To Everyone Panicking About 22-Year-Old Billionaires: MIT's '45 Is Peak' Finding from 2.7M Companies, and 3 Battlefield Maps for AI × Experience Capital

img: A horizontal 5-stage timeline flow diagram. From left: “(1) Panicking over 22-year-old billionaires” → “(2) MIT research reveals 45 is peak” → “(3) Grasp the 3 battlefield maps” → “(4) Move with the 3 steps for those over 35” → “(5) Win with experience capital × AI.” Small arrow icons between each stage, only the final stage emphasized in rose | type: eyecatch | style: Horizontal timeline sequence diagram, white background (#F5F5F5), rose (#c2185b) accents, arrows between each step, 5-stage flow readable at a glance

“Three 22-year-olds became billionaires.” “An 18-year-old founded a unicorn.” Every time I saw headlines like this, honestly, I panicked too.

“Am I already too late?”

But then I read a single piece of research, and the entire landscape flipped. It’s a paper by Professor Pierre Azoulay and his team at MIT (Massachusetts Institute of Technology) Sloan School of Management. They analyzed 2.7 million entrepreneurs using U.S. Census Bureau data. The conclusion went way beyond what I expected.

The average age of the most successful startup founders was 45.

Not 22. Not 29. Forty-five. And the data even shows that a 50-year-old founder is roughly twice as likely to succeed as a 30-year-old founder (MIT News, 2020).

Headlines never pick this up, you know? “22-year-old billionaire” makes a better headline. So today, for everyone panicking about the Gen Z youth wave, I’m laying out the other side of the data. I’ll organize “where veterans are structurally stronger” into 3 battlefields.

I’ve written before about the average age dropping from 40 to 29 and about the 22-year-old billionaires. This piece is the follow-up, fully focused on “okay, so where exactly should those over 35 fight?"

"25 Is the New 30” Is Only Half True

First, so there’s no misunderstanding: the Gen Z youth-wave data itself is real.

The starting point is a report released on January 7, 2026 by Antler, a major global VC (venture capital) firm. The data showed that the average age of AI unicorn (unlisted AI companies valued at $1 billion or more) founders dropped 11 years, from 40 in 2020 to 29 in 2024. Fortune ran the headline “25 is the new 30” (Fortune, 2026-01-07). CNBC covered the same data prominently (CNBC, 2026-01-17).

That much is fact.

But the same report contained another number that nobody put in a headline.

The average age of “non-AI” unicorn founders rose from 30 in 2014 to 33 in 2024.

In other words, the youth wave is happening only in AI. Across startups as a whole, the trend is actually toward “founding later in life.” This isn’t some “Gen Z is taking over the world” story. It’s just that in the special domain of AI, a structure happened to emerge where young people can produce results in a short time.

Layer the MIT research I mentioned on top, and the picture gets even more three-dimensional.

The Azoulay team’s research was posted to NBER (National Bureau of Economic Research) in 2018. The peer-reviewed version appeared in American Economic Review: Insights in 2020 (NBER w24489). They pulled 2.7 million founders from U.S. Census Bureau business statistics. They analyzed the correlation between age at founding and “success” (high growth, IPO, acquisition).

Here are the paper’s key findings.

• Average founding age of the most successful founders (top 0.1% by growth rate): 45.0 years
• Top 1%: 43.7 years
• Top 5%: 42.1 years
• A 50-year-old founder is about twice as likely to succeed as a 30-year-old founder
• A 60-year-old founder shows a higher success rate than founders in their 20s
• The conclusion holds even when restricted to high-tech, startup hubs, and successful exits

In other words, the data flatly contradicts the Silicon Valley myth that “youth is a weapon.”

img: A side-by-side comparison diagram of average founder age in the AI domain vs. all industries. Left column: “AI domain: 40 in 2020 → 29 in 2024 (down 11 years, red downward arrow).” Right column: “Non-AI: 30 in 2014 → 33 in 2024 (up 3 years, blue upward arrow).” Center bottom prominently displays “Average age of most successful founders: 45.0 (MIT research)” | type: comparison | style: White background (#F5F5F5), rose (#c2185b) base, color-coded left/right (AI in reds, all industries in blues), the “45” in the center emphasized in bold

What this reveals is the following structure.

AI domain = proximity to the latest tech matters more than know-how = young generation temporarily favored Non-AI = industry experience, networks, and track record matter = veterans structurally stronger

In other words, the “youth wave” is a localized phenomenon. Across startups as a whole, the structural advantage of veterans hasn’t crumbled.

