AI Boom or Bubble? Dot-Com Lessons for 2025

AI Boom or Bubble? Dot-Com Lessons for 2025

The artificial intelligence boom has become the dominant market story of the current cycle. Are investors asking one major question: Is AI a real structural growth engine, or are markets repeating the same speculative mistakes seen during the dot-com bubble?

To answer that, we need to compare the two periods across several dimensions: technology adoption, investor behavior, liquidity conditions, valuations, macroeconomic indicators, equity-market concentration, the U.S. dollar, crisis impact, government support, infrastructure spending, and monetization risk.

The dot-com era was built around the promise that the internet would rewrite the global economy. That promise was ultimately correct, but the market priced it too early, too aggressively, and across too many weak companies. The AI era has similarities, but it is not an exact copy. AI has real demand, real corporate spending, real profits in the core infrastructure layer, and strategic importance for governments and companies. Still, the outer layers of the AI trade show clear bubble behavior.

Dot-Com Bubble Timeline

• 1990–1994: Foundations

From 1991 to 1993, the commercial internet began to open. NSFNet restrictions eased, and early web browsers appeared. This created the foundation for a new digital economy.

In 1994, Mosaic helped lead to Netscape, while the first online retailers and internet portals began gaining traction. This period created the early belief that the internet would become a mass-market commercial force.

• 1995–1996: Spark and Legitimacy

On August 9, 1995, the Netscape IPO exploded on its first trading day. That event signaled a new era for internet investing.

Between 1995 and 1996, web users grew rapidly, banner advertising expanded, and venture funding accelerated. Companies such as Yahoo! and Amazon scaled up. Amazon was founded in 1994 and later went public in 1997.

• 1997–1998: Acceleration

In 1997, the IPO window widened for “.com” startups. The market embraced the “get big fast” model, where companies prioritized scale over profitability.

From August to October 1998, the Asia/Russia crises and the LTCM meltdown pushed the Federal Reserve toward easier policy. That liquidity helped fuel risk assets and added more momentum to the technology trade.

• 1999: Peak Euphoria

By 1999, tech IPOs reached record levels. Many companies were pre-revenue, but investors still bought them aggressively.

Market metrics shifted away from profitability and toward “eyeballs,” “page views,” and future potential. Dot-com advertising flooded the Super Bowl season, burn rates soared, and infrastructure overbuild began.

• Early 2000s: The Turn

On January 10, 2000, the AOL–Time Warner merger was announced. It became a symbol of peak optimism.

On March 10, 2000, the Nasdaq peaked near 5,048. Soon after, earnings misses and tighter financial conditions began damaging investor sentiment.

From April to December 2000, the IPO window slammed shut. Regulation Fair Disclosure was adopted in August 2000 and became effective in October 2000, limiting selective corporate disclosure.

• 2001–2002: Bust

In 2001, recession pressure hit the economy. Many dot-com companies failed or were acquired for very low prices. The telecom and fiber glut also unwound.

On September 11, 2001, the 9/11 attacks worsened the downturn and deepened the shock to the economy and the technology sector.

On July 30, 2002, Sarbanes–Oxley was enacted after major corporate scandals, including Enron and WorldCom.

On October 9, 2002, the Nasdaq bottomed near 1,114, roughly 78% below its peak.

• 2003–2004: Aftermath and Reset

After the bust, surviving firms refocused on unit economics. Cheap assets helped enable the next technology wave, including broadband, cloud infrastructure, and Web 2.0.

In 2004, the Google IPO marked a more disciplined internet era.

Why the Dot-Com Bubble Inflated

Technological Hype

The dot-com bubble inflated because investors believed the Internet would instantly rewrite every business model. The underlying technology was real, but the market overestimated the speed and profitability of adoption.

Cheap Capital, Easy Money, and Incentives

After the 1998 rate cuts, venture funds were flush with capital. Banks were eager for IPO fees, while stock options encouraged companies to prioritize growth over discipline.

New Metrics and Narratives

Investors embraced slogans such as “get big fast,” winner-take-all logic, and network effects. These ideas were often correct in theory but misapplied in practice.

