Inside the Secondary Market for GaaS Startup Shares

Table of Contents

• What "Secondary" Actually Means in GaaS
• Why a Secondary Market Exists at All
• The Three Channels: SPVs, Brokered Blocks, and Tenders
• How GaaS Secondaries Are Priced
• Why GaaS Shares Trade Differently from SaaS
• The Information Problem Buyers Keep Underestimating
• Transfer Restrictions, ROFRs, and the Paperwork Trap
• A Practical Playbook for Sellers and Buyers
• Insights Most People Overlook
• References

What "Secondary" Actually Means in GaaS

A secondary transaction is simply the sale of existing shares from one holder to another, rather than the company issuing new stock to raise capital. No money goes into the business. When a third engineer at a vertical legal-agent startup sells a slice of vested options to a growth fund, that is a secondary. When a crossover investor buys a $40 million block from a seed VC looking to return capital to its LPs, that is a secondary too.

In the agentic AI-as-a-Service world, this distinction matters more than usual. The primary market, the headline funding rounds tracked obsessively across the GaaS cluster, gets all the press. But the secondary market is where price discovery happens between those rounds, and where the people who actually built these companies turn paper wealth into rent money. For a sector where a Series B can be 18 months apart from a Series C, that gap is long enough for fortunes to feel theoretical.

The mechanics are old. What is new is the speed and scale at which GaaS companies are generating illiquid paper wealth, and the appetite from buyers who could not get allocation in the priced rounds.

Why a Secondary Market Exists at All

Three forces created this market, and all three are sharper in GaaS than they were in the SaaS era.

First, companies stay private longer. The median time from founding to IPO has stretched well past a decade, and the most-hyped agent companies show no urgency to go public when private capital keeps arriving at rising marks. An employee who joined an agent infrastructure startup in its first 20 hires might wait eight years before any conventional liquidity event. Secondaries are the release valve.

Second, the money on the table is enormous and concentrated. GaaS rounds have been unusually top-heavy, a handful of companies command premium valuations and absorb the bulk of capital, a pattern documented across the broader AI funding landscape in Crunchbase's running AI funding analysis. When a 50-person company is marked at several billion dollars, even a 0.2% common stake is life-changing, and the holder has every incentive to find a buyer.

Third, demand is desperate. Plenty of credible funds missed the priced rounds entirely, they passed too early, got squeezed out by oversubscription, or simply did not have the relationship. The secondary market is their back door. That asymmetry, lots of buyers chasing scarce shares in the handful of "winners," is what pushes secondary prices to strange places.

The Three Channels: SPVs, Brokered Blocks, and Tenders

Secondary liquidity in GaaS flows through three fairly distinct pipes, and they attract different participants.

SPVs and syndicates

A special purpose vehicle pools capital from many smaller investors to buy a single position. SPVs have become the retail and prosumer on-ramp into hot agent names, a topic explored more fully in the cluster piece on SPVs and the retail rush into agent investing. They democratize access, but they also stack fees and often sit several layers removed from the company's cap table, sometimes buying from another SPV rather than from a direct shareholder. By the time a "deal" reaches a syndicate email, the markup may be substantial and the information thin.

Brokered blocks

Larger, cleaner transactions move through specialist secondary brokers and platforms. These are negotiated trades, a fund sells a meaningful block to another institution, usually with at least some company cooperation or a recent 409A reference point. Platforms like Forge Global and similar secondary marketplaces have professionalized this layer, providing bid-ask data and settlement infrastructure that simply did not exist for the prior startup generation.

Company-run tender offers

The cleanest mechanism is a structured tender, where the company itself organizes a liquidity window, often alongside a primary round, letting employees and early holders sell a capped amount at a set price to approved buyers. Tenders are sanctioned, transparent about price, and respect the cap table. They are also the subject of their own dedicated cluster article on tender offers and employee liquidity at agent unicorns, because they are quickly becoming the preferred retention tool at the top GaaS names.

How GaaS Secondaries Are Priced

Here is the part outsiders get wrong: there is no single price. A GaaS secondary trades at a discount or premium to the last preferred round, and the spread can be wild.

Common shares almost always trade at a discount to the preferred price set in the last round, because preferred stock carries liquidation preferences, anti-dilution protection, and other rights that common simply does not have. A 20-40% common-to-preferred discount is normal. But in the hottest agent names, FOMO can invert the logic entirely, buyers pay a premium to the last round because they expect the next round to be marked far higher, and the secondary is their only way in.

Three reference points anchor any negotiation: the last preferred round price, the 409A valuation (the company's own appraisal of common stock, usually conservative), and recent comparable secondary trades. The gap between the 409A and the round price is itself a signal. Pricing also leans heavily on revenue multiples, which the market applies differently to usage-based agent revenue than to SaaS subscriptions, a nuance covered in the cluster's revenue-multiples breakdown. McKinsey's work on the economic potential of generative AI is the kind of macro thesis that buyers cite to justify rich marks; it is also the kind of thesis that evaporates fast if a single model-price war compresses a company's margins.

Why GaaS Shares Trade Differently from SaaS

If you priced GaaS secondaries with a SaaS playbook, you would get burned in both directions. Four structural differences drive the divergence.

Revenue durability is genuinely uncertain. A SaaS seat renews or it churns, and you can model retention. A per-task or per-outcome agent contract can evaporate the moment a customer finds a cheaper agent or builds in-house. Secondary buyers who treat usage revenue as if it were subscription ARR are mispricing the durability risk, the central question in the revenue-quality debate running through this beat.

Cost of goods sold is a moving target. Agents are expensive to run; inference is a real, variable cost that scales with usage. A company can post explosive revenue and still have ugly gross margins if model costs are eating it alive. Worse, a foundation-model price cut can swing margins overnight in either direction. SaaS gross margins are boringly predictable; GaaS margins are not.

Concentration risk is brutal. Many vertical agent companies have a handful of marquee logos producing most of their revenue. Lose one, and the growth story breaks. Secondary buyers rarely see the customer concentration data, so they price as if revenue were diversified when it often is not.

The moat question is unsettled. Is the company's edge its proprietary workflow and data, or is it a thin wrapper around a model that any competitor can replicate? That answer determines whether today's mark survives the next 18 months. The "agentwashing" problem that plagues fundraising decks shows up in secondaries too, just with even less disclosure to catch it.

The Information Problem Buyers Keep Underestimating

Secondary buyers operate in the dark, and GaaS makes the dark darker.

In a primary round, the lead investor gets a data room, management access, and a board-level view. A secondary buyer, especially through an SPV three hops from the cap table, often gets a pitch deck that may be a year old and nothing else. They cannot see current burn, current margins, churn, or customer concentration. They are buying a logo and a narrative.

This is dangerous in any private secondary, but in GaaS the underlying metrics move so fast that a 12-month-old deck is nearly worthless. A company that looked default-alive last year on usage revenue might be burning hard now after a margin shock. The SEC's guidance on the risks of pre-IPO investing is dry but worth reading before wiring money into any of these structures, particularly the warnings about unverified valuations and stale information. The single best discipline a secondary buyer can impose is to refuse any deal without a recent 409A and, ideally, some form of company-sanctioned information.

Transfer Restrictions, ROFRs, and the Paperwork Trap

Even when buyer and seller agree on price, the deal can die in the documents.

Almost every venture-backed company's stock is subject to transfer restrictions. The most common is a right of first refusal (ROFR): before a holder can sell to an outside buyer, the company, and sometimes existing investors, can step in and buy the shares on the same terms. A ROFR can take weeks to clear and can vaporize a deal entirely if the company decides it wants the shares itself. Many GaaS companies also flatly prohibit common-share transfers without board consent, precisely to keep their cap tables clean and their information closely held.

This is why SPVs proliferate: when you cannot transfer the underlying shares, you sell an economic interest in a vehicle that holds them, which sidesteps some restrictions but introduces new layers of counterparty and fee risk. Sellers should read their stock agreements before promising anything, and buyers should confirm the chain of title is real and not a synthetic interest dressed up as ownership. A surprising number of "secondary" offers circulating in the agent space are forward contracts or profit-sharing arrangements rather than actual share transfers, a distinction that becomes very expensive at exit.

A Practical Playbook for Sellers and Buyers

For sellers, usually employees, the rules are simple but easy to ignore. Know your transfer restrictions before you market anything. Wait for a company-sanctioned tender if one is plausibly coming, because tenders are cleaner, faster, and usually priced better than a back-channel block. Model the tax hit, since secondary proceeds can trigger ordinary income or capital gains depending on your instrument and holding period, and the bill can be enormous. And do not sell your entire position into the first FOMO bid; partial liquidity that takes risk off the table while keeping upside is almost always the smarter move.

For buyers, discipline beats access. Insist on a recent 409A and any sanctioned information you can get. Triangulate the price against the last round, the 409A, and real comparable trades rather than trusting an SPV organizer's markup. Underwrite the margin and durability risks specific to GaaS, inference cost exposure, customer concentration, model-dependency, not just the topline growth. And map the full fee and structure stack: an SPV-of-an-SPV with a markup at each layer can mean you are paying a 50% premium to the "headline" secondary price before the company has grown a dollar. In a sector this volatile, the entry price is the one variable you fully control.

Insights Most People Overlook

The secondary price is a better fear gauge than the round price. Priced rounds are negotiated, lagging, and often propped up by signaling incentives. The secondary bid-ask spread, by contrast, reflects what real buyers will pay today with their own money. When secondary discounts on a hot GaaS name suddenly widen, that is an early-warning indicator of a coming down round long before the cap table reprices, watch the secondary, not the press release.

Employees selling is not always a red flag, but a cluster of insiders selling is. A single engineer cashing out for a house down payment tells you nothing. A coordinated rush of early employees and angels offering blocks at steepening discounts tells you the people with the best information are quietly de-risking. Aggregate insider selling behavior is one of the few signals that pierces the GaaS information fog.

SPV proliferation around a name is a liquidity signal that cuts both ways. When dozens of syndicates are marketing a single agent company, it means demand is hot, but it also means original shareholders are eager enough to sell that supply is flowing. Saturation of SPV offers often precedes a plateau in the primary mark, because the marginal enthusiastic buyer has already been absorbed.

The 409A-to-round gap is the cleanest tell about real value. Companies have an incentive to keep the 409A low (cheaper option grants) and the round price high (better optics). When that gap is unusually wide on a GaaS name, the secondary buyer paying near the round price is buying the optimistic number while the company's own appraiser is signing off on something far lower. Few buyers ever ask to see both.

Forward contracts dressed as secondaries are the sleeper risk. A meaningful share of "agent unicorn secondaries" floating around are not share transfers at all, they are forwards or synthetic exposure that only convert to real ownership at a liquidity event, if the counterparty performs. In a sector where some of these companies will not survive a margin shock, counterparty risk on a forward is a way to lose money even when you "picked the winner."

References

• Crunchbase News, AI Funding Coverage and Analysis
• Forge Global, Private Market Secondary Marketplace
• McKinsey, The Economic Potential of Generative AI: The Next Productivity Frontier
• U.S. SEC, Investor Alert: Pre-IPO Investing