The problem is, AI news is all you see, so it’s easy to mistake this for “Gen Z is strong across all founding domains.” I thought the same thing at first. When your social media timeline fills up with Gen Z AI founders, your own age suddenly starts to feel like a burden.

But look at the data calmly, and there are domains where, if you choose the battlefield right, veterans are overwhelmingly stronger. Let me organize those into three.

”3 Structural Reasons” Veterans Are Strong

Before talking battlefields, let me lock down three structural reasons veterans are strong.

Reason 1: Industry-experienced founders succeed at 2x the rate

A particularly important finding from the Azoulay paper is the effect of industry experience. Founders with 3+ years of work experience in the same domain were 2x more likely to reach the “top 0.1% high-growth company” tier compared to those without (HBR, 2018).

This makes intuitive sense. People who’ve felt firsthand “what the problem is,” “who the bottleneck is,” and “where the money flows” in an industry. When they design a solution, the precision is dramatically different from the start. There are clearly domains where ten Gen Z genius programmers bundled together can’t beat one veteran with the industry’s tacit knowledge.

Reason 2: B2B SaaS founders cluster in their 40s

According to SaaStr’s aggregated data, in the B2B SaaS world (subscription software services for businesses), the average founder age at Series A is 41. At Series B, it’s 43 (SaaStr).

Why? Because B2B SaaS sells to “industry decision-makers.” The difficulty of a 20-something programmer pitching a hospital decision-maker, vs. a veteran with 20 years in healthcare pitching the same person — completely different. Enterprise sales is a trust game. Building “I can entrust this to this person” depends directly on years in the industry.

Reason 3: AI amplifies “experience density”

This is the most important point in this whole piece.

Before AI, even when veterans had knowledge, turning it into a product was time-consuming. The walls of “I can’t code, I can’t design, I’m weak at marketing” were real.

Now that AI handles all of that, a veteran’s “experience density” gets directly leveraged. The “the one who acts wins” mindset I’ve always preached works here too. The moment a person with experience starts using AI, they can build a valuable service several times faster than a Gen Z’er starting from zero with AI.

There’s data showing SMBs have a 91% chance of recouping AI investment within a year. The more industry experience you have, the more easily AI usage connects to “the core of the work.” The recoupment rate likely skews even higher.

In other words, AI isn’t the veteran’s enemy — it’s the veteran’s strongest weapon.

The reason Gen Z gained an advantage isn’t that they’re special. It’s that AI commodified the parts where veterans used to dominate (code, design, operations), so younger people could finally step onto the same arena. Meanwhile, in the parts that haven’t been commodified — industry knowledge, networks, trust, judgment born of experience — the veteran’s edge is unshaken.

Now to the main argument. So where, specifically, should those over 35 fight? From my client work and my own experience, here are the three battlefield maps.

Battlefield Map ① The “Translator Position” of Industry × AI

The strongest battlefield is this one.

The world has plenty of “people who get the industry” and plenty of “people who get AI.” But the number who get both is shockingly small. This empty zone is the veteran’s main battlefield.

Concretely, coordinates like these.

• 15 years in healthcare × designing diagnostic-support tools with Claude/ChatGPT
• 10 years in real estate × automating property valuation with AI
• Manufacturing procurement experience × optimizing supplier selection with AI
• Law firm administration experience × contract review with AI
• Care facility floor experience × shift optimization + family reporting with AI

These combinations are physically impossible for Gen Z to assemble. Because “15 years in the industry” is the prerequisite. A 22-year-old programmer simply can’t understand healthcare-floor workflows at a 15-year resolution. Conversely, a healthcare veteran of 20 years can pick up AI prompting basics in two weeks, frankly.

In other words, the “rarity” of the combination overwhelmingly favors the veteran side.

What’s more, on this battlefield, rip-off Gen Z consultants don’t make it through the door. Because even when they pitch, decision-makers turn them away with “you don’t even know our industry.” Decision-makers contract with independent consultants from within their own industry. This becomes a structural barrier to entry that protects veterans.

Three concrete things to do.

1. List “the routine work everyone hates” in your industry
2. Build a prototype that automates that work with AI (Claude, ChatGPT, Notion AI, etc.)
3. Demo it to former colleagues and former clients, and hand it over for free with “give it a try”

Free is fine to start. If even one company adopts it, that becomes a case study. Once you have a case study, the next client comes through industry connections. This is the veteran’s royal road.