Infrastructure Boom

The late 1990s also saw massive overbuild in fiber networks and data centers. Capacity expanded far ahead of real demand.

How the Dot-Com Bubble Burst

Unsustainable Valuations

Many internet startups had extremely high stock prices despite little or no revenue and no viable business model.

Shift in Sentiment

By early 2000, analysts and investors began questioning whether dot-com companies could ever become profitable. This led to a wave of selling.

Easy Money Flow and Accounting Irregularities

Low interest rates and abundant venture capital in the late 1990s created speculative investment in technology startups. This “easy money” environment encouraged rapid expansion and risk-taking without solid fundamentals.

Lack of Proper Profitability

Many firms had negligible sales and no clear path to profit.

Scandals and Shocks

Corporate fraud and macro shocks destroyed confidence. Enron, WorldCom, recession pressure, and 9/11 all damaged investor trust.

Lack of Tech-Specific Regulation

Regulators were unfamiliar with the fast-evolving internet sector. This led to an underestimation of the risks tied to digital business models and extreme valuations.

Milestones in AI Popularity

• 2016–2021: Setup, Not Yet “The Trade.”

AI advances such as AlphaGo and transformers have built the technological foundation. However, equity leadership was still broad technology, not a focused “AI trade.” The market inflection came later.

• November 2022–June 2023: Inflection and First Leg Up

On November 30, 2022, ChatGPT launched. That event triggered mass-market awareness and created a new earnings narrative around AI.

By May–June 2023, Nvidia joined the $1 trillion market-cap club, becoming the first chipmaker ever to do so. Its stock was up more than 170% year-to-date, driven by AI data-center demand.

• Second Half of 2023: Narrow Leadership and AI Narrative Dominance

The “Magnificent Seven” mega-cap technology companies drove most of the S&P 500’s gains. The group rose 75.7% in 2023, compared with the S&P 500’s 24.2% gain.

Investors poured into S&P 500 trackers as Magnificent Seven strength attracted global flows toward the U.S. A record 27% of global equity ETF flows went into S&P 500 funds in 2023.

• 2024: From Hype to Spending Cycle

Corporate earnings calls mentioning “AI” surged and stayed elevated quarter after quarter, with more than 200 companies per quarter citing AI.

Hyperscalers outlined massive AI data-center capital expenditure plans for fiscal year 2025. Microsoft, for example, signaled roughly $80 billion in FY25 spending.

• 2025: Broadening and Supply-Chain Pull-Through

Mentions of “AI” on S&P 500 earnings calls hit a 10-year high. In Q2 2025, 287 companies mentioned AI, a record level.

NVIDIA topped $4 trillion in market capitalization in July 2025, becoming a symbol of the AI build-out trade.

Memory and HBM bottlenecks also showed deepening supply-chain demand. SK hynix said HBM was sold out for 2024 and almost all of 2025.

Wall Street expected AI capital expenditure to keep climbing. Industry trackers saw 2025 data-center capex at record highs, with estimates in the $300 billion-plus range.

Despite some broadening beyond mega-cap technology, the Magnificent Seven still represented roughly 35% of the S&P 500 by late 2025. AI remained the market’s organizing narrative.

AI Popularity in Short

There are three key stages in the AI boom:

1. Visible investor traction:

Late 2022 to mid-2023, driven by ChatGPT and Nvidia reaching a $1 trillion valuation.

2. Sustained narrative dominance:

From 2023 to 2025, the Magnificent Seven led the market, and AI mentions on earnings calls reached record levels.

3. Real-economy follow-through:

From 2024 to 2025, hyperscaler capital expenditure, semiconductor demand, and HBM supply tightness confirmed that the AI boom had moved beyond pure market hype.

Milestone Mapping: Dot-Com Era vs AI Age

The dot-com era and AI era share a similar rhythm: early technological breakthrough, rising investor excitement, market leadership concentration, infrastructure expansion, and then the question of whether real monetization can justify valuations.

The dot-com cycle moved from early commercial internet adoption to IPO mania and then collapse. The AI cycle moved from research breakthroughs to mass adoption after ChatGPT, then to hyperscaler spending, Nvidia dominance, chip shortages, and broad corporate AI adoption.

