses making money from it? Will that money pay back the cost of building the data centres? And have share prices run ahead of all three? Its research suggests the biggest companies' AI investment currently earns more than a 15% return, but could fall below that from 2028. It describes that as a loss the largest companies can absorb, not a threat to the wider financial system (IoT Analytics, October 2026).

Concern

What the evidence shows

Our view

Spending is running ahead of cash

Alphabet and Amazon spent more on investment than their businesses brought in during the second quarter. Alphabet's free cash flow, the cash left over after investment, was negative $9.1bn in the third quarter.

A real pressure point, eased by record backlogs and very strong finances. We watch closely how quickly signed business turns into sales.

Stress in AI borrowing

Some debt issued by smaller AI data centre companies has traded at yields of around 15%, a sign that lenders see higher risk (Financial Times, 16 October 2026).

The clearest area of excess. It calls for being selective, not for walking away from the theme.

A handful of companies driving the market

J.P. Morgan estimates that 42 AI-linked companies produced around two-thirds of the S&P 500's gains since early 2024.

This can make markets bumpier, but it is less worrying when the biggest companies are also the most profitable.

Falls in other markets

China's STAR 50 index fell 40% in under four months, wiping out around $480bn of value.

The fall was steepest where speculation by individual traders ran ahead of real profits, which supports choosing companies on their underlying strength.

Returns may disappoint

Independent research suggests returns on the biggest companies' AI investment could drop below a 15% target from 2028.

Possible for parts of the build-out. Cash-rich companies could absorb those losses: returns would suffer, but the wider financial system would not be at risk.

Sources: NextMSC, using Q2 2026 company data; Alphabet Q3 2026 results, 29 October 2026; Financial Times, 16 October 2026; J.P. Morgan via NPR, 20 October 2026; Reuters, 26 October 2026; IoT Analytics, October 2026.

Overall, we see pockets of excess, as in every major technology shift, rather than one big bubble. There will be setbacks, and some business models will not last, but the infrastructure and profits behind the AI theme are far more durable.

In short

There are pockets of excess, mainly among smaller, debt-funded companies, but not a bubble across the whole market.

What Would Change Our View

A conviction is only worth holding if it is tested against the facts. These are the signs we watch most closely, and where each one stands today.

What we watch

Why it matters

Where it stands now

Cloud sales growth

Shows whether interest in AI is turning into paying customers

Speeding up: Azure +48%, Google Cloud +103%, AWS +39% in the latest quarter

Growth in signed future business

Shows how much future business is already under contract

Rising: Google Cloud's grew from $514bn to $561bn in one quarter

Spending compared with cash coming in

Shows how much of the build-out companies can pay for themselves

Above 100% at Amazon and Alphabet; 65% at Microsoft

Borrowing conditions for AI infrastructure

Companies that rely on debt are where problems would show up first

Pressure limited to smaller companies

Access to electricity

Power and grid connections now limit how fast new data centres can open

Grid connections remain a bottleneck in key markets

Sources: company Q3 2026 results; NextMSC; Financial Times, 16 October 2026; IEA, April 2026.

We would rethink our view if cloud growth slowed for a sustained period, if backlogs started to shrink, or if borrowing problems spread from smaller companies to the largest ones. None of these is happening today.

Where We See Lasting Value: Chips, Data Centres, Power and the Grid

01

Chips

The engine of AI. Nvidia's data centre sales rose 117% in a year.

02

Data centres

Still in short supply, with signed demand into 2028 and beyond.

03

Power

Data centre electricity use is expected to roughly double by 2030.

04

The grid

Getting connected is now a bottleneck, making grid assets valuable.

Rather than trying to guess which individual AI apps will win, we focus on the infrastructure that every AI platform needs. Computer chips remain the engine: Nvidia's data centre sales rose 117% on the year before to $89 billion in the three months to July 2026, and it kept 75 cents of every dollar of sales after production costs (Nvidia, 26 August 2026). Data centre space is still in short supply, with Amazon and Alphabet building for signed demand that stretches into 2028 and beyond (Q3 2026 results).

Electricity and the power grid are less visible but just as important. Data centres used 17% more electricity in 2025, growing more than five times faster than the world's overall electricity use, and the International Energy Agency expects their demand to roughly double by 2030 (IEA, April 2026). Getting connected to the grid has become a major bottleneck for new data centres, which makes power lines, energy storage and power generation valuable in their own right. These are the areas where demand for AI turns into real, contracted and often regulated income, whichever AI platforms end up leading.

Investing in AI for the Long Term

We do not expect a smooth ride, and there will be ups and downs along the way. Big new technologies rarely move in a straight line. Sharp falls, disappointing results and worrying headlines have been part of every major investment cycle.

