$KPG.ax using AI one can deduce following:
1/
Public markets have re-rated global accounting & professional services down 30–65% on AI fears in 12 months.
Private capital is paying full price.
One of these markets is wrong.
2/
The AI re-rating, 12 months to May 2026:
• KPG: -65%
• Wolters Kluwer: -58%
• Judges Scientific: -54%
• Constellation Software: -44%
• CBIZ: -43%
• Accenture: -42%
• Intuit: -29%
The cohort: anything that bills hours, sells professional info, or templates knowledge work.
3/
Striking: Constellation Software — the gold standard of decentralised acquisition compounding — is down ~44%.
Q1 2026: 20% revenue growth, record capital deployment.
The selloff isn’t fundamental. It’s category sentiment.
4/
The AI-immune serial acquirers tell the real story:
• Halma (safety sensors): +47%
• Brown & Brown (insurance brokerage): flat-positive
• Diploma (industrial distribution): held up
The serial acquirer model hasn’t been re-rated. End-markets perceived as AI-exposed have.
5/
Meanwhile, on 5 May 2026:
New Mountain Capital acquired Grant Thornton Australia for A$1B.
• ~3.3x revenue
• 15x EBITDA
• Zero discount for “AI disruption”
One of the most sophisticated PE firms in the world. A$1B of conviction.
6/
Not isolated. PE activity in mid-market accounting has accelerated globally:
• New Mountain (GTA)
• Cinven, Bain, H&F, TowerBrook — all active in accounting
None pricing AI displacement. Multiples consistent with pre-AI-rerating public comps.
7/
Apply the benchmark to KPG:
• ~9x forward EBITDA (public)
• 15x EBITDA (private market reference)
• ~60% discount to implied private market value
Same business model. Same end-market. Same asset class. Different price.
8/
Public markets have priced AI displacement as the base case.
Private acquirers, with more diligence resources and longer horizons, have priced it as not.
One of these markets is wrong.
9/
Pattern is familiar:
• 1990s outsourcing fears
• Post-Enron consulting selloff
• Dot-com media collapse
• Post-GFC financial services
Each produced a multi-year window where patient capital was rewarded for ignoring consensus pessimism.
10/
Whether AI impairs accounting revenues is an open empirical question.
What’s not open: the public market has priced that scenario with conviction. The private market has not.
The price is on the screen. The benchmark is public. The arithmetic is straightforward.
1/
Public markets have re-rated global accounting & professional services down 30–65% on AI fears in 12 months.
Private capital is paying full price.
One of these markets is wrong.
2/
The AI re-rating, 12 months to May 2026:
• KPG: -65%
• Wolters Kluwer: -58%
• Judges Scientific: -54%
• Constellation Software: -44%
• CBIZ: -43%
• Accenture: -42%
• Intuit: -29%
The cohort: anything that bills hours, sells professional info, or templates knowledge work.
3/
Striking: Constellation Software — the gold standard of decentralised acquisition compounding — is down ~44%.
Q1 2026: 20% revenue growth, record capital deployment.
The selloff isn’t fundamental. It’s category sentiment.
4/
The AI-immune serial acquirers tell the real story:
• Halma (safety sensors): +47%
• Brown & Brown (insurance brokerage): flat-positive
• Diploma (industrial distribution): held up
The serial acquirer model hasn’t been re-rated. End-markets perceived as AI-exposed have.
5/
Meanwhile, on 5 May 2026:
New Mountain Capital acquired Grant Thornton Australia for A$1B.
• ~3.3x revenue
• 15x EBITDA
• Zero discount for “AI disruption”
One of the most sophisticated PE firms in the world. A$1B of conviction.
6/
Not isolated. PE activity in mid-market accounting has accelerated globally:
• New Mountain (GTA)
• Cinven, Bain, H&F, TowerBrook — all active in accounting
None pricing AI displacement. Multiples consistent with pre-AI-rerating public comps.
7/
Apply the benchmark to KPG:
• ~9x forward EBITDA (public)
• 15x EBITDA (private market reference)
• ~60% discount to implied private market value
Same business model. Same end-market. Same asset class. Different price.
8/
Public markets have priced AI displacement as the base case.
Private acquirers, with more diligence resources and longer horizons, have priced it as not.
One of these markets is wrong.
9/
Pattern is familiar:
• 1990s outsourcing fears
• Post-Enron consulting selloff
• Dot-com media collapse
• Post-GFC financial services
Each produced a multi-year window where patient capital was rewarded for ignoring consensus pessimism.
10/
Whether AI impairs accounting revenues is an open empirical question.
What’s not open: the public market has priced that scenario with conviction. The private market has not.
The price is on the screen. The benchmark is public. The arithmetic is straightforward.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
