The Other Billion: How Retention Pools Come Out of a Sale Price

Nvidia confirmed on Thursday that it will acquire Hugging Face, the platform where developers share and deploy open AI models, for $12.93 billion. The number that matters more for anyone who owns a business is the split underneath it: roughly $11.9 billion goes to Hugging Face's investors, and up to $1 billion is set aside as an equity-based retention program for employees who join Nvidia. That structure, one price for the company and a second pool for the people, shows up in middle-market deals every week, usually with far less disclosure. How it is negotiated changes what a seller actually takes home.

What Nvidia disclosed

The deal is Nvidia's second largest on record, behind its $20 billion purchase of Groq assets in December. Hugging Face hosts more than 3 million models and is used by roughly 18 million developers and 200,000 companies. Its last priced round, in August 2023, valued the company at $4.5 billion on $235 million of new money from a group that included Google, Amazon, Salesforce, and Nvidia itself. Investors who came in at that valuation are looking at a bit under three times their money in three years.

The retention pool is the unusual part. At up to $1 billion, it is about 7.7 percent of the headline value. Willis Towers Watson's M&A retention studies put the typical retention budget at less than 2 percent of purchase price: about 1.4 percent for strategic acquirers and 2.1 percent for private equity buyers in the most recent survey. Pearl Meyer's review of more than 1,200 public-company transaction filings found retention pools clustered between $600,000 and $10 million, and shrinking as a share of deal size as deals get larger. Nvidia went the other direction, and the reason is plain: it is buying a developer community and the team that built it, not a factory. If the team leaves, the asset is worth less.

The closing timeline reinforces the point. The deal is expected to close in the first half of 2027, pending regulatory review. That is a nine to twelve month window in which competitors have a clear list of people to call. A retention program that only starts at closing does not solve that problem, which is why buyers increasingly structure pre-closing and post-closing tranches.

The question every seller should ask: who funds it?

In a middle-market sale, the retention pool appears in the letter of intent, often as a single line. The wording determines who pays for it.

If the pool is funded by the buyer on top of the purchase price, the seller's proceeds are unaffected. The buyer has decided that keeping the team is worth an extra 2 percent and is paying for it. That is the Nvidia structure as disclosed: $11.9 billion to investors, plus up to $1 billion for employees.

If the pool is funded from the purchase price, the seller is paying it. A $50 million headline with a $2 million seller-funded retention pool is a $48 million deal for the owners, before the other closing-day deductions such as debt payoff, transaction expenses, and the working capital adjustment. Buyers sometimes describe this as "the seller's decision to reward the team," which is accurate only if the seller understood the pool was coming out of their side of the table.

A third version is the most common and the least visible. The buyer offers a price, then says the offer assumes a retention program of a certain size will be in place, and asks the seller to establish it before closing. The seller pays the bonuses as a transaction expense, the buyer gets the benefit, and the price never changes. The pool has been funded from proceeds without anyone using those words.

The fix is procedural. Before signing a letter of intent, ask for the pool as a dollar amount and ask whether it sits on top of or inside the enterprise value. Then model net proceeds both ways. On a typical lower middle-market deal, where pools run 1 to 3 percent of enterprise value, the difference is the same order of magnitude as the working capital peg negotiation, and it gets a fraction of the attention.

When the owner is also the key employee

For founder-led businesses, the retention conversation is not about the staff. It is about the founder.

A buyer that wants the owner to stay for two or three years has two ways to pay for that. It can pay purchase price, which the owner receives at closing as capital gain. Or it can move some of that consideration into a retention award, an earn-in, or a rollover equity grant that vests over time. From the buyer's side, the second approach is better on every dimension: the money is tax-deductible as compensation, it is forfeitable if the owner leaves, and it aligns the owner's payout with the buyer's operating plan.

From the owner's side, the trade is worse than it looks. Purchase price is taxed at long-term capital gains rates, currently a top federal rate of 20 percent plus the 3.8 percent net investment income tax for most sellers. A retention bonus is ordinary income, taxed at up to 37 percent plus payroll taxes, and it is not paid at all if the owner is gone before the vesting date. A dollar moved from price to retention loses roughly 15 to 20 cents of after-tax value even if every condition is met, and it carries forfeiture risk on top.

This does not mean an owner should refuse all retention structure. Buyers have legitimate reasons to want continuity, and a seller who refuses any post-closing commitment often gets a lower price. It means the owner should treat retention consideration as a separate negotiation, and should ask that the buyer's first offer state the price without any owner-retention component, so the cost of adding one can be measured.

What the terms should say

Once the size and funding are settled, the terms of the pool determine whether it does its job. The seller wants the pool to reach the people who made the business valuable and to pay out on reasonable terms. The buyer wants flexibility.

WTW's most recent study noted that acquirers increasingly hold back part of the pool as "dry powder," making smaller initial awards and reserving the balance for whoever turns out to be a flight risk. That is sensible from the buyer's chair. From the seller's chair, it means the pool disclosed at signing may never be fully paid, and the people who kept the business running through diligence may see less than they expected.

The same study found that C-suite officers are twice as likely to be offered retention agreements as other employees. In a founder-led business the equivalent of the C-suite may be a plant manager or a lead salesperson, and the seller is better placed than the buyer to say who belongs on the list.

The forward look

Three things are likely to make retention economics more prominent in deal negotiations over the next year.

The first is the shift toward people-dependent targets. Software, services, healthcare, and anything with an AI component are businesses where the workforce is the asset, and the Nvidia pool is a visible signal that the 2 percent ceiling has moved for those sectors.

The second is longer sign-to-close periods. Every additional month between signing and closing is a month in which the seller risks delivering a diminished business, and buyers will push retention costs toward sellers as a hedge.

The third is the WTW finding that retention awards are moving from cash toward equity. Stock in a public strategic can be attractive. An equity award from a private equity platform is a claim on a company employees cannot value, with a payout date they do not control.