The comparison does not prove that AI must collapse like the dot-com. But it does show that technology revolutions often create financial excess before the full economic value is realized.

Is It Fair to Say AI Is Inflated?

AI market inflation has been driven by a mix of hype, investment momentum, and speculative expectations. That echoes patterns seen in past technology bubbles.

• Valuation Bubbles

Startups and public companies saw valuations rise sharply. Some research suggests the AI bubble may be up to 17 times larger than the dot-com bubble in relative terms.

• Investor Overexcitement

High-profile breakthroughs such as ChatGPT and AlphaFold triggered a wave of optimism. Investors rushed in, often valuing companies based on potential rather than proven profitability.

• Overinvestment in Infrastructure

Massive spending on data centers, GPUs, and cloud platforms may have outpaced actual demand in some areas. Supply surged while real-world adoption lagged, creating a possible mismatch.

• FOMO Among Corporations

Non-tech companies rushed to adopt AI because they feared being left behind. This led to rushed deployments and inflated expectations.

Economic Indicator Comparison: Dot-Com vs AI Era

Inflation

Inflation was very different across the two technology cycles.

In the late 1990s, inflation was quiet, sitting around 2%. After the dot-com crash, inflation actually moved lower. That allowed the Federal Reserve to cut rates and flood the system with liquidity. Cheap money first inflated the bubble, then softened the landing after the crash.

Today’s AI boom is happening after the biggest inflation shock in decades. Prices moved toward 9%, and even after cooling, inflation remains around 3%. The Fed cannot easily return to “free money mode” without risking another inflation spike.

That means the AI boom is growing in a much tighter money world than the dot-com boom. The market is still euphoric about AI, but credit is more expensive, capital is more disciplined, and Washington is less willing to rescue speculators if the market cracks.

The AI story may still produce a bubble in certain names. Structurally, however, it is happening under harder money conditions than in 1999. That could prevent the same kind of “spray money at any idea” behavior — or, if AI still inflates to 1999-style extremes, it will mean the hype was strong enough to beat even tight policy. That would be even more dangerous. Because, in fact, this market growth has been achieved at a much higher cost, and if the project fails, the losses will be much greater.

Labor Market Comparison: Dot-Com vs AI Era

The labor-market backdrop is also different.

The dot-com cycle rose and fell in a mostly normal economy. The AI cycle sits on top of an economy that was bailed out on a historic scale and then pushed toward automation, reshoring, and AI investment as strategic priorities.

That changes how long the story can last — and who gets hurt if it cracks.

U.S. GDP Comparison: Dot-Com vs AI Era

U.S. GDP growth in the 1990s and early 2000s was mostly steady and mostly strong. Outside the 2001 slowdown, the economy kept expanding. The dot-com bubble was built on top of a broadly healthy and accelerating U.S. economy.

The last decade looks different. Growth has been slower. The economy suffered a full collapse in 2020, then a massive one-time spike in 2022, mostly driven by stimulus and reopening. After that, growth drifted lower again.

Today’s AI boom is happening in a world that is more fragile, more policy-driven, and structurally slower.

In the 1990s, the narrative was: technology is unlocking a productivity boom inside a strong economy.

Today, the narrative is: AI will become the productivity boom we are missing.

The first one was confirmation. The second one is a promise.

S&P 500 Comparison: Dot-Com vs AI Era

The S&P 500 went vertical into 2000, then lost roughly half its value and did not reclaim those highs for years. That is what a classic bubble looks like: the market priced in a perfect future, reality disappointed, and investors gave up for a long time.

The AI trade today is not just story and hype. But that does not mean there is no risk.

Today’s highs are heavily concentrated in a handful of ultra-cap technology and AI names. The S&P 500 has become more top-heavy than normal.

If that top layer stumbles on AI monetization, margins, regulation, export controls, or capital-expenditure payback, the downside could still be violent.

Nasdaq 100 Comparison: Dot-Com vs AI Era

During the dot-com bubble, the Nasdaq 100 went almost straight up into 2000, then collapsed by more than 80% and stayed down for years. That was a classic speculative bubble. Prices were far ahead of any real earnings base, leaving little underneath to support them.

