The Numbers Don’t Lie: AI Investment Has Reached Dot-Com Peak Territory

The enormity of the volume of capital flooding into artificial intelligence is staggering, and increasingly familiar to anyone who lived through the late 1990s.

• Total corporate investment in AI hit $252.3 billion in 2024, with private investment jumping 44.5% year over year. (Stanford AI Index, 2025)
• Over 70,000 AI-related companies now operate globally, with thousands offering nearly identical large language model (LLM) wrappers or ‘AI-enhanced’ versions of existing SaaS tools.
• NVIDIA’s market cap surpassed $3 trillion in 2024, briefly making it the most valuable company on Earth — driven almost entirely by AI demand.
• At the peak of the dot-com bubble in 2000, the NASDAQ had risen 400% in five years. The Philadelphia Semiconductor Index rose over 500% between 2019 and 2024.

The parallels are not subtle. During the dot-com era, the mantra was ‘get big fast.’ Today’s equivalent is ‘move fast and add AI.’ In both cases, the underlying assumption was that market size and growth rate could substitute for sustainable competitive advantage.

We’re Witnessing the LLM Wars Already Consolidating

Consider the large language model market. Technically the engine room of today’s AI economy. In 2022, when ChatGPT launched, the race was on. Within 18 months, there were hundreds of competing foundation models: GPT-4, Claude, Gemini, Llama, Mistral, Falcon, Grok, and dozens of regional and enterprise variants. By late 2024, the pattern had already begun to emerge. Smaller players are quietly shutting down, pivoting, or being absorbed. Inflection AI, once valued at $4 billion and backed by Bill Gates, Eric Schmidt, and Reid Hoffman, sold its core team to Microsoft in a deal that looked, to many observers, like a fast and loose acquisition. Stability AI, pioneer of Stable Diffusion, faced financial distress and leadership upheaval. Character.AI, valued at $5 billion, saw its founders return to Google.

So much shifting in such a short time. This is not a failure of the technology. It comes down to a failure of differentiation. When every player is building on the same foundational models and the same public datasets, what exactly is your moat? What makes you special at all? And this is a key discussion topic for consumers of AI as well. What makes what you create with AI special?

The Cell Phone Wars Showed Us Exactly How This Ends

For those who find the dot-com comparison unrelatable, consider a more recent and instructive precedent: the smartphone industry. In 2007, the iPhone launched and ignited an explosion of mobile innovation. By 2010, there were dozens of major smartphone manufacturers: Nokia, BlackBerry, HTC, Motorola, Sony Ericsson, LG, Palm, and Samsung — all fighting for the same consumer. Hundreds of smaller OEMs and regional players added to the chaos.

By 2023, the global smartphone market was effectively a duopoly: Apple and Samsung control over 50% of global shipments by value, with Xiaomi, OPPO, and a handful of others splitting the remainder. Nokia is a shadow of itself. BlackBerry is gone as a hardware company. HTC sold its phone division to Google. LG exited the market entirely.

What separated the survivors? Not just better hardware. The decisive factor was patent portfolios and platform lock-in. Apple’s fierce protection of its iOS ecosystem and design patents, Samsung’s massive semiconductor and display IP, Google’s Android platform – all of these were protected by a vast portfolio of foundational mobile patents that kept the giants alive when there were no more “new” smartphone consumers.

During the smartphone wars, over 250,000 patent applications related to mobile technology were filed between 2007 and 2015. The companies that filed and enforced were strategically the ones that survived consolidation. The AI industry is now entering the same model. The question every AI company, and every enterprise investing in AI, should be asking is “what is our IP strategy?” There’s no sustainability when everything, everywhere has and is powered by AI systems.

So, Is the AI Bubble Going to Burst? Here’s Our Position.

Yes, and no. Here is where we take a clear stance. The AI bubble, in the sense of the all-encompassing, hype-driven, capital-at-any-valuation wave, will correct itself fast. It doesn’t have a choice. The tech infrastructure that it integrates with is finite. When Goldman Sachs analysts question whether AI infrastructure spending will ever produce equal returns, and when enterprise buyers begin demanding ROI from the AI tools they’ve been piloting for two years, the market will separate clarity and realism from smoke and mirrors.

