Goldman Sachs buys companies in the $500 million to $2 billion range and puts more than 110 operational partners behind them. Most firms competing for those deals cannot hire one such person, let alone a bench of them.
The usual middle-market pitch is proprietary sourcing and a local network. Michael Bruun's pitch is the balance sheet, the 45,000 employees and the vendor relationships of the bank itself, rented out to companies far too small to command them on their own.
"Nobody else has Goldman Sachs as their wingman or wing woman in their investment strategy."
Bruun has spent nearly 23 years at Goldman Sachs and is now its global co-head of private equity, which means he is the person deciding what a firm that both advises and finances companies should do when it buys them instead.
The full interview is covered here so you can skip it. 38 minutes of audio, 17 minutes of reading.
Here are the 11 insights that matter.
👤 Guest: Michael Bruun, Global Co-Head of Private Equity at Goldman Sachs, who has spent nearly 23 years at the firm and invests in middle-market buyouts of $500M to $2B
🎙️ Host: David Weisburd, who hosts How I Invest and co-hosts The 10X Capital Podcast
📰 Published: 16 September 2026 on YouTube (How I Invest with David Weisburd)
🔴 YouTube | ⏱️ 38 min | ✅ Time saved: 21 min
Key Takeaways
Goldman's middle-market edge is 110-plus operating partners the target companies could never hire directly
The operators came for the chance to work with Goldman Sachs again, not for the company size
Talent compounds and ideas do not, which is why Bruun ranks a hire above a customer introduction
Every AI project in the portfolio falls into one of two buckets: scaling revenue or improving margin
His call-center example is Jevons paradox in practice — the same headcount, far more calls handled
The firm hedges base rates years into the future so that investors are only paying for value creation
A clean data set is now a due-diligence item and gets priced into the return model
One company arrived with 25 separate ERP systems and no data lake
Goldman's direct private equity funds have mostly skipped continuation vehicles, because strategic buyers pay more
Bruun has stopped asking whether a portfolio CEO believes in AI and started asking whether the rest of the leadership team does
AI is pulling valuation multiples in different sectors toward each other, because the same playbook now works in all of them
1. Skin in the Game
Weisburd opened on Lloyd Blankfein's phrase about skin in the game, and why it shaped Bruun's early career at the bank.
Bruun's reading is that investing alongside a client is a different relationship from advising one. Goldman Sachs advises, arranges financings, makes markets in stocks, bonds and commodities, and invests its own money — and the last of those changes the other three.
"And I think what Lloyd meant when he mentioned skin in the game was that we can be several things for clients and give an even more comprehensive service." He said very few firms can do it and none at Goldman's scale.
He pointed readers at Blankfein's recent book, Street Wise, which he said sets out the same point.
On incentives inside his own team, the answer is carry plus personal money. Compensation is annual pay plus carried interest in the fund, and the firm asks its own people to invest in the funds alongside outside investors.
"We think that carry is important to incentivize great investment outcomes but we also think that you need to have a lot of skin in your game at the most personal level."
2. Goldman as the Wingman
Weisburd put the competitive problem directly: the core middle market, $500 million to $2 billion, is the most contested part of private equity, and Bruun has formidable competitors in it.
"Nobody else has Goldman Sachs as their wingman or wing woman in their investment strategy." He conceded the competition first, then gave three answers.
The first is sourcing and value creation through the bank's network.
The second is headcount: "We have more than 110 operational partners that drive value creation in our strategy."
The third is information. "We're 45,000 employees at Goldman Sachs." The firm is in most markets in the world, and Bruun said those signals flow back into the investing business — which matters most, on his account, in a period of geopolitical uncertainty and a blurrier macroeconomic picture.
The vehicle for the operators is what the firm calls its value accelerator, built over the past ten years and organized into centers of excellence. The idea was to take the operational resources normally found in mega-buyout funds and apply them in the middle market.
Why senior operators agreed to work on smaller companies is a Goldman-specific answer: many had worked with the bank before as clients — it had advised them or financed them — and coming back was a way to stay with the institution while doing operating work with visible effect.
3. Talent Beats Ideas
Weisburd described the model as taking an executive a middle-market company could never recruit and fractionalizing that person across several portfolio companies, giving each one 10% or 20% of someone otherwise out of reach.
Bruun's addition is that the executive arrives with a network: "But I would say that talented person, that talented executive usually has a network or he or she has a network and so they will bring in even more talent. They're a talent magnet."