The Nasdaq 100 in the AI era shows something different. This cycle is not only story and hype. There is real profitability, real cash flow, and real global capital crowding into the same winners.

Therefore, anyone waiting for an exact 2000-style implosion should recognize that the price action so far suggests this is not the same beast.

U.S. Dollar Index Trend: Dot-Com vs AI Era

The U.S. dollar also tells an important story.

During the dot-com era, the dollar was strong during the boom. After the bust, the dollar clearly broke down. The world effectively said: the party is over.

In the AI era, the dollar has also been strong during the AI and technology dominance phase. It has pulled back, but it remains structurally strong. The world is still treating the U.S. as the place to be.

That means even if AI is partly a bubble, the global capital base underneath it looks stickier today than it was after 2000. This supports the argument that the AI cycle can run longer than the dot-com cycle, even if corrections happen.

In the late 1990s, the dollar surged alongside the tech boom because global investors were buying the “new economy” story. After the dot-com crash, that dollar strength faded. Money left, and the U.S. no longer looked uniquely magical.

In the current AI cycle, the dollar spike was not only about optimism. It was also about fear: Federal Reserve tightening, geopolitical shocks, and supply-chain risk.

The key difference is that even after conditions cooled, the dollar remained relatively strong. Capital has not rotated out of the U.S.

That tells us two things.

First, AI is not just an equity-market narrative. It sits on top of global capital flows that still treat the U.S. as the safest and most productive destination.

Second, that support can keep valuations elevated longer than many investors expect, because the rest of the world still prefers to park capital in dollar assets.

U.S. Coincident Economic Activity Index: Dot-Com vs AI Era

The U.S. Coincident Economic Activity Index from the Philadelphia Fed also adds important context.

During the dot-com run-up, the real economy was strong and getting stronger. The index climbed almost in a straight line into 2000. Then it rolled over in 2001–2002, showing when the technology crash finally reached Main Street.

But the important twist is what happened afterward. By 2003–2006, the index surged higher again.

The lesson is clear:

• A sector bubble can form inside a very healthy economy.
• The real economy may soften only after the bubble pops, not before.
• The broader economy can recover and reach new highs, even if many bubble-era companies never come back.
• That message is important for AI.
• Strong coincident data today does not prove that there is no bubble.

A future AI correction, if it happens, does not automatically mean the end of the cycle forever.

The real question is this: after the hype shakes out, does the AI build-out become the next structural engine of growth, the way the internet and then housing and credit did after 2002?

9/11, COVID, and Markets

Global crises shape technology differently depending on context and timing.

The late 1990s saw a frenzied dot-com boom that ended abruptly around 2000. The September 11, 2001, attacks hit during that bust and further shocked the economy and technology sector.

Nearly two decades later, as AI and digital technologies were expanding, the world was hit by the COVID-19 pandemic in 2020. These two crises had very different impacts on technology.

9/11 and the Tech Downturn

The 9/11 attacks intensified an already declining technology market. The dot-com bubble had already begun bursting in 2000, and 9/11 effectively marked the end of that era of exuberance.

The lasting technology legacy of 9/11 was a greater focus on security technology, including biometric IDs and intelligence-related data mining, along with a more cautious business environment.

The crisis created a sense that unforeseen threats could “wreck the party” at any time. That arguably tempered the freewheeling innovation spirit for several years.

It took time for the technology sector to regain momentum. The rise of Web 2.0 around 2004 and beyond helped revive investor interest and entrepreneurial activity.

COVID and the AI Acceleration

After 9/11, money became scarce. Technology hiring slowed. The U.S. shifted into a security and war footing. The internet continued to evolve, but the bubble itself stayed dead for years.

After COVID, money did not retreat — it detonated. Liquidity surged, everything digitized, and AI shifted from “experimental” to “mandatory.”

COVID did not kill the AI story. COVID is one of the reasons the AI story scaled so quickly.

When people say AI is just another dot-com bubble, they need to ask whether the last bubble received an accelerant like COVID. In this cycle, the crisis poured fuel on the technology instead of choking it.