Gen Z does things like “throw up an LP (landing page) and run social ads to acquire customers.” Veterans don’t. They land the first client through network and trust. This is overwhelmingly faster.

img: A 2-axis matrix diagram of “industry experience” and “AI skill.” Vertical axis = industry experience (high/low), horizontal axis = AI skill (high/low). Four quadrants: “(1) Industry veteran + AI weak (many),” “(2) AI strong + doesn’t know industry (Gen Z type),” “(3) Bottom zone (general public),” “(4) Translator position — strongest territory (few, the sweet spot).” Only quadrant (4) is emphasized in rose | type: diagram | style: White background (#F5F5F5), 4-quadrant matrix, only quadrant (4) filled in rose (#c2185b), relative dot sizes in each quadrant express the population scale

Battlefield Map ② Automating Your “Customer Assets”

If Map ① is “going outward,” Map ② is “going deeper inward.”

Veterans have an asset Gen Z couldn’t build in any number of years. The existing customer list, network, and trust relationships. These are pure time-built assets that only accumulate with years.

The people you exchanged business cards with during your corporate days, the companies you’ve previously done business with, the people who’ve followed you on social media for 5 years. All of these are assets only veterans hold.

Battlefield Map ② is the strategy of automating and extending these assets with AI.

Concrete examples.

Example A: Veteran former salesperson

• 300-company customer list from the past 5 years
• AI auto-collects each company’s latest news, hires, and earnings
• Once a month, send “Given this move at your company, here’s a proposal we can make”
• 30% of existing customers reply “let’s talk” → consulting contracts worth several million yen monthly

Example B: Veteran from a former marketing department

• 100 marketing contacts among former colleagues and former clients
• AI analyzes each company’s social posts and auto-generates improvement suggestions
• A ¥30,000/month “AI assistant for your marketing team” service
• 50 contracts = ¥1.5M/month

Example C: Veteran former recruiter

• A list of 500 candidates from past interviews
• AI infers candidates’ career updates and likely job-change timing
• Sends hiring companies a monthly “this person looks ready to move”
• ¥500K per placement, 3 per month = ¥1.5M/month

This strategy is absolutely impossible for Gen Z. They physically can’t hold “5 years of customer lists,” “100 former colleagues,” or “500 interviewed candidates” in their 20s.

What’s more, since these are “people you already have trust with,” customer acquisition cost (CAC) is essentially zero. You can build an economic structure that doesn’t depend on marketing.

The key here is to design so that the list is “run by AI.” Manually following 300 companies a month is impossible. Only by automating with AI does it become a business one person can run. AI is your weapon here.

Did you keep the business card files from your corporate days? If you threw them out, that’s a real, real shame. I have a client who hit ¥1M/month within 3 months of going independent, and his biggest asset was “the client list from his corporate days.”

Battlefield Map ③ Regulated, High-Ticket “Who Can Take Responsibility” Game

The third battlefield is a domain Gen Z structurally cannot enter.

Healthcare, finance, legal, real estate, pharmaceuticals, construction, public sector — these are regulated industries. Licenses, credentialed professionals, and track records are required, and the bar to new entry is one notch higher.

And it’s exactly here that AI utilization is most strongly demanded. Diagnostic efficiency, contract review, property valuation, architectural design — full of work that gets 10x faster with AI.

But in this domain, “someone who can take responsibility for what AI produces” is essential.

A doctor takes responsibility for AI diagnoses. A tax accountant signs off on AI tax calculations. A lawyer approves AI contract reviews. Without this, the field can’t actually use it.

In other words: credentialed professionals, former credentialed professionals, and people with 10+ years of practical experience in regulated industries. They can monopolize the battlefield as “the final approver of AI.”

Three concrete patterns.

Pattern ①: Former accountant — AI bookkeeping outsourcing

• CPA license + AI bookkeeping automation
• ¥30,000/month, fully outsourcing SMB accounting
• 50 companies solo = ¥1.5M/month

Pattern ②: Former legal staff — AI contract review

• 10 years of legal experience + AI for contract drafts and risk extraction
• ¥30,000/contract spot review
• 50 contracts/month = ¥1.5M

Pattern ③: Former banker — AI business plan consulting

• Bank loan-screening experience + AI to auto-generate business plans
• Shape them so banks approve
• ¥200,000/case, 10 cases/month = ¥2M

None of these are doable by 20-something Gen Z. You can’t acquire credentials and track record except through years.

