Tech is back in favor, but the tape still lacks a clear "big top" signal
The tape looks more constructive than it did a few months ago, but not yet euphoric. U.S. futures are higher on the 4 Ts - Tehran, Trade, Taiwan and Tariffs - and tech stocks are again the market's hottest trade. That helps explain why investors are being pulled back into the trend just as XLK is up 18% since March 30. For now, that reads more like momentum than a clean fundamental breakout.
This does not look like a classic "big top." Full-blown euphoria usually comes with extreme breadth, loose valuation discipline, and a more obvious rotation out of quality. What investors are seeing instead is a more selective risk-on backdrop, with the tech meltup continuing alongside some easing in geopolitical tension. That is not the same as saying chasing concentration here is a bad idea; it is just a less favorable risk-adjusted entry.
Kabra's hesitation comes from the cash-flow lag in AI capex
Kabra's point is not that the AI buildout is unreal. It is that XLK is up 18% since March 30 before the balance-sheet proof he wants has arrived. His timing call centers on a simple lag: capex is front-loaded, while cash generation tends to show up later.
Free cash flow is the trigger he is watching
Hyperscalers' free cash flow has been declining every quarter since early 2024, and Societe Generale expects aggregate hyperscaler free cash flow to turn positive back into the positive in early 2027. Until that inflection is visible, cap-weighted tech remains exposed to a market that is still paying for future monetization before the cash-flow recovery is confirmed.
Why equal-weight S&P 500 looks more attractive to him
Kabra's preference for equal-weight S&P 500 exposure is a portfolio-construction call, not a rejection of AI. If the biggest payoffs are still ahead, equal-weight offers broader U.S. growth exposure and less dependence on a narrow group of capex-heavy names clearing the monetization hurdle first.
What would signal a real topping pattern - and what would change the call
The practical question is not whether tech still has supporters. It is what evidence would change the allocation call.
Warning signs to monitor
A more convincing "big top" would likely show up in market internals before valuations fully break. In today's tape, that would mean the broad risk-on trade starts to outrun the AI core. A useful warning sign is when cyclical leadership and policy headlines drive gains while concentration in the biggest AI spenders cools, especially with markets already responding to the 4 Ts: Tehran, Trade, Taiwan and Tariffs, plus Tech.
Watch these specific signals: - Cisco surged 16% on strong guidance while semis were mostly lower, a split that can suggest enthusiasm is broadening away from the highest-quality AI beneficiaries. - The S&P 500 has already bounced close to 45% since the April trough. If that rebound starts to stall while cap-weighted tech struggles to hold recent highs, momentum may be weakening from the inside out.
What would invalidate Kabra's caution
The clearest invalidation would be an earlier-than-expected cash-flow turn among the biggest AI spenders. Kabra's timing call hinges on hyperscaler free cash flow reheading back into the positive in early 2027. If that happens sooner, or if the capex-to-sales profile improves faster than expected, waiting could prove more costly than staying concentrated.
A narrower way to stay exposed
While waiting for that signal, the cleaner stance is not "no tech." It is lower-concentration exposure. Kabra favors broader U.S. equity exposure through equal-weight S&P 500, and Goldman's work also points to a backdrop where cyclicals and a wider market base have helped the rebound, with the index having bounced close to 45% since the April trough.
- Keep core exposure in the broader U.S. equity proxy rather than chasing concentration.
- Add selective tech if hyperscaler free cash flow turns positive earlier than expected.
- Treat a stall near recent highs after the 45% rebound as the more actionable warning sign.