Greenspan’s Legacy Is Still in Tech Valuations

For today’s AI and software builders, the important question is not whether the late-1990s tech bubble happened. It is why the valuation habits that grew up around that era still feel familiar.

The short answer is that Greenspan-era monetary policy helped normalize a way of thinking about growth assets that modern founders, investors, and board members still live with: if the future is far enough away, discount rates, liquidity, and central-bank behavior can matter more than present-day profitability. That habit did not disappear when the dot-com bust arrived. It moved into later software cycles, then into SaaS, then into today’s AI-heavy market.

The result is a valuation culture that can swing between two extremes. In easy-money periods, long-duration tech gets priced as if growth and optionality will outrun cost of capital forever. When rates rise or growth slows, those same models can reprice brutally. The legacy is not just “cheap money.” It is a lasting willingness to treat distant cash flows as if they were close enough to finance reality away.

The Greenspan-era lesson was not subtle

The intellectual backdrop matters. Greenspan’s period as Fed chair is often remembered for “irrational exuberance,” but one paper in our sources argues that the Fed did not actually try to restrain late-1990s stock valuations. It found “no empirical evidence” that policy attempted to moderate them and instead concluded that the Fed accommodated high valuations. ¹

That finding fits the broader “Greenspan put” literature. The market came to believe that central banks would step in when asset prices fell hard, even if that was never an official policy. Frederick Mishkin’s line captures the problem well:

"I don't know about a Greenspan put, but there was some element of that — and it is very hard to dissipate that impression."

— Frederick Mishkin ²

That impression matters because markets do not need a formal promise to behave as though a promise exists. Once investors expect central-bank backstops, risk appetite changes. The sources in this set repeatedly point to the same mechanism: lower perceived downside encourages leverage, duration extension, and valuation regimes that lean on future growth more than current cash generation. ^{2, 3, 4}

Greenspan also supplied the era’s philosophical language. In a retrospective on the period, he was quoted as saying that economic value had shifted toward “conceptual values” created by scientific insights and technology. That framing did not just describe a productive economy. It gave cover to a market narrative in which intangible output deserved special valuation treatment. ⁵

"The creation of economic value in recent decades has shifted toward conceptual values – that is, those created by new scientific insights and technology – with far less reliance on physical volumes."

— Alan Greenspan ⁵

Why this mattered for valuation models

Once the market accepts that technology is different, valuation discipline changes with it.

A DCF model is still a DCF model, but its assumptions become more elastic when the asset being modeled is a software platform, a cloud business, or now an AI workflow layer. High-growth tech is especially sensitive to discount rates because so much of its terminal value sits far in the future. That is the simple math behind the late-ZIRP boom in SaaS and the broader re-rating of long-duration equities. ^{6, 7}

LongYield’s summary is direct:

"Software equities — with their recurring revenue streams, high gross margins, and cash flows weighted heavily toward future years — became the purest expression of a long-duration equity trade."

— LongYield ⁶

That description could just as easily be applied to the latest AI infrastructure and application wave. Builders know the story: early losses are tolerated if the future market is large enough, margins are high enough, and the discount rate is low enough. What the Greenspan era helped normalize was the market’s comfort with that tradeoff. If capital is cheap and central banks are perceived as supportive, investors become more willing to pay for possibility.

A later source on valuation methodology shows that analysts moved toward DCF for high-growth sectors after the high-tech bubble, but the caveat is important: DCF is theoretically robust and practically fragile because it is highly sensitive to discount-rate and long-term-growth assumptions. ⁸ That fragility is exactly why monetary policy has outsized influence on tech multiples. Small rate changes produce large changes in present value when the cash flows are pushed years into the future.

The real legacy is behavioral, not just mathematical

The easiest mistake is to think the legacy here is just lower rates or more liquidity. The deeper effect is behavioral.

One paper in the set argues that when central banks suppress nominal volatility, economic actors shift from idiosyncratic innovation risk toward systemic financial risk. In that environment, leverage and macro asset-price strategies can outcompete risky innovation. ⁴ That is a useful lens for modern tech, because it helps explain why funding markets can reward narrative, scale, and financial engineering long before they reward durable business quality.