The veteran’s “license × AI” combination will probably remain one of the strongest battlefields for at least another 10 years. Because regulatory loosening takes time. As long as the law says “a credentialed professional’s signature is required,” Gen Z can’t enter.

This isn’t a conservative way to fight. If anything, “credentialed professionals who maximize AI for efficiency” can take the most innovative position within their industry.

3 Steps for Those Over 35 to Start Today

If you’ve read this far thinking “okay, so what do I start with?” — here are all three steps.

Step 1: Inventory your experience capital (time required: 90 minutes)

Take pen and paper to a café. Leave the phone behind. Write down these three things.

1. Every “industry/role I had 3+ years of experience in” across your career
2. Within those, every “task I could do faster than other people”
3. From your network back then, the list of “people I can still reach today”

A lot of people stop here and think “I haven’t really done much.” Wrong. Don’t compare to others — compare to the inexperienced. You can compare your current self to your 20-something self. The fact that you can now do everything that “stumped you on day one” — that’s experience capital.

Step 2: Decide one angle to automate with AI (time required: 1 week)

From the list in Step 1, pick one based on these conditions.

• Recurring as ongoing work (once-a-year doesn’t count)
• You can do it fast (2x+ the speed of others)
• AI can partially automate it (handing it all to AI is impossible)
• Worth tens of thousands to hundreds of thousands of yen per case (too cheap and you can’t do enough volume)

For example: “monthly automation, with AI, of the prospect-list analysis I’ve done for the past 5 years.” Or “automate 70% of monthly closing work at accounting firms with AI.”

If you get greedy and try to “automate every task,” you’ll fail. Narrow to one. The one who acts wins, but you can’t hit anything without aiming.

Step 3: Sell to the first one person (time required: 2 weeks)

From Step 1’s list, pick three “people I can still reach today” and say this.

“I built this thing recently with AI — would you try it and give me feedback? Free for the first 3 months.”

You don’t need an LP or social ads here. Land your first client through your network — that’s the veteran’s royal road. Once one person uses it, build a case study from them and spread it through the industry.

If you reach out to three people and not one bites, return to Step 2 and rework the service. Normally, 1-2 of 3 will bite. That’s how much demand exists for “industry × AI” services.

img: A horizontal flowchart laying out the 3 steps for those over 35. Step 1: “Inventory experience capital (90 min, pen and paper)” → Step 2: “Decide one AI automation angle (1 week, 4 conditions)” → Step 3: “Sell to the first one person (2 weeks, 3 people from network).” Each step shows time required and concrete actions, ending with a goal marker “First contract in roughly 1 month total” | type: diagram | style: Horizontal flow diagram, white background (#F5F5F5), rose (#c2185b) accents, 3 steps connected by arrows, checkbox-style sub-items beneath each step

Wrap-Up — Don’t Panic and Pretend to Be Young

Don’t panic at the 22-year-old billionaire news. That news is fact. But it’s a phenomenon in “the special domain of AI,” not the story of founding overall. The conclusion MIT pulled from 2.7 million people is unambiguous. The average age of the most successful founders is 45. A 50-year-old is twice as likely to succeed as a 30-year-old.

That’s the reality the data shows.

So don’t try to imitate Gen Z’s playing style. Their strengths are “AI native,” “zero preconceptions,” and “fast moving.” These are the privileges of the 20s. If those of us over 35 fight on the same arena, we can’t win.

Instead, choose a battlefield where you fight with the veteran’s strengths. All three battlefield maps I laid out today are domains Gen Z physically cannot enter.

• Battlefield ①: The translator position of industry × AI
• Battlefield ②: Automating customer assets
• Battlefield ③: The regulated, high-ticket “who can take responsibility” game

All three are strategies that leverage “assets that only accumulate over years” using AI. This is the veteran’s main battlefield. For those who want to set their scale even larger, the Soonicorn ($500M-$999M zone) framing as a new intermediate goal is also worth a look.

I’ve been independent for several years myself, and “experience from my corporate days” is still what works the most. New customers I picked up via social media still rank below the people I exchanged business cards with 5 years ago, in both pricing and trust. Experience is an asset, and you can’t throw it away.

The moment you think “am I already too late,” don’t ride that panic — look at the data. 45 is peak. 50 is twice 30. Your “years” are, from now on, your largest possible leverage.

There’s no reason not to act. The one who acts wins. The battlefield map only starts functioning as a real map for those who actually move.

Reaching out to those first three people — let’s put it in this week’s calendar.