Importantly, 68% of U.S. businesses increased their investment in AI during the COVID-19 crisis, either by deploying new AI tools or expanding existing projects.

COVID’s Long-Term Tech Impact

Unlike 9/11’s dampening effect, COVID-19 created a technology surge with long-lasting effects.

By forcing the adoption of teleconferencing, online education, telehealth, and automation, the pandemic permanently shifted consumer and business behavior toward digital tools.

Many emergency measures became new norms. Companies adopted greater operational resilience, automation, and flexible work arrangements that are expected to remain part of the post-pandemic world.

COVID-19 did not create a one-time technology bump. It accelerated the Fourth Industrial Revolution, including AI, IoT, cloud computing, and automation. Society moved further into a digitally driven paradigm faster than expected.

Technology firms, both large and small, thrived in aggregate during the pandemic compared with other sectors. Investor capital flowed heavily into technology and AI startups solving new needs, from remote-work software to biotechnology.

This was very different from the post-9/11 period, when technology entrepreneurship went dormant for several years.

Crisis Comparison: 9/11 vs COVID

9/11 in 2001

The 9/11 attacks happened after the dot-com bubble had already burst. The attacks deepened the technology downturn.

Investor confidence collapsed, many startups failed, and innovation shifted toward security, surveillance, and defense technologies. The broader atmosphere became one of fear and regulation.

Consumer-technology progress slowed as attention shifted toward safety and stability. It took several years for entrepreneurial energy and risk appetite to return.

COVID-19 in 2020

The pandemic had the opposite effect. It accelerated technological and AI innovation.

Digital transformation jumped ahead by years. Remote work, automation, cloud infrastructure, and AI became essential.

Funding surged for cloud, data, and health-technology solutions. Technology became the foundation of economic resilience.

The Larger Lesson

Crises act as inflection points for innovation.

When technology is mature enough to solve urgent problems, as it was during COVID-19, it flourishes.

When technology is speculative or fragile, as internet investing was in 2001, it contracts and redirects.

Both events proved that necessity drives invention. 9/11 reinforced resilience and security. COVID-19 showed technology’s power to adapt, accelerate, and redefine the modern world.

The Core AI Questions

Is AI a Bubble Right Now?

The AI market has bubble behavior in the outer layers, but the core is not a classic fantasy bubble.

A true bubble happens when valuations disconnect from real earnings, capital floods in on hype, and investors assume instant transformation.

That described the dot-com era, where unprofitable startups and irrational IPOs collapsed once funding stopped.

Today, AI is different. We are not in a fully delusional 1999-style bubble. Instead, we are in an AI infrastructure boom with speculative excess around the edges.

It is risky, but it is rooted in real technological and economic value.

Why the Core AI Sector Is Different

The inner layer of AI is economically grounded, even if parts of it may be overpriced.

• Real Profits

Companies such as Nvidia generate massive revenues from actual demand for AI infrastructure.

• Real Customers

Major technology companies, including Microsoft, Alphabet, Meta, and Amazon, are spending tens of billions of dollars to deploy AI now, not hypothetically.

• Strategic Importance

AI has become a national and corporate necessity. It is increasingly treated like energy, semiconductors, or defense infrastructure. This is especially important after COVID, as companies and governments focus on automation, productivity, and resilience.

This means the inner AI layer — chips, cloud infrastructure, and foundation models — is not imaginary.

Where AI Shows Classic Bubble Traits

• The outer AI layer still shows classic bubble behavior.
• Some startups are overvalued without defensible advantages.
• Many companies are engaging in “AI-washing,” adding the AI label to every product.
• Infrastructure expansion is moving ahead of proven returns.
• Markets are also showing an unrealistic belief in overnight industry disruption.
• This is where AI looks most like a bubble: not in the core technology, but in the speculative edges around it.

How Close Are We to a Full-Blown AI Bubble?

We are likely in the “late middle innings” of the AI cycle. The market is inflated, but not yet at peak euphoria.

Signs of Rising Bubble Risk

Extreme concentration:

A small group of mega-cap technology companies dominates market performance. This is a hallmark of late-stage bubbles, where everything revolves around one narrative.

Assumed infinite demand:

Markets are pricing in endless AI growth, including expanding workloads, rising spending, and fast monetization. Straight-line optimism rarely survives.

Second-derivative risk:

Even a mild slowdown in AI spending growth could sharply hit supplier revenues because valuations assume perfection.

Signs We Are Not Yet at Terminal Bubble Conditions

Profitable leaders:

Unlike many previous bubbles, the core AI leaders generate massive profits and can self-fund through tighter liquidity conditions.

Expanding real use cases:

AI is now embedded across industries, including customer support, code writing, logistics, compliance, and drug research. That proves genuine utility beyond hype.

Policy tailwinds:

Governments treat AI as a strategic priority. They subsidize compute infrastructure and protect chip supply chains. This is very different from the dot-com era’s “cool website” logic.

Bubble Scale Rating

The AI market looks like a partially bubbly super-cycle. It is driven by genuine technological demand, but inflated by excessive expectations.

A reasonable rating is roughly 7–8 out of 10 on the bubble scale.

Valuations and hype are stretched, but the cycle is still anchored by real profits, real adoption, and geopolitical importance.

If AI Bursts, What Actually Breaks?

If the AI boom bursts, the damage may not come from one source. It could come from several pressure points: energy constraints, policy reversal, infrastructure overbuild, monetization disappointment, narrative concentration, geopolitical shocks, and macro stress.

Energy and Power Constraints

AI growth depends on massive physical resources: land, electricity, and cooling.

Governments are increasingly treating AI data centers as strategic infrastructure, prioritizing energy for AI, defense, and surveillance.

This creates rapid “build now” spending and assumptions of endless demand. However, if a recession hits or public backlash grows over energy costs and power-grid strain, funding could be capped. That would damage the growth assumptions behind the AI infrastructure cycle.

In the short term, political pressure supports capacity expansion.

In the medium term, public pushback and spending freezes become real risks.

That volatility — from urgent buildout to sudden restraint — is classic bubble fuel for the AI infrastructure cycle.

Government Money and the AI Industrial Policy Race

Governments are treating AI as a strategic industry, similar to semiconductors or green energy. They are subsidizing and protecting it.

Short-Term Effect

Multiple nations are funding the same types of infrastructure: data centers, chip fabs, and model labs.

Investors may assume that the state will always underwrite these projects.

Easy money can fuel overvaluation and weak players, especially when funding shifts from business merit to political narratives.

Bubble Risk

When political priorities change, such as after a new government takes power or fiscal discipline returns, projects dependent on state support can collapse quickly.

Companies once seen as “national champions” may become stranded assets.

That would reveal that many ventures were politically sustained rather than commercially viable.

The bottom line is that government-backed enthusiasm can create a policy-driven bubble. It remains stable while politics support it, but it becomes vulnerable to a sudden rug pull when subsidies fade.

Monetization Disappointment Risk

The biggest potential trigger for an AI market correction is monetization disappointment at the top.

Markets currently assume AI will:

• Generate massive new revenues for hyperscalers.
• Justify higher software prices.
• Cut labor costs across industries.
• These assumptions fuel soaring valuations and index concentration.
• But if major companies hint at slower adoption, high inference costs, or hesitant customers, even mild “not yet” language could erase trillions of dollars in market value.
• The irony is that bubbles rarely burst because belief in the technology disappears. They burst when belief in the timeline breaks.
• Overhyped expectations of immediate monetization create the setup for abrupt repricing once reality proves slower than the story.

The Narrative Capture Risk

A major late-stage bubble risk is narrative capture. This happens when AI becomes the single national obsession.

If AI is framed as the only route to progress or power, other industries lose investment and attention. Capital floods into AI while sectors such as manufacturing, retail, and media are dismissed as obsolete.

The market then becomes dangerously concentrated around one story.

This creates systemic fragility. When every valuation depends on the same narrative, any disruption — regulation, lawsuits, policy shifts, or safety concerns — can trigger a broad market shock.

Once AI dominates both political and investment narratives, the system becomes all-in on one pillar. Any crack in that pillar can turn a hot market into a full-blown collapse almost overnight.

Geopolitical and Macro Risks That Could Burst the AI Boom

Several global and policy dynamics could inflate and then burst the AI bubble.

