Not for Sale, Then Sold: What Cognex's Offer for RealSense Teaches Owners

Cognex, the Massachusetts machine vision company, agreed on September 22 to buy RealSense for about $500 million in cash. RealSense was not for sale: its CEO says the company was preparing to raise another round of funding when Cognex approached three months ago, and that the board turned down the first proposal before accepting improved terms. For owners who get an unsolicited call from a buyer, the deal is a useful case study in what gives a seller leverage when there is no auction, and what a headline number does and does not include.

A Fourteen-Month-Old Company With a Price Tag

RealSense started in 2014 as a research effort inside Intel, building depth-sensing cameras that let robots judge distance and shape. Intel spun it out as an independent company in July 2025, when, according to CEO Nadav Orbach, quarterly sales were about $8 million and there was internal debate about shutting the unit down. The new company raised $50 million from a group led by a semiconductor-focused private equity firm, with Intel keeping roughly 20 percent.

Fourteen months later, RealSense expects 2026 revenue of $80 million to $90 million, more than 50 percent above the prior year. It has been profitable for two quarters. Cognex, with about $1 billion in annual revenue, is paying entirely from cash on hand.

At the midpoint of that revenue guidance, $500 million works out to about 5.9 times this year's revenue. Cognex did not frame the deal around current earnings at all. Its pitch to investors was market access: robotic perception is a market Cognex estimates at $600 million today, growing more than 25 percent a year to about $1.6 billion by 2030.

The Leverage of a Real Alternative

The most useful detail for private owners is the sequence. Cognex approached first. RealSense had a plan that did not involve a sale. The board said no to the first number. Cognex came back with better terms.

That sequence works only when the alternative is real. A company that is preparing a funding round has, in effect, a competing valuation in progress: investors are about to put a price on the business, and management can compare any offer against what it would own after raising capital and executing its plan. When a buyer knows the seller can walk away, a first offer is treated as a first offer.

Most private owners who receive an unsolicited approach do not have that. They have a business that runs well, no formal plan to sell, and no current view of what the company is worth. The call arrives, the number sounds large, and the owner has nothing to measure it against. Buyers know this. An inbound offer is often priced to see whether the owner will anchor on it.

The preparation is not complicated: know your normalized earnings (profit after removing one-time items and owner-specific expenses), have a recent independent valuation range, and have a credible answer to the question "what happens if you don't sell?" For RealSense, the answer was a funding round and continued growth. For a founder-owned services business, it might be a dividend recapitalization, a minority investment, or simply another three years of cash flow. The alternative has to be real enough that walking away is plausible.

Pricing a Company That Just Became Profitable

RealSense also illustrates how buyers price a business whose profits are too new to lean on. Two quarters of profitability is not enough history for an earnings multiple to carry the valuation. EBITDA (earnings before interest, taxes, depreciation and amortization, the usual starting point for private company pricing) would produce either a very large multiple or a meaningless one.

So the buyer priced on revenue and on what the revenue could become inside a larger company. Orbach acknowledged as much, telling the Israeli outlet Calcalist that "the multiple might indeed not be high" relative to the enthusiasm around robotics, but that the company had results to show. At roughly six times forward revenue for a business growing more than 50 percent, the price reflects growth without paying for the full hype cycle around physical AI.

For owners, the lesson cuts both ways. A company with fast revenue growth and thin margins will be priced on revenue and strategic fit, which favors buyers who can plug the product into an existing sales force. A financial buyer, pricing on stand-alone cash flow, would likely have offered less. That is why the identity of the buyer matters as much as the multiple: the same business is worth different amounts to different owners, and an unsolicited approach tells you which kind of buyer has already done the math.

$500 Million or $600 Million: What the Headline Includes

Reports on the deal cited two numbers. Cognex's announcement said about $500 million. Calcalist described a transaction valued at approximately $600 million. Both are accurate, because they measure different things.

The $500 million is the purchase price. On top of that, Cognex plans a three-year cash retention program of $56.5 million at target for RealSense employees, subject to performance adjustments, and about $50 million in restricted stock units under Cognex's own equity plan. That is roughly $106.5 million going to the team over time, conditioned on staying and performing. It is part of what Cognex is spending, but it is not paid to the shareholders who own the company today.

This distinction matters for any owner comparing offers. A buyer can present a larger total by including employee retention, earnouts, or rolled equity, while the cash that reaches the owners at closing stays the same or shrinks. The reverse is also true: a buyer who funds retention separately, rather than carving it out of the purchase price, is effectively adding value to the deal. The only way to compare two offers is to reduce each to the same question: how much cash, and how much contingent value, reaches each owner, and when.

What Was Left Out of the Sale

One more detail deserves attention. Before closing, RealSense will spin out its facial authentication product line into a separate company, led by about 25 existing employees. Cognex is buying the robotics perception business, not the biometrics business.

Carving a product line out of a sale is common, and it is often the right answer when a buyer only wants part of a business. But it has consequences the owner needs to plan for: the spun-off unit needs its own capital, contracts, intellectual property rights and people, and the shareholders need to agree on who owns it afterward. A carve-out also affects price. If the excluded business consumes cash, removing it can make the remaining company look more profitable. If it has value, the owners keep that value outside the deal.

For a private owner, the relevant question when a buyer wants only part of the company is simple: what is left, who will run it, and is it viable on its own? Settle those answers before signing.

When a Pre-Emptive Offer Makes Sense

Accepting an unsolicited offer means skipping the auction, and with it the competitive tension that often pushes price up. That is a real cost. But it is not always the wrong choice. BDO's 2026 private equity survey, released the same day as the Cognex announcement, found that 82 percent of fund managers expect deal prices to rise because too much capital is chasing too few quality companies. In that environment, a buyer who moves early to take a company off the market may be willing to pay for the certainty.

A pre-emptive deal makes the most sense when the buyer is a strategic acquirer that can use the business in ways others cannot, when the terms are clean (all cash, no financing condition, a short path to closing), and when the seller has tested the offer against a real alternative. Cognex's deal checks each box.