By Josh Kilen, Founder & CEO, Cascade Digital Marketing
I spend a lot of time thinking about AI.
Not just because we use it at Cascade Digital Marketing to enhance our clients’ campaigns—though we do, and I’ll explain how in a moment. I think about AI because, after eight years of helping professional service firms navigate digital marketing’s endless hype cycles, I’ve developed a pretty good nose for what’s real and what’s vapor.
And AI? It’s both.
Let me explain what our research uncovered—and why it matters for every business owner reading this.
The Numbers That Keep Me Up at Night (In a Good Way… Mostly)
Here’s what caught my attention: In 2025, worldwide spending on artificial intelligence will hit $644 billion—a staggering 76% jump from 2024. That’s not speculative money chasing moonshots. Well, actually, some of it absolutely is. But more on that in a second.
When I dug into the research for this piece (because at Cascade DM, we don’t guess—we know), I found something fascinating: More than 1,300 AI startups are now valued over $100 million each. Nearly 500 have crossed the billion-dollar mark. Ten AI-focused startups with zero total profit gained almost $1 trillion in market value in just 12 months.
Read that again. Zero profit. One trillion dollars in valuation.
If you’ve been in business longer than five minutes, alarm bells should be ringing. This is textbook bubble behavior—the kind of irrational exuberance that preceded the dot-com crash in 2000.
But here’s where it gets interesting (and where my perspective diverges from the doomsayers): This time might actually be different.
Why This Bubble Has Substance (Unlike Your Cousin’s NFT Collection)
The dot-com bubble burst because thousands of companies with zero revenue, zero profit, and zero viable business models were valued like they’d already conquered the world. Pets.com famously spent $27 million on marketing—including a Super Bowl ad—before collapsing less than a year after its IPO.
Today’s AI boom has something the dot-com era didn’t: real, profitable companies with actual cash flow betting their futures on this technology.
Microsoft. Google. Amazon. Meta. Nvidia.
These aren’t starry-eyed startups burning venture capital. They’re printing money from existing businesses while simultaneously investing hundreds of billions into AI infrastructure. Microsoft alone is spending over $10 billion on OpenAI and building data centers at a pace that would make 1950s highway planners blush.
At Cascade Digital Marketing, we see this distinction playing out in real-time with our clients. When we deploy AI-enhanced marketing campaigns—using natural language processing to analyze customer motivations or machine learning to optimize ad targeting—we’re building on infrastructure that actually exists and works.
This isn’t science fiction. It’s Saturday morning.
The Research That Changed My Mind
I’ll be honest: Six months ago, I was skeptical about AI’s near-term business impact.
Then we ran a deep-dive analysis for several clients in the professional services space (lawyers, financial advisors, consultants—the folks who make up the bulk of our clientele). We wanted to understand: Are AI tools actually improving marketing ROI, or are they just shiny objects?
The data surprised me.
AI-enhanced content analysis helped us identify emotional trigger words that increased conversion rates by 23% for one personal injury law firm. Predictive analytics reduced cost-per-acquisition by 31% for a wealth management client by identifying high-value prospects earlier in their journey. Automated competitive intelligence gave a consulting firm insights that previously required eight hours of manual research—delivered in 90 seconds.
These aren’t marginal improvements. They’re step-function changes in efficiency and effectiveness.
But—and this is crucial—they only worked because we’d done the hard, unglamorous research work first. We’d mapped customer motivations, analyzed competitive positioning, and developed clear messaging strategies. AI amplified our research-first methodology; it didn’t replace it.
This is where most of the AI bubble enthusiasm goes wrong.
The $400 Billion Question: Who’s Actually Making Money?
Here’s the stat that should worry AI investors (but excite AI users like us): Tech companies will spend roughly $400 billion on AI infrastructure in 2025—data centers, chips, model development. Yet consumer spending on AI services is only about $12 billion annually.
That’s a 33:1 ratio of investment to revenue.
One commentator quipped it’s like funding an economy the size of Singapore while currently earning revenues the size of Somalia. Gartner and BCG studies indicate fewer than 10% of corporate AI initiatives are yielding measurable financial returns today. An MIT survey found 95% of 300 AI developments hadn’t turned a profit despite tens of billions in spending.
So are we in a bubble? Absolutely.
Will it pop? Probably.
Does that mean AI is doomed? Not even close.
I think what we’re witnessing is the classic pattern of transformative technologies: massive over-investment in the short term, a painful shakeout, then steady, world-changing growth over the long term. Railroads went through this in the 1840s. Electricity in the 1920s. The internet in 2000.
The question isn’t whether AI will transform business—it will. The question is when the valuations will match the actual profits.