But AI itself is not a fad – this is very important. The dot-com crash did not kill the internet. It killed bad business models, those that were not protected or positioned for IP survival. Amazon survived. Google was born at the end of the dotcom plateau. The same will be true of AI. The correction will not erase the technology; the technology is the foundation and is required for AI to exist. It will concentrate value in the hands of companies that built real, defensible competitive advantages and the single most defensible advantage in a technology market is intellectual property.

AI patent filings at the USPTO more than doubled between 2002 and 2018 alone — from 30,000 annual applications to over 60,000 — with filings accelerating dramatically in subsequent years. The race to own foundational AI IP is already well underway.

The Real Issue: Most Companies Have No Idea What AI IP They Already Own or What They’re Giving Away

Here is the critical insight that most organizations are missing right now, and it may be the most important thing you takeaway. As enterprises rush to deploy AI; embedding it in products, workflows, customer interactions, and R&D pipelines as much as they possibly can to make all processes as efficient as possible, they are simultaneously generating novel innovations and exposing themselves to IP risk on two main areas:

First, the innovations being created through AI-assisted R&D represent patentable advances that, if left undocumented and unfiled, will enter the public domain or be claimed by competitors. Second, the AI tools companies are using, particularly when those tools are trained on or exposed to proprietary data, create complex questions of IP ownership and disclosure that most legal teams are completely unprepared for.

This is not a hypothetical risk. It is an active one. And it is playing out right now, invisibly, in R&D labs, engineering teams, and product development cycles across every industry. We have a lack of visibility and lack of imagination issue with regard to AI IP.

This is precisely the inflection point where IP.com becomes not just valuable, but essential. IP.com has spent over two decades building the infrastructure that helps organizations understand, track, and strategically navigate the intellectual property landscape. In the AI era, that mission is more urgent than ever. For innovation leaders and patent professionals navigating the AI shakeout, IP.com provides three critical capabilities:

1. Prior Art & Landscape Intelligence:

Before committing R&D resources, you need to know what’s already been claimed. InnovationQ+™ from IP.com delivers AI-powered prior art search and competitive landscape analytics, powered by the proprietary Semantic Gist® engine, so innovation teams can identify white space, avoid costly patent disputes, and make smarter filing decisions from the start.

2. Competitive Patent Monitoring:

Knowing what your competitors are filing and in which technology domains is strategic intelligence, not a nice-to-have. InnovationQ+™ gives innovation teams direct visibility into competitor patent portfolios, technology investment trends, and emerging market entrants, so you can pivot, differentiate, or accelerate with confidence.

3. Invention Disclosure & Documentation:

The innovations your teams create today need to be captured before they’re lost or claimed by someone else. IQ Ideas+™ streamlines the entire journey from idea to invention disclosure, using dual AI engines to evaluate novelty, score patentability, and auto-generate structured disclosures ready for review, cutting screening time by up to 80–90% compared to traditional methods.

The dot-com bubble did not punish innovation; it punished innovation without a plan. The smartphone industry competition did not punish ambition; it punished ambition without IP protection.

The AI bubble correction, when it comes and in whatever form it takes, will follow the same logic. The organizations that emerge stronger will be those that treated intellectual property not as a legal afterthought, but as a core strategic asset along the way while integrating AI, using AI, and innovating with AI in their workflows. The question for every innovation leader reading this is not ‘Will AI be disrupted?’ It’s ‘When the dust settles, will my organization own anything?’

IP.com exists to make sure the answer is yes, but only if you act before your competitors do. Schedule a demo and let’s start building your IP strategy today.

Sources:

• Stanford HAI AI Index 2025 — https://hai.stanford.edu/ai-index/2025-ai-index-report/economy
• USPTO “Inventing AI” Report (2020) — https://www.uspto.gov