History also shows how costly it can be to sell just because shares look expensive. Shares already looked pricey in the mid-1990s, yet an investor who sold for that reason in June 1997 would have missed a further rise of about 70% in US shares before the 2000 peak (MAS Economics, October 2026). For long-term investors, the real question is not whether AI shares will have more setbacks (they almost certainly will), but whether the companies behind them will be bigger, more profitable and more essential ten years from now. On today's evidence, we expect they will, although nothing can be guaranteed.

“We do not invest for the swings of a single quarter. For long-term investors, the more important decision is to stay on the right side of this change, while being careful about which parts of the AI industry you invest in.”

Payal Trehan, Chief Investment Officer and Founder, Redstone Capital

Frequently Asked Questions

Are We in an AI Bubble, and Is It Bursting?

We do not believe so. A bubble forms when share prices lose touch with what companies can realistically earn, yet the leading AI companies are reporting record sales, rising profits and hundreds of billions of dollars of signed future business. October did bring sharp falls, including a drop of about 13% in Alphabet shares on 29 October, their worst day since 2004 (CNBC), and a 40% fall in China's STAR 50 index in under four months (Reuters, 26 October 2026). We see this as the market adjusting its expectations, not a collapse. There are pockets of speculation, mainly in some privately funded companies and those that rely heavily on borrowing, but the core of the market is earning its value.

How Is Today Different from the Dot-Com Crash?

In March 2000, the Nasdaq-100 cost about 60 times the profits its companies were expected to make, and many leading companies had little in the way of sales. Today it costs about 22 times expected profits, and the biggest AI companies are making record profits backed by signed contracts.

What Happens if AI Shares Fall Sharply?

It would depend on where the losses land. Speculative companies and those that rely heavily on borrowing would likely be hit hardest, while companies with real profits and strong finances have historically held up far better and often come out with a bigger share of their market. A fall in share prices would not mean AI had failed: a bubble is about the price paid for a share, not the value of the technology itself. The internet thrived after 2000, and we would expect AI use to keep growing through any downturn.

Should I Sell My AI Investments?

That depends on your goals, how long you plan to invest and the rest of your portfolio, and it should not be decided on headlines alone. This article is not personal advice, so please speak to your adviser, who can assess what is right for your circumstances.

Investment Desk Insight, Every Week

This article is based on the research behind our weekly client reports. Every Monday, Redstone Capital clients receive a report from our Investment Desk covering the week's market moves, company results, central bank decisions and world events, along with the Desk's view on what they could mean for long-term investors.

The aim is simple: to give clients a clear, calm perspective when markets are at their noisiest, and a direct line to the people forming those views. If you would like to receive the Investment Desk's research as a Redstone Capital client, please speak to one of our advisers.

Talk to us

If you would like to talk about how your portfolio is positioned for today's markets, we offer a free 30-minute consultation to review your finances and long-term investment plans.

→ Book a Discovery Call

Important Information

Marketing communication. This article is issued by Redstone Capital Ltd, which is regulated by the Dubai Financial Services Authority (DFSA). It reflects the views of the Redstone Capital Investment Desk as at 30 October 2026, which may change without notice. It is provided for general information only and does not constitute personal advice, investment research or an offer or solicitation to buy or sell any security. Before making any investment decision, please speak to your adviser, who will assess whether it is suitable for your circumstances.

Key risks. The value of investments, and any income from them, can fall as well as rise, and you may get back less than you invest. Past performance is not a reliable indicator of future results. Investments concentrated in a single sector or theme, such as technology or artificial intelligence, can be more volatile than diversified portfolios.

Forward-looking statements. Statements about future markets, earnings or economic developments are opinions and expectations, not guarantees, and actual outcomes may differ materially.

Third-party data and holdings. Figures are drawn from the named third-party sources as at the dates shown, which we believe to be reliable. Companies are mentioned for illustration only and this is not a recommendation to buy or sell them. Redstone Capital Ltd, its employees and its clients may hold positions in securities mentioned.

Distribution. This material is not directed at any person in any jurisdiction where its publication or availability would be contrary to local law or regulation.

Start The Conversation

Let's talk about your wealth

Managing your wealth well takes clarity, strategy and the right expertise. A 30-minute call with us is the simplest way to get started. Wherever you are in the world, we'll help you see your financial picture clearly and show you what a coordinated strategy could look like.

Book a Discovery Call

Clarity on your position

A clear view of where you stand across pensions, investments and cross-border assets.

Expert perspective

Honest insight from advisors who specialise in international and cross-border financial planning.

No obligation

Just a conversation. No pressure, no commitment, no cost.