My bet? We’ll see a correction around 2026-2027. Weak AI startups with no path to profitability will collapse. Even the giants might see stock pullbacks if AI revenue doesn’t materialize as quickly as investors hope.
But the survivors—and the underlying technology—will emerge stronger.
The Hidden Chokepoint Nobody’s Talking About
Want to know what really keeps me up at night? It’s not whether AI works. It’s whether we can power it.
This part of our research genuinely shocked me.
AI data centers are energy monsters. They accounted for about 4% of U.S. electricity usage in 2024, and that number is expected to more than double by 2030. A single generative AI query uses 10 times the electricity of a standard Google search.
The U.S. Department of Energy released a chilling report in mid-2025: If current trends continue—retiring old power plants without fast enough replacement—blackouts could increase 100-fold by 2030.
One hundred times.
That’s not a typo.
The grid simply isn’t ready for AI’s voracious appetite. In Northern Virginia (a major data center hub), grid constraints are delaying new data center connections because substations can’t handle the load. Getting a new AI facility hooked into the electric grid now takes an average of four years, with some projects waiting up to seven years.
Microsoft’s solution? They inked a $16 billion deal to restart the Three Mile Island nuclear plant. Google is buying up hydropower agreements. Meta is building multi-gigawatt “AI campuses” with dedicated power infrastructure.
This is where the rubber meets the road for the AI boom.
If energy supply can’t keep pace—due to construction delays, high costs, or political opposition—it becomes the ultimate bottleneck. Companies might face higher operating costs or limits on scaling their AI models, which would directly curb the profits they can realize from AI services.
For business owners, this means two things:
- AI services will likely get more expensive as power costs rise
- Early adopters who figure out efficient AI workflows now will have a competitive advantage when the infrastructure catches up
At Cascade Digital Marketing, we’re already thinking about AI efficiency—not just AI capability. It’s why we focus on targeted, research-driven applications rather than throwing AI at every problem.
What This Means for Your Business (The Part You Actually Care About)
Okay, enough macro-economics. Let’s talk about what you should do with this information.
First: Don’t ignore AI because it’s overhyped. The hype is real, but so is the technology. Goldman Sachs projects AI could increase global GDP by 7%—around $8 trillion—over the next decade through productivity gains. Even if they’re off by half, that’s a massive economic shift.
Second: Don’t adopt AI just because everyone else is. Remember that 90% failure rate for AI initiatives? Most fail because companies implement AI solutions looking for problems, rather than solving specific problems with AI-enhanced tools.
Third: Focus on AI applications that amplify your existing strengths. This is the Cascade DM philosophy, and it’s why our AI-enhanced campaigns deliver an average 5.7X ROI for clients.
Here’s how we’re actually using AI in our client work:
- Deep customer motivation analysis: AI-powered natural language processing analyzes thousands of customer reviews, support tickets, and social media conversations to identify emotional patterns and pain points human researchers might miss. We then use those insights to craft messaging that resonates.
- Predictive audience targeting: Machine learning models analyze behavioral data to predict which prospects are most likely to convert, allowing us to focus ad spend on high-probability opportunities rather than spray-and-pray tactics.
- Content optimization at scale: AI tools help us test dozens of headline variations, identify semantic keyword opportunities, and ensure content answers the questions prospects are actually asking—not the questions we think they’re asking.
- Competitive intelligence automation: Rather than spending hours manually tracking competitor campaigns, AI monitors and analyzes competitor activity across channels, alerting us to strategic shifts we should respond to.
Notice what’s missing from that list? “AI-generated content we publish without human review.” “AI chatbots that frustrate customers.” “Black-box algorithms we don’t understand.”
We use AI as a research tool and efficiency multiplier—not as a replacement for strategic thinking.
The Five-Year Outlook: Boom, Bubble, or Both?
So where does this leave us heading into 2030?
Based on everything our research uncovered, I believe we’re looking at a “swoosh” pattern: continued strong growth in AI investment through 2025-2026, a market correction (possibly sharp) around 2026-2027 as expectations collide with profitability realities, then sustained, transformative growth through 2030 and beyond.
The bubble will pop. Unprofitable AI startups will fail. Even major players might see stock corrections if AI revenue lags. There could be a period—6 to 18 months—where “AI” becomes a dirty word in investor circles, much like “dot-com” did in 2001.
But the boom is real. The companies that survive the shakeout will possess genuinely valuable AI capabilities. Productivity gains will materialize. New business models will emerge. Early adopters with efficient AI workflows will dominate their niches.
Goldman Sachs, JPMorgan, and other Wall Street analysts recently argued that AI investment is sustainable and could unlock an $8 trillion productivity increase. I think they’re directionally correct, even if the timeline is optimistic.
Here’s my prediction: By 2030, we won’t talk about “AI marketing” any more than we currently talk about “internet marketing.” AI will simply be embedded in how business gets done—just as email, websites, and mobile apps are today.
The winners will be companies that:
- Adopt AI early to learn and refine workflows
- Focus on specific, measurable use cases (not AI for AI’s sake)
- Build AI on top of solid foundational strategy (research, positioning, messaging)
- Prepare for higher operational costs as AI energy demands drive prices up
The losers will be companies that either ignore AI entirely or chase every shiny AI tool without strategic coherence.
Why “Research First” Matters More Than Ever
At Cascade Digital Marketing, our tagline has always been: “Other agencies guess. We know.”
That research-first philosophy is more critical now than ever.
AI tools are incredibly powerful. They can analyze data faster than any human, identify patterns we’d never spot, and execute tasks at inhuman speed. But they can’t tell you which problems to solve or why your customers actually buy from you.
That requires human insight, strategic thinking, and—yes—old-fashioned research.
When we onboard a new client, we spend the first 30 days doing deep discovery:
- Mapping customer motivations through interviews and data analysis
- Analyzing competitive positioning
- Developing messaging frameworks rooted in psychological triggers
- Identifying the highest-value opportunities for growth
Only then do we layer in AI-enhanced execution: smarter targeting, optimized content, automated monitoring, predictive analytics.
This is how you avoid becoming part of that 90% AI failure statistic.
The firms that will win over the next five years aren’t the ones with the most AI tools. They’re the ones with the clearest strategy, amplified by AI execution.
The Bottom Line: Bet on the Boom, Prepare for the Bubble
If you take away one thing from this article, let it be this: AI is simultaneously overhyped and underestimated.
Overhyped in terms of near-term profits and stock valuations. Many AI companies are ridiculously overvalued based on speculative future earnings that may never materialize.
Underestimated in terms of long-term transformative potential. Five years from now, businesses that haven’t integrated AI into their operations will be at a severe competitive disadvantage—like companies without websites in 2005.
At Cascade Digital Marketing, we’re not waiting to see how this plays out. We’re actively researching, testing, and implementing AI tools that deliver measurable ROI for our clients today—while remaining clear-eyed about the limitations and challenges ahead.
We’re betting on the boom. But we’re prepared for the bubble.
Because in marketing, as in life, the winners aren’t the ones who guess right. They’re the ones who know—through research, data, and disciplined execution—what actually works.
And right now? When used strategically? AI works.
Ready to explore how AI-enhanced marketing can transform your professional services firm? We’re accepting a limited number of new clients for Q1 2026. Our research-first approach means we take time to understand your business before deploying any AI tools—because strategy always comes before technology.
Schedule your free strategy session and let’s talk about what’s actually possible when you combine human insight with artificial intelligence.
Josh Kilen is the founder and CEO of Cascade Digital Marketing, a research-driven digital marketing agency serving professional service firms since 2016. Based in Tacoma, Washington, Cascade DM has helped over 200 clients nationwide achieve an average 5.7X ROI through AI-enhanced, research-first marketing campaigns. When not analyzing AI trends, Josh can be found hiking Mount Rainier or perfecting his espresso technique.
FAQ
Is AI a bubble?
The author argues both: by the numbers it shows textbook bubble behavior and will probably correct, but the underlying technology is real. He points to $644 billion in worldwide AI spending in 2025, a 76% jump from 2024, and notes that ten AI startups with zero total profit gained almost $1 trillion in market value in 12 months. His bet is a correction around 2026 to 2027, followed by sustained growth.
How is the AI boom different from the dot-com bubble?
Unlike the dot-com era, when thousands of companies had no revenue or viable model, today’s AI boom is backed by real, profitable companies with cash flow, such as Microsoft, Google, Amazon, Meta, and Nvidia. They are funding AI infrastructure that already exists and works, including Microsoft’s reported spending of more than $10 billion on OpenAI.
Why do most corporate AI initiatives fail?
The post cites Gartner and BCG findings that fewer than 10% of corporate AI initiatives yield measurable financial returns today, and an MIT survey that found 95% of 300 AI developments hadn’t turned a profit. The author says most fail because companies adopt AI looking for problems to solve rather than applying it to specific problems on top of solid strategy and research.
What energy risk could limit the AI boom?
AI data centers accounted for about 4% of U.S. electricity usage in 2024, a figure expected to more than double by 2030, and a single generative AI query uses ten times the electricity of a standard Google search. The post notes a 2025 Department of Energy report warning blackouts could increase 100-fold by 2030, and that connecting a new AI facility to the grid now takes about four years.
