This Is Ledger
Briefing · Macro desk

The 2026 bull case rests on an AI capex bet that must clear

Strategists see the S&P at 7,000-7,300 next year. The tail risk is not inflation — it is whether the ROI on the AI build ever arrives.

L
By The Ledger Desk
AI synthesis · Published 2 Jul 2026 · 1 source at the time
Sources ↓
Forecast spectrum

12 named voices on the record

0%
50%
100%
HSBC
HSBC
HSBC
HSBC
Wall Street strategists
Wall Street strategists
Wall Street strategists
Bank of America
HSBC
Bank of America
Bank of America
Bank of America
HSBCmedium

Will OpenAI receive at least $207,000,000,000 in new equity funding by 2030?

Position: YES

caliber 70
HSBCmedium

Will OpenAI generate around $220,000,000,000 in revenue in 2030?

Position: YES

caliber 70
HSBCmedium

Will OpenAI record cumulative operating losses of at least $500,000,000,000 from the start of 2025 through 2030?

Position: YES

caliber 70
Wall Street strategistsmedium

Will the S&P 500 close 2026 at or above 7,300?

Position: YES

caliber 70
Bank of America57%

Do 57% of surveyed fund managers expect the S&P 500 to be above 7,000 in 2026?

Position: YES

caliber 65
Wall Street strategistsmedium

Will the S&P 500 close 2026 at or above 7300?

Position: YES

caliber 65
Wall Street strategistsmedium

Will the S&P 500 close 2026 at or above 7300?

Position: YES

caliber 60
Bank of Americamedium

Do a majority of surveyed fund managers expect the S&P 500 to close 2026 above 7000?

Position: YES

caliber 60
Bank of America57%

Do a majority of surveyed fund managers expect the S&P 500 to be above 7,000 at the end of 2026?

Position: YES

caliber 60
HSBCmedium

Will OpenAI receive at least $207,000,000,000 in new equity funding by 2030?

Position: YES

caliber 60
HSBClow

Will OpenAI report roughly $220 billion in revenue in calendar year 2030?

Position: YES

caliber 55
Bank of Americamedium

Do fund managers rank higher inflation / higher interest rates as the top risk to another good 2026 at 45%?

Position: YES

caliber 55
Key numbers

What anchors the cluster

OpenAI requires at least $207 billion in new equity funding until 2030, projects $220 billion revenue in 2030, $500 billion operating losses from 2025 to 2030, and $76.5 billion losses in 2030.

OpenAI faces $500 billion in operating losses from start of 2025 until 2030, with $76.5 billion in 2030.

OpenAI projected revenue reaches around $220 billion in 2030.

OpenAI requires at least $207 billion in new equity funding until 2030, when free cash flow may turn positive.

The consensus for 2026 is comfortable and crowded. Wall Street strategists cluster around 7,000 to 7,300 on the S&P 500, and the BofA fund manager survey shows the same base case

: another year of gains, powered by the same tech complex that carried 2024 and 2025. The interesting question is not whether the consensus is right. It is what happens if the capital cycle underwriting it stalls — because the index and the AI build have become the same trade.

Start with the risk hierarchy the buy-side has actually written down. According to the BofA fund manager survey, 45 percent name higher inflation and interest rates as the top risk for 2026, and 26 percent name an AI capex slowdown. That second number is the one worth staring at. A year ago, a slowdown in hyperscaler spending was not a mainstream tail. Now more than a quarter of managers treat it as the dominant scenario risk — ahead of unemployment above 5 percent (17 percent), a US-China trade rupture (9 percent), and the midterms (2 percent). The market has internalised that the index and the capex cycle are the same object.

From capital-light to asset-heavy

The structural point, made cleanly by Maverick Equity Research, is that TMT (technology, media and telecom) is no longer the capital-light margin machine of the 2010s. It is becoming asset-intensive — data centres, custom silicon, power contracts, cooling — and that shift will compress margins in the near term regardless of how the demand side resolves. This matters for index math. An S&P 500 whose earnings weight is concentrated in a handful of hyperscalers is now an index whose margin trajectory depends on depreciation schedules and utilisation rates, not just gross profit on software seats. The multiple the market pays for that cash flow should, in principle, be lower. It is not.

The sharpest stress test of that future is OpenAI, and HSBC's numbers are the ones to hold in view. The bank estimates OpenAI needs at least $207 billion in new equity funding through 2030, against cumulative operating losses of roughly $500 billion between 2025 and 2030, with a $76.5 billion loss in the terminal year even as revenue reaches around $220 billion. Free cash flow, on HSBC's arithmetic, turns positive only around 2030. That is a private company absorbing more equity over five years than the market capitalisation of most S&P 100 members — and it is the demand anchor for a meaningful share of the capex the index is being rerated on.

The index and the AI build have become the same trade. If one clears, both clear. If one stalls, both stall.

The Ledger Desk

The dossier is one-sided: every named forecaster in the cluster sits on the bullish side of 2026 at medium conviction, and HSBC's OpenAI numbers are presented as base-case rather than stress. That unanimity is itself the signal. Readers should treat the operationalisable claims accordingly — Wall Street strategists carrying the S&P above 7,300 by year-end 2026, HSBC's $207 billion OpenAI equity raise clearing by 2030, and OpenAI's $220 billion 2030 revenue print — as a linked chain, not independent bets. The 26 percent of managers flagging a capex slowdown are pricing the scenario in which the chain breaks at its weakest link. That is the trade worth watching, not the consensus target.

Briefings are synthesised by the Ledger Desk from multiple sources cited in the sidebar. They are distinct from Articles, which are written by named contributors and carry a tracked Calibration Index. The Desk does not currently carry a Brier score; this is a deliberate choice for the v0.1 editorial layer and will be revisited.

Source map

Where the material came from

  • Maverick Equity Research
Cited

Sources

6 articles