Goldman Sachs: AI Market Optimism Fading, Early August Rally Partially Reversed
Nashnova编辑部
Goldman Sachs says the brief AI optimism rally in early August has partly reversed. Strong Q2 hyperscaler earnings eased some fears but failed to set a new direction for the AI narrative, and stretched valuations leave the market exposed to any challenge to the bull case.
AI stocks have run up hard — can the valuations hold?
Goldman argues that the implied valuations baked into AI-related stocks keep climbing.
This means → justifying today's prices requires increasingly optimistic assumptions about AI's ultimate economic payoff.
In plain terms = the market has already priced in most of the good news. The margin for error is shrinking, and any disappointment can trigger a sharp pullback.
How does Goldman track the market's AI views?
Goldman built a framework that watches semiconductor stocks and hyperscaler stocks move together — or apart — to decompose sentiment into two types of shocks.
"Aggregate" shocks: the market raises or lowers its view of AI's total value — both groups move in the same direction.
"Distribution" shocks: the market's view of total AI value stays the same, but the expected split of who captures the profits shifts — the two groups move in opposite directions.
Using daily return data from 2025 onward, Goldman finds that distribution shocks are significantly more volatile than aggregate shocks. This reflects a deeper divide: the market argues more fiercely over who wins from AI than over whether AI wins at all.
What has the AI narrative looked like over the past 18 months?
Goldman identifies seven phases, each marked by swings in aggregate sentiment and profit-distribution expectations:
Phase 1 (Jan–Apr 2025): the DeepSeek episode triggered a negative aggregate shock; profit expectations shifted from semis to hyperscalers. Phase 2 (Apr–Sep 2025): a large positive aggregate shock, distribution still favoring hyperscalers.
Phase 3 (Sep 2025–Mar 2026): hyperscaler concerns mounted, negative aggregate shock, distribution swung back to semis. Phase 4 (Mar–Jun 2026): positive aggregate shock plus a large distribution tilt toward semis.
Phase 5 (Jun–Jul 2026): semis were downgraded, negative aggregate shock, distribution shifted back to hyperscalers. Phase 6 (late Jul–early Aug 2026): a brief positive-sentiment rally. Phase 7 (early Aug 2026–now): that rally partly reversed, and distribution tilted back toward semis once more.
Why do these swings matter beyond the AI sector?
Goldman finds that even amid frequent macro shocks, AI narrative shifts carry independent explanatory power over returns in the Nasdaq, emerging markets, South Korea, Taiwan, and Japan.
This means → AI is no longer just a sector theme — it has become a standalone driver of equity markets across multiple countries and regions.
What comes next?
Goldman says it is watching the last major tech company to report earnings this week, looking for a potential catalyst.
The key question: can this report re-establish a more durable direction on either the "aggregate value" or "profit distribution" axis — rather than another short-lived swing?
市场有风险，内容仅供研究参考，不构成投资建议。