This is also where venture capital enters the story. A discussion paper excerpt in the sources criticizes the idea that VC merely “insures” entrepreneurs against risk, but even that critique is useful here because it shows how persistent the moral-hazard debate has been. When capital is abundant and downside feels socialized by policy, founders and investors may behave as if the cost of failure is lower than it really is. ⁹

The ZIRP-era source pushes this logic into operating culture. It says companies grew accustomed to the idea that “growth at any cost” could function as a valuation methodology, and that cost became finance’s problem rather than engineering’s concern. ^{10, 11}

"Growth at any cost became not just a strategy but a valuation methodology."

— Ithron ¹⁰

That line is not about one funding cycle. It is about the operating norm that followed cheap capital into software organizations. If you are a founder, this matters because product, hiring, and go-to-market decisions all absorb the assumptions embedded in the funding environment. If you are an investor, it matters because the same assumptions tend to leak into comparables, benchmarks, and exit expectations.

The modern echo is SaaS, then AI

The cleanest modern evidence comes from SaaS. One 2026 index study shows how near-zero discount rates inflated SaaS multiples to 16.9x revenue in late 2020 and 2021, then how the Fed’s 2022-2023 tightening cycle compressed them by roughly 60% by end-2022. It also notes a further 2026 re-rating driven not by rates but by AI disruption fears and slower growth. ⁷

Another valuation source says plainly that, during the near-zero-rate years, “immediate cash flow generation became less important.” When short-term Treasury yields moved above 5%, multiples for loss-making companies fell fastest. ¹²

That is the modern version of the same old rule: if the discount rate is low enough, future growth can justify almost anything. If the discount rate rises, the illusion breaks.

"Near-zero discount rates mathematically inflate the present value of future cash flows. A company expected to generate $1B in free cash flow in 10 years is worth vastly more when discounted at 0.5% vs 5%."

— SaaS Capital Index ⁷

The important point for builders is that this is not a software-only distortion. It is a capital-formation distortion. The venture and public markets that price AI companies today still rely on the same mechanics: a belief in very large future markets, a willingness to underweight current losses, and a tolerance for narrative-heavy operating plans so long as the macro backdrop is supportive.

Why the old regime may have outlived its usefulness

Here the historical analogy becomes less comforting.

The late-1990s boom ran on a mix of productivity optimism, cheap capital, and an assumption that technology was altering the economy’s value creation function. The current AI cycle has real productivity promise too, but the sources here warn against treating technology analogies as blueprints. A 2026 a16z discussion explicitly says older shifts such as PCs to mobile are “non-predictive heuristics rather than reliable blueprints.” ¹³

That caution should be taken seriously. The fact that past rate environments inflated tech valuations does not mean today’s AI market must follow the exact same path. It only means the market infrastructure is still primed to make the same mistake: pricing long-duration promises as if they were nearer-term cash flows.

The private markets evidence suggests the same tension. Secondary-market infrastructure now gives insiders and late-stage holders more ways to realize paper gains even as companies remain private longer. The panel summary in our sources warns that democratized access is often driven by people with a direct economic interest in selling their positions. ¹⁴

"The move toward democratized private-market access is being steered by participants who have a direct economic interest in selling their own late-stage positions to new entrants, so listen to the cautions about 'size' and 'FOMO' more than the marketing on potential upside."

— All-In Podcast, via 1 Minute Signal coverage ¹⁴

That is the hidden bridge from Greenspan to now. When policy makes money feel abundant for long enough, markets build habits around liquidity, exits, and duration. Those habits outlast the policy regime itself.

What builders and investors should take from this

Three practical implications stand out.

First, valuation discipline should be stress-tested at multiple discount rates, not just the one that flatters the current market. In a long-duration business, a few points of rate change can dominate the model. ^{6, 7}

Second, beware of narrative carryover. If a market has spent years rewarding growth at the expense of economics, the next cohort of investors will often inherit those assumptions as if they were facts. That is how “temporary” macro conditions become permanent operating habits. ^{10, 11}

Third, separate product truth from financing truth. A product can be genuinely useful and still be badly priced. That distinction gets blurred in periods when policy and liquidity make every category look expandable. The 1990s taught the market that tech could be economically transformative. The years since have taught it something less flattering: transformative sectors can still be over-discounted, over-levered, and over-owned.

The Greenspan legacy, then, is not that central banks caused every tech bubble. It is that they helped create the mental model in which high-growth technology could be valued as a claim on a very distant future, with policy backstops and liquidity assumptions doing a lot of the work.

That model still shapes AI. The question is whether today’s builders and investors recognize it before the next repricing does.