The first effect is on the existing management team, which sees what good looks like in a specific part of the playbook. Sometimes the management team is changed; often it is simply shown a standard.
Ranked against the other things a sponsor can hand a company, talent wins. "Giving them a new customer, that's very powerful, or giving them a great introduction." And then: "But giving them talent is what really moves the needle."
The exchange's argument for why is about attention rather than insight. A convincing pitch about an AI stack survives about an hour before the next priority displaces it. A person hired to do that work spends every week on it, and that compounds in a way an idea does not.
Bruun added that the same holds away from the visible parts of a business — the supplier relationships and back-office work customers never see. His examples of access that changes outcomes are the leading large language model vendors now and critical suppliers during the pandemic.
4. Volatility Favors Veterans
Asked why the operating playbook works better in volatile periods, Bruun said it is about decision speed resting on experience.
The playbook works best in volatile times, he said, "because you need to make decisions faster and those decisions needs to be made on a sound set of observations and based on a lot of experience."
His formulation is binary: either you have been through a crisis or you have not, and one crisis prepares you for the next. "We learn more as humans when we're in volatile times."
"Best example right now we're faced with maybe the most consequential moment in private equity driven by AI." Nobody has prior experience of AI specifically, he said, but senior operators have lived through large shifts in technology road maps and business models.
He called the upper middle market a talent arbitrage. The operators he works with have run bigger things, so they bring a larger toolbox to a smaller company — and he said they find it more fun, because decisions move faster there.
Bruun's personal version of the same point is that the usual career arc is to become, each year, slightly closer to the oldest person in the room. Twelve to fifteen years into his career that reversed: building the operating group surrounded him with people who had seen more than he had.
Weisburd's framing was the venture diligence question of whether a manager was around in 2008. He said he started his own career that year, so he could answer yes, but that thirteen or fourteen years of bull market followed.
5. Rates Are Back to Normal
Bruun's own crisis timeline runs through two of them before the long bull market.
"I hit the great financial crisis I guess four years into my career and that sort of reset a lot of expectations." Working mostly in Europe, he then went into the 2011 sovereign debt crisis, and only after that came the decade-long run to 2021 and 2022.
What ended, on his account, was the reliance on multiple expansion and cheap debt. Value now has to come from compounding earnings and cash flow, which he called going back to basics and said has been healthy for the industry.
His claim on where rates actually sit: "If I add the base rate that we're experiencing in most places in the world plus the credit spreads, it looks much closer to the average that you have seen for several decades than anything that we obviously experienced in that bull market."
Weisburd put the stronger version — that pricing in falling rates was the industry's mode of operation for a decade and is a strategy no one should run. Bruun did not accept the premise that firms modeled falling rates; he allowed that the industry grew complacent about rates staying low through a long period of quantitative easing.
What he does now is hedge, aggressively and far out. He said the firm hedges base rates for many years ahead, and tries to take out regulatory risk, credit-spread risk, foreign-exchange risk and supply-chain risk. He was explicit that nothing can be hedged forever — three to five years is the realistic window, and he does not assume more.
The reason for stripping all of that out is what is left. "So what we're left with and what investors in our strategies are paying for is our raw value creation capability and our raw strategy capability."
6. What Strategics Pay For
Weisburd asked whether assets sell better to strategic buyers simply because profits are higher, or because of something the spreadsheet does not hold.
Bruun's answer is that the decisive items are not in the financial statements: "I think there's a lot that's not in the spreadsheet like one of the most important things that you can't actually see in financial ledgers right now is does this company have a good tech stack?"
"Does this company have a homogeneous set of data? Is it a clean data set? These are some of the things that will be very consequential for the way companies adopt AI."
Those items are now a diligence line and a budget line. What has to be changed, and what fixing it costs, is written into the return model before the deal is signed.
His example of what bad looks like is a company with 25 separate ERP systems and no data lake. Fixing that over the holding period is what he means by future-proofing a company, and it is why a strategic buyer values it more highly afterward — though he said it is no guarantee.
Carve-outs are the case he likes most, and for a reason that sounds like a drawback. A business being carved out has usually been strategically neglected, so a great deal is below the standard he wants.
"It's an opportunity to put the best tools into the company instead of changing a lot of legacy stuff." Weisburd called it a forcing mechanism, which Bruun accepted.
