Reviewing the "Five Major Bubble Indicators," Goldman Sachs believes "the current situation is more like 1997 rather than 1999, and the AI bull market still has a second half."
For investors closely watching whether the AI-driven rally in U.S. stocks has entered bubble territory, Goldman Sachs has given a clear answer: it’s not a bubble yet, at least not on the scale of the "macro-bubble" seen in 1999-2000.
According to Chasing Wind Trading Desk, on November 9, Goldman Sachs published a report stating that current macro fundamentals resemble those of the mid-bubble years of 1997 or 1998, rather than the peak in 1999. The critical imbalances which ultimately led to the bubble’s collapse — such as widespread investment overheating, deteriorating corporate profits, and a sudden surge in leverage — have not appeared.
This means that, despite high valuations, the AI-driven bull market may still have a second half to play. Exiting too early could mean missing out on considerable further gains. However, risks are accumulating, and investors should begin setting up hedging strategies. Unlike 1999, current credit spreads and market volatility remain low, providing a more cost-effective window for investors to manage risk through options and other tools.
The Five Key Indicators of a “Macro-Bubble”
Goldman Sachs first clarified one point: simply high valuations do not equate to a “macro-bubble.” True macro-bubbles, such as the late-1990s internet bubble, involve not only severely overvalued asset prices but also macro imbalances that have a significant impact on the real economy. The report systemically reviewed five key macro and market characteristics of the 1990s:
Massive investment frenzy: By early 2000, investment in tech equipment and software as a share of GDP soared from a little over 3% at the start of 1995 to a record 4.5%. Total non-residential investment as a share of GDP rose from about 11% in 1992 to nearly 15% in 2000.
Profitability peaking then declining: Despite ongoing gains in productivity, corporate profit margins had already peaked by late 1997. A tight labor market drove wages higher, eroding corporate profits.
Leverage spiking sharply: Surging investment and falling profits forced the corporate sector from surplus to deficit, with a substantial rise in corporate debt and a worsening balance sheet health.
External crises spurred capital inflows and Fed rate cuts: The Asian financial crisis and Russian default in 1997-1998 led to large capital flows from emerging markets into the U.S. To counteract financial pressures, the Federal Reserve cut rates by 75 basis points at the end of 1998, further fueling the stock market.
Credit and volatility markets send warnings: From mid-1998, even as the stock market accelerated upwards, credit spreads and Nasdaq volatility rose in tandem, reflecting that other asset markets had started repricing for risk.
Why hasn’t the current AI boom “bubbled up” yet?
The report compared today’s AI boom against the above five indicators, concluding that signs of macro imbalance are far from the late-1990s level.
Investment has started, but scale and breadth lag: Though capital expenditure by the "AI super-giants" is expected to double since the launch of ChatGPT, AI-related investment as a share of GDP remains much lower than the telecom peak of 2000. In terms of duration and breadth, the current investment frenzy is much milder.
Profitability remains solid: Corporate profit margins are stable, with no signs of deterioration. Productivity has rebounded, while wage growth is slowing, leading to a sharp drop in unit labor costs—completely opposite to the late 1990s.
Financial health is relatively robust, leverage is controllable: Unlike the late 1990s, when the corporate sector fell into deficit, today’s companies still enjoy financial surpluses. Large tech firms mainly fund capital expenditures from free cash flow rather than debt; balance sheets are generally strong.
External environment differs, no catalyst for capital influx: The U.S. current account deficit is large but stable, with nothing like the externally driven, large net capital inflows of the late 90s.
Credit spreads and volatility remain low: With low leverage, credit spreads are still extremely narrow. Equity market implied volatility has not shown sustained increase. The report states credit spreads and volatility today resemble 1997, not 1999.
The party is still in 1997, but the 1998 turning point is approaching
Overall, Goldman Sachs believes the macro footprint of the current AI boom is fundamentally different from the late-stage (1999-2000) tech bubble, and much more like the early stage (1997-1998). The report notes that AI’s estimated contribution to GDP growth is currently only about 0.3 percentage points, resembling the early stage of the 1990s tech boom.
However, the report sharp-eyed detected potential signals of a “1998-style” turning point:
Acceleration in investment plans: Capital expenditure plans by AI giants and private companies hint at continued rapid growth in AI-related investments.
Financial balance nearing inflection: The corporate sector’s financial surplus is being eroded, nearing a deficit for the first time in 20 years. The asset and liability advantages of tech giants are no longer as prominent as before.
Debt financing rising: The proportion of debt-financed data center investment is increasing; new AI deals are boosting demand for debt issuance and driving complex arrangements similar to “vendor financing.”
External funds and loose policy: The Fed has begun a "preventive" rate-cut cycle; meanwhile, governments from the Middle East and Japan have announced over $4 trillion in investment commitments, which, if realized, could play a role similar to late-1990s external capital inflows.
How to balance offense and defense in the AI bull market’s second half?
The report says, since the current market is more like 1997, calling a bubble and exiting too early is dangerous—history shows this risks missing the period of highest returns. However, valuations have already gone ahead of macro fundamentals, and the market cap growth of AI-linked firms far exceeds their predicted capital income present discounted value (PDV).
So for investors, the key is how to balance ongoing participation with risk protection. The report offers concrete advice:
Using options for offense-defense rotation: Unlike the high spreads and volatility in 1998-2000, today’s low-volatility environment is fertile ground for options strategies. Investors can consider using lower-cost bullish option structures to keep capturing upside while limiting downside risk.
Set up risk hedges ahead of time: History suggests that even in bull markets, credit spreads can widen as debt rises. Thus, over the coming one to two years, it’s reasonable to position for potentially wider credit spreads or higher long-term equity volatility as an effective hedge during an ongoing AI bull market.
Watch out for bi-directional rate risks: If AI investment boom continues, corporate financing needs will compete with government deficits for capital, possibly pushing up long-term rates. But if AI excitement is derailed, history shows both policy rates and long-term yields could eventually fall sharply.
In summary, Goldman Sachs’ report gives investors a complex roadmap: the AI party isn’t over, but smart investors should start listening for changes in the music and be ready to face the next phase of risks and opportunities.
Risk Warning and DisclaimerThe market carries risks, investments should be made cautiously. This article does not constitute personal investment advice, nor does it take into account individual users’ specific investment goals, financial situations, or needs. Users should consider whether any opinions, views, or conclusions in the article are suitable for their particular circumstances. Investment decisions are at your own risk.