• Great-Power Conflict

Heightened geopolitical rivalry could trigger panic investment in AI and computing capacity. This could create a melt-up followed by a crash once demand normalizes.

• Tech Sovereignty and Chip Controls

Export restrictions could lead to front-loaded stockpiling of chips and infrastructure. That may later result in oversupply and correction.

• Energy and Land Politics

Declaring AI data centers a national priority could drive a building frenzy. Later, that buildout could meet public backlash over energy costs, land use, and resource strain.

• Subsidy Distortions

Industrial-policy money can support weak companies and allow unviable projects to scale. When subsidies end, collapse can follow.

• Monetization Disappointment

If hyperscalers reveal that AI investments are not yet paying for themselves, market sentiment could reverse sharply.

In essence, these risks combine geopolitical urgency with economic overreach. That creates the perfect setup for a fast, policy-fueled AI boom that can end in a hard correction.

Final Thoughts: AI Boom or Bubble?

Right now, AI is not quite at peak euphoria. But geopolitics can accelerate the bubble instead of killing it.

If Washington, Beijing, and other major powers decide that AI and chips are national security priorities, money can flood into the sector. Capacity can become overbuilt. Valuations can go vertical.

That looks like strength — until one thing happens: growth slows even slightly.

That is when markets discover which part of AI is a real new industry and which part is wartime FOMO funded by governments.

The first part survives.

The second part blows up.

FAQ

Is AI currently in a bubble?

AI has bubble behavior in the outer layers, but the core is not a classic fantasy bubble. The core AI infrastructure layer is supported by real revenues, real customers, real capital expenditure, and strategic government interest. However, overvalued startups, AI-washing, and unrealistic monetization expectations show clear signs of speculative excess.

How is the AI boom different from the dot-com bubble?

The dot-com bubble involved many companies with little revenue, weak business models, and no clear path to profitability. The AI boom includes highly profitable leaders such as Nvidia and major customers such as Microsoft, Alphabet, Meta, and Amazon. However, both periods share similar risks: concentration, overvaluation, infrastructure overbuild, and excessive optimism.

Why did the dot-com bubble burst?

The dot-com bubble burst because valuations became unsustainable, investor sentiment shifted, companies lacked profitability, easy money dried up, and shocks such as recession pressure, corporate scandals, and 9/11 destroyed confidence.

Could AI stocks collapse as the Nasdaq did after 2000?

A full repeat is not guaranteed because today’s AI leaders have stronger earnings and cash flow than many dot-com companies did. However, the downside could still be severe if AI monetization disappoints, capital expenditure slows, regulation tightens, or investors lose confidence in the timeline of AI adoption.

What is the biggest risk to the AI boom?

The biggest risk is monetization disappointment. Markets assume AI will generate massive revenues, justify higher software prices, and reduce labor costs. If major companies suggest that AI returns are slower than expected, valuations could reprice sharply.

Why does energy matter for AI?

AI requires huge amounts of electricity, land, and cooling capacity. Data-center expansion depends on energy availability. If electricity costs rise or public opposition grows, AI infrastructure spending could face delays, caps, or political resistance.

What role do governments play in the AI boom?

Governments treat AI as a strategic industry, similar to semiconductors, defense, and energy. Subsidies and industrial policy can extend the boom, but they can also create a policy-driven bubble if funding supports projects that are not commercially viable without state support.

Is the AI boom supported by real demand?

Yes, the core AI boom is supported by real demand from hyperscalers, cloud providers, chip buyers, enterprises, and governments. The issue is not whether AI is real. The issue is whether current valuations assume too much too soon.

Are we close to peak AI euphoria?

The AI cycle appears to be in the late middle innings. Bubble risk is elevated, but the market has not yet reached terminal euphoria. A reasonable bubble-risk rating is around 7–8 out of 10.

What would survive if the AI bubble bursts?

The real infrastructure, productivity tools, profitable chip and cloud leaders, and useful AI applications would likely survive. The weakest layer — overvalued startups, politically funded projects, AI-washed companies, and infrastructure built ahead of demand — would be most vulnerable.