The cost shows up as a J curve. The work is real money out the door during the early period, in exchange for faster growth and better margins afterward.
7. Continuation Vehicles
Weisburd quoted Mark Sotir, his own former partner who now runs a family office, on what Sotir called the absurdity of the private equity model: grow an asset for three or four years, spend the last year dressing it up, then sell it to a competitor. The show's own figure was that continuation vehicles have reached $110 billion.
Bruun started with the emotional case for them, which he said he has felt himself. Deal teams fall in love with an asset they have built and would like to own it forever, partly because then they do not have to go and buy another one.
The constraint is the fund's own contract: "The thing is that your investors is entering into a timebound relationship with you." Distributions have to come along the way, not all in year ten — and with colleagues' money and the bank's balance sheet in the fund, he said there are plenty of people reminding him of that.
Goldman runs large businesses that buy into continuation vehicles, and Bruun was complimentary about them — the advantage there is choice, because those teams can pick among many.
His own direct private equity strategies have mostly not used them, and the reason is price: "It's just that we have mostly sold our assets to strategic buyers who acquired those assets with synergies and therefore we're able to pay the highest price for the asset."
He named a simpler alternative he likes: selling a minority stake to a partner, which takes some money off the table while keeping the firm invested and can bring in a partner who helps with acquisitions.
"So nothing against CVs I just think with most things in life you should diversify your sources of liquidity."
Weisburd's counter was that a competitor is implicitly a strategic buyer, and one that will value an asset above its intrinsic value — which is close to the definition. Bruun agreed that a strategic with synergies should usually prevail, and said he does not expect a significant share of his exits to run through continuation vehicles based on the pipeline he can see.
8. AI Is Revenue or Margin
Asked what AI adoption inside portfolio companies actually looks like, Bruun started with the chief executive and then gave a deliberately crude framework.
"I have not met a company where AI wasn't important." He allowed that some blue-collar businesses may be undisruptable in their core work, but said how a company finds customers and what it learns from its own operations still apply.
The advice to portfolio companies is to start narrow: "Don't try and spread AI anywhere in the business in an uncontrolled way." He told them to focus on a finite set of tangible outcomes first, for safety and security reasons as much as delivery ones.
"What we've seen is that AI falls into fundamentally two categories. Either it's scaling revenue or it's improving your margins." He said people will call the framework too basic, and defended it on the ground that growth at higher margins is what the firm wants either way.
On revenue, the common ground across sectors is client service — more relevant sales conversations, and surfacing problems the client has.
His call-center example is Jevons paradox in practice. Contact centers, he said, have roughly the same number of people as before: "They're just handling way more calls because they're now AI enabled."
The concrete case is an eyewear business losing appointments. "We have a company that is in the contact lens space or spectacle space and they have noticed that many of their clients didn't showed up at the scheduled appointment time." A morning reminder call fixes two things at once: staff are not left idle, and "And secondly, every time you enter the store, it's an opportunity to provide better service to you and potentially do a sale of contact lenses or products."
The playbook then travels, which he thinks is the real change. Selling contact lenses turns out not to be very different from selling insurance or cybersecurity, so the same approach moves quickly from one portfolio company to the next.
The second-order effect is on valuation. "For many years, if you were a healthcare investor, you didn't speak to the tech investor or you didn't speak to the financial services investor. Because of AI, all investors can learn from each other." He said that is producing convergence in valuation multiples across sectors, on the argument that a software company and a financial services company with the same growth, margin, cash flow and acquisition potential should not be priced very differently.
On margins, the gains are in repetitive work and in code, the latter because the firm owns companies that write their own software.
His conclusion is that scale will separate the platforms: "I think means that the platforms that have the capabilities to invest the most into these resources should have breakout performance." Goldman's own claim to that scale is purchasing power — "We are a customer with 12,000 engineers at Goldman Sachs" — plus a growth equity business, a private equity business and newly formed companies built to help mid-sized firms adopt AI.
9. Adoption Starts at the Top
Weisburd pressed on implementation: a dedicated AI team, or every employee spending a share of their time on it?
Bruun said the answer varies by culture but the starting point does not. "The CEO needs to make it incredibly well understood by all employees at the relevant firm that AI is an absolute must and it's something that will change the way we do our business and we embrace that change."
The firm built a school for it: "We have stood up an AI university where we're literally taking the CEOs of the portfolio through that university." He described a short curriculum whose purpose is to make the chief executive a better conversational partner for their own board and staff.
