Broadcom's boss called it a casino. Tomorrow, ASML shows us its hand.

When the man selling the chips says the AI trade looks like a casino, it would be silly not to listen.

On Monday in Singapore, Broadcom CEO Hock Tan suggested that "a lot of investors have no clue what they're investing in", setting "very extreme expectations" before discovering that the trade resembles gambling. His verdict on corporate adoption was even blunter: companies are burning through AI tokens without finding meaningful returns, like children spending their arcade money in ten minutes and learning nothing.

Then comes the awkward bit. Tan runs one of the businesses making a fortune from the spending he questions.

Broadcom's AI semiconductor revenue hit $16.7bn in its latest reported quarter, up 221% year on year and 54% sequentially. Total revenue rose 86% to $29.6bn, while free cash flow reached $13.7bn. For the current quarter, Broadcom expects AI semiconductor revenue of $21.7bn, up 236% year on year.

Its longer-term ambition is more striking still: AI revenue of $230bn in 2028, according to recent reporting on its outlook.

This is not a management team quietly preparing for the bubble to burst. It is a company enjoying extraordinary demand while warning that the economics of the wider ecosystem remain unproven.

So, which one is it: a casino, or the greatest infrastructure build-out of our time?

Q3 numbers from ASML tomorrow will provide useful insights for what to expect in 2027 and beyond.

Its machines are indispensable to the manufacture of leading-edge chips, giving ASML an unusually powerful position in the AI supply chain. Chipmakers can debate which model wins, but advanced semiconductor production still depends on equipment that is extraordinarily difficult to replicate.

The headline numbers should be strong. In July, ASML guided Q3 net sales of €11bn–€12bn and raised its 2026 FY sales outlook to €43bn–€45bn, with a gross margin of 54%–56%.

But those are not the numbers I would focus on. Net bookings and customer demand for 2027 and 2028 are critical to understanding the future of the AI infra trade, as new orders offer a more immediate indication of what customers are prepared to commit to next. That distinction matters when chipmakers are expanding capacity to meet demand they expect from AI infrastructure.

ASML said in July that order intake had been extremely strong in the first half of the year. It planned to increase 2027 capacity for low-NA EUV and DUV immersion systems by around 30% against 2026 levels, while investigating further increases for 2028. Management was responding to customer commitments and improving visibility into longer-term demand.

Although encouraging, the stakes have been raised, as capacity expansion is really just a bet that demand will persist. Customer returns are not guaranteed.

A strong order number, supported by a credible 2027 outlook, would suggest that customers are still committing capital rather than merely talking up AI. A disappointing number, or weaker commentary on future orders, would raise questions about whether the spending cycle is approaching a more difficult phase.

Neither outcome settles the bubble debate on its own. Semiconductor equipment orders are lumpy, delivery schedules can shift, and orders do not automatically translate into profitable AI applications. But bookings, capacity plans and customer commentary together offer a more useful signal than another round of bullish forecasts.

There is also ASML's valuation. Its standalone position in EUV does not make its shares immune to disappointment. If investors are pricing in years of uninterrupted growth, even excellent results can fall short of expectations.

Tan can sound sceptical while benefiting from the infrastructure build-out. Broadcom sells custom accelerators and networking equipment into that spending cycle; it does not need every enterprise AI experiment to succeed for its own revenue to grow. But even Broadcom's exceptional growth cannot tell us whether customers across the ecosystem will ultimately earn acceptable returns on their investment.

That is the distinction between demand for the picks and shovels and the economics of the gold rush.

Tan is playfully managing expectations without surrendering the growth narrative.

He acknowledges the bear case before investors can use it against him and distinguishes real demand for chips from the much less certain question of how customers will monetise AI. Tan does all of this while reporting extraordinary growth and projecting an even larger business ahead.

The lesson is not simply to sound cautious. It is to be precise about what the business can prove, what remains uncertain and which indicators support management's confidence. This level of candour only really works from a position of strength and when all of the numbers are going in the right direction.

Tomorrow, ASML gives investors another piece of that evidence.

A casino can run hot for years. But the test is whether customers keep coming back.

Disclosure: I am a long-term holder of ASML and other semiconductor shares.

Bellway delivered 9,695 completions in FY2026, up 10.8%, while housing revenue rose more than 13% to £3.14bn. Adjusted operating profit is expected to be around £320m, but the margin fell to approximately 10% from 10.9%, reflecting a higher proportion of lower-margin bulk sales. The forward order book declined to 4,206 homes from 5,307 a year earlier.

The tension is familiar: volume growth is supporting the top line, but lower-margin sales and a smaller forward pipeline complicate the recovery story. For investors, the next question is whether private demand can strengthen enough to rebuild the order book without relying so heavily on bulk transactions. The announced £50m buyback adds a capital-allocation angle, but it does not resolve the underlying demand question.

This morning, Brent crude was trading around $105 a barrel after reaching $106+ on Monday, with geopolitical supply risks competing with the prospect of additional Russian diesel exports. The direction matters beyond energy stocks: sustained oil-price pressure could feed inflation, complicate central-bank decisions and raise operating costs for energy-intensive infrastructure.

For the AI investment thesis, energy is not a footnote. Power availability and electricity costs affect data-centre economics. Higher oil prices do not mechanically determine electricity prices, but a broader energy shock could add another cost and financing headwind to an already capital-intensive build-out.

This newsletter is published by James Lea / CorpCast for general information, educational and journalistic purposes only. Nothing in it constitutes financial, investment, legal or tax advice, or a personal recommendation, and it is not an offer, invitation or inducement to buy, sell or hold any security, cryptocurrency or other financial instrument.

Any views, opinions, forecasts or interpretations expressed are the author's own personal opinions at the time of publication, do not reflect the views of any business or entity the author is affiliated with, and may change without notice. No account is taken of any reader's personal circumstances, objectives or financial situation, and nothing here should be relied upon as the basis for any investment decision.

The author may from time to time hold long or short positions in securities, cryptocurrencies or other assets referenced in this newsletter, and may transact in them at any time without further notice. This is a standing disclosure covering all issues and is not repeated or itemised per asset.

Investing and trading involve substantial risk of loss, including the potential loss of your entire investment. Past performance is not a reliable indicator of future results. Figures, data and third-party information are believed reliable but are not guaranteed to be accurate, complete or up to date.

James Lea / CorpCast is not authorised or regulated by the Financial Conduct Authority. Always conduct your own independent research and consult a suitably qualified, FCA-regulated financial adviser before making any investment decision.