Photo by Steve Johnson on Unsplash

A potential 400% growth of cloud spending by the world's most valuable AI startups may have single-handedly masked a critical fact about the tech market: possibly most of the cloud growth is driven by AI loss-making startups. This dependence creates a precarious foundation for the stock market, where the momentum of the Magnificent Seven now hinges on how long billions in subsidized funding can prop up the massive costs of the AI arms race.

And that subsidized growth for AI loss-making startups many times is funded by the very cloud companies themselves that are showing that accelerated revenue growth.

Said simply, much of the Magnificent Seven cloud companies' are driving their own revenue growth from their own meaningful investments in AI loss-making startups.

That's setting up a reality check that either AI startups turn a profit at some point or the cloud companies get called out on their own investment dollars driving their own revenue growth.

We're about to show that the revenue growth that these cloud companies are self-generating is driving a potentially meaningful portion of the overall revenue growth.

How long can that cycle last?

Arguments have been made that this time is different than the dot-com boom-bust. But we’re about to show that the main driver of the AI boom is similar to the dot-com boom-bust. Similar to then, today’s market is mostly driven by loss making startups with no near-term profit timetable to fund the current spending.

Therefore, similar to the dot-com bust, it may only be a matter of time before AI sees the same cliff before the final takeoff, similar to the past.

Our analysis concludes that a tiny fraction of the cloud market is so disproportionately aggressive in its spending that it’s creating the illusion of broad-based enterprise health. And within the analysis you'll see that the cloud companies investment dollars itself is a key driver for their own revenue growth.

1. The Startup Hyper-Acceleration: 400% Growth is Real, But Not Profitable

The 400% growth factor is a proxy for the sudden, massive increase in demand for GPUs and compute capacity. The main proof was primarily revealed in Microsoft’s recently reported quarter and the financial burn of the segment's leaders are probably similar:

• OpenAI's Cost Proxy: Microsoft's financial reports for Q1 FY26 confirm Azure's revenue surged 40% YoY, attributing the momentum directly to demand for AI services. The rising net loss absorption related to the OpenAI investment is a key indicator of the underlying costs. When comparing the ‘other expense’ line with the previous quarter it shows a cost jump of over 100% sequentially, validating the extreme acceleration. If over 100% growth is from one quarter, then four quarters could give us expectations of over 400% growth in OpenAI spend year over year.

• Anthropic's Hard Data: Competitor Anthropic's internal spending confirms this trend. Their cloud compute costs were calculated to have grown about 300% Year-over-Year (from late 2023 to late 2024), based on reported monthly spend figures. This one year growth validates the huge growth OpenAI showed as seen through the Microsoft’s financial statements.

For round numbers, taking OpenAI’s more than 400% annual spend growth and Anthropic’s roughly 300% annual spend growth, we are using 400% growth as a proxy.

But since the growth appears to accelerate, this estimate may not be aggressive.

2. The Core Calculation: 11.5% of Customers Drive 46% Cloud Growth

To understand the market's dependence, we must first establish the size of the loss-making AI segment relative to the entire public cloud.

Step A: Calculating Loss-Making AI Market Spend Share Of 11.5%

The roughly $90 Billion figure for the total annualized spend of loss-making AI is a derived from explicit contracts and loss proxies:

Compared to the widely cited industry analyst forecast of a $780 Billion total global public cloud market in 2025, AI loss making startups thus may account for roughly 11.5% of total cloud revenue ($90 B divided by $780 B). We don't need precision in these estimates because it's here to help us formulate an understanding of the true driver of the cloud market, the Magnificent Seven, and so the stock market.

Step B: Calculating the Growth Contribution From Loss-Making AI Startups Of 46%

This small base with massive acceleration creates a disproportionate contribution to total cloud revenue:

Loss-Making AI Contribution = Loss Making AI Market Share X (multiplied by) Loss-Making AI’s Growth In Spend.

46% = 11.5% X 400%

Conclusion: 46% of cloud revenue growth may have been generated by a single segment that is currently operating at a massive loss and its investors are the very cloud companies reporting this revenue growth.

3. The Shocking Implication: The Market's Growth is somewhat of an Illusion

The most alarming conclusion is revealed when we compare loss making AI startups 46% growth contribution to overall cloud growth.

The entire public cloud market grew by approximately 28% in the last reported quarter. If loss making AI startups accounted for anywhere more than the 28% (like our 46% number) then maybe all the cloud growth last quarter may have been accounted by loss making AI startups. And that revenue growth was manufactured to a meaningful extent by the cloud companies themselves investing in these AI loss-making startups.

Based on this math, the remaining 88.5% of the cloud business consisting of stable, profitable enterprises—the true barometer of broad tech adoption— may be growing at a negligible rate, maybe even contracting.

The record cloud figures for Microsoft Azure, AWS, and Google Cloud are not a sign of broad enterprise health but rather they are a direct financial reflection of the AI funding cycle.

4. The Systemic Risk: When Does the Music Stop?

The entire thesis of the Mag 7 dominating the stock market is now dependent on the longevity of this funding loop:

1. Big Tech injects billions in funding/credits into AI startups which drives their own cloud revenues.

2. Startups immediately convert that funding into compute spend, paying AWS and Azure driving cloud revenues.

3. Hyperscalers book the payment as cloud revenues and revenue growth, which drives their stock valuation but they never saw cash inflow from it. Just the opposite, if anything they had cash outflows to invest in these companies.

The Fatal Flaw: The loss-making AI startups generating this ~46% growth have mounting losses and no clear timeline to turn a profit. The cost of training and running the next generation of models continues to outpace revenue by a significant margin.

• Risk to the Mag 7: If venture capital funding tightens, or if a major model builder decides to prioritize efficiency over spending, the 400% growth from loss-making AI startups (which may be driving a disproportionate level of the overall cloud spend) can evaporate quickly. If any funding of the $90 Billion in AI spend were to take a pause, cloud growth rates would decelerate rapidly, potentially triggering a massive re-evaluation of the entire tech sector’s prospects.

• And if the investment community starts to call out these cloud companies for driving their own revenues, it can cause meaningful stock market risk.

The current stock market rally is fundamentally reliant on the sustained, $90 Billion loss of a small group of subsidized customers. Investors must ask: How long can Big Tech afford to pay its own customers to buy its products before the market forces a reckoning?

Arguments have been made that this time is different and AI has better underlying fundamentals than the dot-com boom-bust. But there are similarities as well where the entire market is excited based on accelerating cloud growth rates. But those cloud growth rates, we think, based on some of the math above, is based on a similar driver as the dot-com boom-bust; loss-making companies and self-funded revenue growth are driving it.

More By This Author:

Melt-Up Or Top? Market Indicators And 2 Confirmation Signals To Watch
AI Hype Vs. Dot-Com Reality: We Analyze Valuations, Growth, And The Looming Risk Of A Burst
China’s Rare Earth Gambit: Give Up Taiwan Or Lose Rare Earths