He pointed to Goldman's own internal program, One GS 3.0, as the same message delivered to its own people.
Implementation has to run in both directions, and the bottom-up half needs room to fail. In a secure environment, he said, a thousand flowers have to bloom: "if humans are not allowed to experiment it is unlikely that we will get to the most relevant and most creative solution."
The uncomfortable part is the leadership team, not the chief executive. "At the same time, I think CEOs are also faced and this might at times be uncomfortable with the task of deciding whether the rest of their XCOM or leadership team is actually ready for this transformation."
His two tests are specific: is the chief technology officer experimenting with AI, and is the head of sales using AI tools to find cross-selling opportunities rather than working the way the company did before late 2022. People who do not embrace the change, he said, are unlikely to remain the relevant leaders — and he said the firm is actively having that conversation with its chief executives.
Then the bottleneck moves to supply. A chief technology officer who wants to implement a model still has to get access to it, and Bruun said getting a management team onto a vendor's map is where Goldman's name does the work — the vendor is servicing that company in partnership with the bank.
The demand for this is visible in attendance. He said seminars with management teams and with limited partners draw far bigger audiences and far more follow-up when the topic is AI.
10. The War for Talent
Weisburd restated Jevons paradox in his own terms — a division between what humans are good at and what AI is good at — and then put a second paradox to Bruun.
Weisburd's claim: "There's also another paradox which is the faster the technological change, the less the technology matters and the more it's the culture." Because AI is prompted in English, he said, everyone starts with the same basic operating principles, so the most flexible and ambitious team eventually beats a more AI-native but rigid one.
"So the question becomes not which AI tool do I implement but which leadership do I implement that will find the AI tool because the AI tools are changing every single day." Optimizing on the tool, he said, leaves you behind at the next technological step six months later.
Bruun agreed and gave the multiplier that makes it urgent: "In a weird way, it used to be that a great employee operating a great culture could achieve a lot, but just think about what a great employee with nine agents can achieve, right?"
His conclusion is the line the episode is named for: the war for talent "is bigger than it's ever been." The right person with a change mindset moves faster because they are working with AI, not merely because they are fast.
The test he applies to a chief executive is whether they are waiting it out — "are they themselves embracing this change or do they keep their head in the sand and hope that AI will be over in a few quarters from now because it's not."
He widened the point beyond technology. Political agendas and market volatility are also moving faster, in his view partly because everyone has access to more information than three years ago, and the people comfortable operating in constant change are the ones who will run these businesses.
What he does not think changes is the value set. "We always emphasize things like partnership, client service, integrity, excellence. I don't think AI is going to change any of that."
11. Enjoy the Journey
Weisburd closed by asking for one piece of timeless advice Bruun would give himself at the start of 23 years at Goldman Sachs.
His answer was about attention, not strategy. Early in a career the focus is entirely on reaching the next destination.
"If I could give myself an advice today, enjoy the journey a little bit more and learn from that journey."
He tied it back to the rest of the conversation. Being present in the work, he said, is also what produces innovation, because the AI question is mostly about how to make what you are already doing better or reimagine it entirely.
Bonus Insights
Weisburd's own example of the fractionalized executive came from his hometown, Indianapolis, where a company in a smaller city cannot attract a senior operator at any price it can pay.
Bruun's argument for carve-outs includes a point about legacy systems that is easy to miss: the painful part, the clean break, is what makes replacing the tooling possible at all.
He named the three risks he most wants out of a deal alongside rates — regulatory, foreign-exchange and supply chain — and said the pandemic's rewiring of supply chains is why the last one stays on the list.
Bruun's bottom line is that a private equity return now depends less on financial engineering than on whether the company can absorb AI, and that absorbing AI is a leadership question — which is why he thinks the competition for operators who can run a business through fast change is the tightest he has seen.
Products, Companies & Tools Mentioned
Goldman Sachs (Bruun's employer of 23 years: 45,000 employees, 12,000 engineers, and the source of the operating bench and vendor access he says is the strategy's edge)
The 10X Capital Podcast (Weisburd's other show)
Books & Resources Mentioned
Street Wise – Lloyd Blankfein (Blankfein's recent book, which Bruun said sets out the skin-in-the-game characteristics of Goldman Sachs)
Watch the full episode:
If this was worth your time, send it to someone closer to the industry than you are.
Get the latest market chatter as it happens: