NextFin

Investor Frenzy for AI Strips Safeguards From Convertible Bonds

Summarized by NextFin AI
  • U.S. convertible issuance hit $85.5 billion by early June 2026, already 90% of 2024's full-year total, with AI-linked issuers making up over 30% of the global market.
  • AI issuers are waiving key protections like the fundamental-change put right and make-whole premium, accepting thinner documentation in exchange for cheap capital.
  • CoreWeave priced $3.5 billion at a 1.75% coupon, while Tempus AI and Onto Innovation issued zero-coupon convertibles, and Alphabet raised $19.3 billion via mandatory structures.
  • The erosion of safeguards raises second-order risks: AI convertibles could lose their bond-floor protection if equity and credit markets fall simultaneously.

NextFin News - Investors chasing artificial-intelligence exposure are handing over one of the bond market's oldest safeguards in exchange for a seat at the table: the right to demand their money back if the company that sold them convertible debt is taken over. Across the record-breaking convertible market of 2026, AI-linked issuers are pricing deals with thinner documentation — waiving the "fundamental change" put right, dropping the make-whole premium that compensates holders for forced early conversion, or issuing mandatory structures that remove the holder's choice altogether. The concessions capture the mood of a market that has priced AI as a one-way bet, and they are exactly the kind of terms that have preceded the last time a hot sector flooded the convertible market and left investors holding the bill.

The Deal: Cheap Cash for Issuers, Thinner Protection for Holders

U.S. convertible issuance reached about $34 billion in the first four months of 2026, more than double the same period a year earlier, according to Bank of America Global Research and Barclays Research. That pace puts the market on track to surpass the 2025 full-year record of more than $120 billion. By early June, issuers had raised $85.5 billion across 127 deals — already 90% of the $96.1 billion raised across 265 deals in all of 2024, and roughly 61% of the $140.5 billion record set in 2025, according to Dealogic data compiled by Mergermarket. The second quarter alone produced $49.5 billion across 52 transactions, the largest quarterly total on record, with June delivering $21.4 billion across just seven transactions — the busiest month by volume despite the fewest deals.

Roughly half of this year's issuance is tied in some way to AI. Computers and electronics companies — the backbone of the AI buildout — raised $56.8 billion across 43 deals, accounting for 66.4% of 2026 issuance volume. AI-adjacent issuers now make up more than 30% of the global convertible bond market, up from low single digits a few years ago, according to BofA Global Research.

The price of that access is falling. CoreWeave, the cloud infrastructure company at the center of the AI financing boom, sold $3.5 billion of convertibles due 2032 at a coupon of just 1.75%, upsizing the deal from an initially proposed $3 billion. Tempus AI raised $400 million through a six-year convertible with a zero coupon and no principal increase at maturity — investors receive nothing unless the stock climbs to $69.26, about 40% above where it traded when the deal was sold in May. Onto Innovation priced $1.3 billion of 0.00% notes due 2031. And some of the largest names in the market have gone further, tapping mandatory structures that force conversion into equity at maturity, removing the holder's option entirely. Alphabet raised $19.3 billion in June through two series of mandatory convertible preferred stock; combined with Super Micro Computer's mandatory deals, the two companies raised more than $23 billion in structures that hand equity to investors whether they want it or not. Oracle added a $5 billion mandatory raise in the first quarter.

"A lot of it is to build out capital expenditure, particularly AI, and that's unusual and not something we've seen in previous cycles," said Michael Youngworth, managing director and head of global convertibles at Bank of America Securities. "Companies are not necessarily coming for a specific need, but because money is cheap."

What Is Being Stripped, and Why It Matters

To understand what investors are giving up, it helps to know what a convertible bond is supposed to protect. A convertible note is a hybrid: it pays a fixed coupon and principal like a bond, but can be exchanged for shares if the stock rises to a predetermined price. That conversion feature embeds a call option on the issuer's stock, which is why convertibles can be priced with coupons far below straight debt. In return for accepting a lower yield, holders have historically received two safeguards that distinguish convertibles from plain equity.

The first is the "fundamental change" put right. Under a standard indenture, a fundamental change includes a change of control, the sale of all or substantially all of the issuer's assets, a transaction in which existing stock is replaced by consideration that is less than 90% exchange-traded securities, or a delisting or insolvency. When such an event occurs, noteholders may require the issuer to repurchase their notes for cash at 100% of principal, plus accrued interest. It is the bondholder's escape hatch if the company is sold to a private buyer, loaded with debt, or ceases to be publicly traded.

The second is the "make-whole" premium. If an issuer exercises its soft-call right or forces an early conversion, holders receive a temporary increase in the conversion rate — extra shares designed to compensate them for the remaining option value they lose when the note is extinguished ahead of schedule. The make-whole is most valuable early in a note's life and when the stock is near the conversion price; it is the mechanism that keeps an issuer from calling away the option just as it becomes in the money.

Both protections are now being waived, narrowed, or priced out of deals as issuers test how much investors will accept. The dynamic is a clean read on market psychology: when demand for exposure is intense and the underlying theme is believed to be unstoppable, investors stop paying for insurance against the theme failing. The fundamental-change put is, in effect, a bet that the company might be acquired or fail; the make-whole is a bet that the stock could rise sharply and the issuer could force conversion. In an AI market where the dominant view is that every issuer will either be bought at a premium or see its shares multiply, both hedges look like unnecessary costs to be negotiated away.

The irony is that the protections being discarded are precisely the ones that matter in the scenarios where a convertible is supposed to act like a bond rather than a stock. If an AI infrastructure company is acquired by a private-equity consortium and taken private, the put right is the only thing standing between a noteholder and illiquid paper in a non-public entity. If a struggling issuer refinances on punitive terms or forces an early conversion to clean up its balance sheet, the make-whole is the compensation for stolen optionality. Strip both away, and the instrument's "bond floor" becomes a promise from a company that has just demonstrated it can rewrite the contract.

That said, the erosion is selective rather than universal. CoreWeave's upsized $3.5 billion offering, priced through a private placement in April, still allows noteholders to require the company to repurchase the notes at 100% of principal plus accrued interest if a fundamental change occurs. Penguin Solutions' convertible notes carry the same repurchase right. The most aggressive terms are concentrated among the most sought-after names, where investors compete hardest for allocation — which is precisely why they are a warning signal rather than a market-wide indictment.

The Mechanism: How Frenzy Becomes Terms

The transmission channel from AI enthusiasm to weaker documentation runs through three distinct groups of market participants, each with a different incentive.

Issuers are facing the largest capital-spending program in a generation. Data centers, power infrastructure, and chip capacity require upfront cash measured in the tens of billions, and the window to raise it cheaply is the constraint on the entire buildout. Convertibles solve the problem better than either alternative: straight debt is expensive in a high-rate environment — benchmark 10-year Treasury yields have reached a 16-month high — and equity issuance dilutes shareholders whose support management needs to keep funding the cycle. A convertible lets an issuer borrow at a coupon near zero while deferring dilution to a future date when, the company's bet, the stock will be higher.

Hedge funds, the traditional marginal buyers of convertibles, are not the ones demanding the safeguards. Their strategy is convertible arbitrage: buy the bond, short the stock, and harvest the implied volatility embedded in the conversion option. For a hedge fund running a relative-value book, the fundamental-change put is a tail-risk feature it rarely prices meaningfully, and the make-whole matters only in scenarios its short hedge already covers. What the hedge fund wants is a liquid, volatile name with a rich option premium — and AI delivers both.

Long-only investors are the new marginal money, and they are the ones conceding the protections.

"Long-only investors are buying for exposure to semiconductors — the hottest part of the market right now, driven by AI capital spending," said Venu Krishna, managing director and head of U.S. equity strategy at Barclays.
For a long-only portfolio manager under pressure to show AI exposure, a convertible offers a politically defensible compromise: it sits in the fixed-income sleeve, carries a bond-like label, and participates in the upside. The manager's benchmark anxiety — the fear of missing the AI rally while holding "safe" assets — is a more powerful force than the fine print of an indenture. A protection that pays off only in a takeover or a collapse is easy to sacrifice when the consensus view is that neither will happen.

The result is a market where the people most eager for the product are the least equipped to value its protections, and the people who understand the protections have no reason to insist on them. "I think deals are trading well and investors are happy to get exposure to AI through the convert market, so I'd expect that to continue," said Bryan Goldstein, head of convertible bonds origination at Matthews South. Then came the caveat, in a

"Right now everything is very strong and there's a lot of optimism. There needs to be the right balance: terms that are attractive for issuers but that also work for investors."

Cyclical Frenzy or Structural Shift? Both, and That Is the Problem

The critical question for anyone reading this market is whether the erosion of protections is a cyclical symptom of a hot issuance window — the kind of thing that reverses when sentiment cools — or a structural change in how growth companies finance themselves. The answer is that both forces are at work, and they point in opposite directions across different time horizons.

On the cyclical leg, the evidence is unmistakable. Convertible markets have seen this movie before. In 2024, crypto companies flooded the convertible market; issuance collapsed alongside crypto prices, and investors who bought on the theme were left with impaired credits. The pattern is a textbook late-cycle signal: when documentation quality deteriorates at the same time as issuance volume hits records, the marginal deal is being done for access, not for value. Credit spreads near decade lows, coupons at zero, and investors accepting structures with no yield at all are the fingerprints of a market that has stopped underwriting credit and started underwriting a narrative.

On the structural leg, the AI capital cycle is different in scale and duration from the crypto episode. This is not a speculative asset class financing itself; it is the infrastructure layer of what investors believe is a general-purpose technology, funded by companies with real revenue, real customers, and multi-year contracted demand. Mandatory convertibles — which automatically convert into equity at maturity — have become an established tool for very large, strategic raises rather than a niche structure. When the largest, most creditworthy names in the market institutionalize a structure that removes holder optionality, that is a regime change in financing behavior, not a cyclical aberration. The convertible and convertible-like instrument has moved from being a fallback for companies that cannot access cheap straight debt to being a primary funding channel for the AI capital cycle.

Separating the two legs matters because it determines what happens when the music stops. A cyclical wave reverts quickly: spreads widen, coupons rise, and protections come back within a couple of quarters. A structural shift does not revert on its own — the market simply learns to operate with thinner documentation, and the next downturn tests whether investors understood what they bought. The dangerous combination is a structural shift in financing behavior riding on top of a cyclical peak in investor enthusiasm. It means the terms being written today will still be in force — five, six, seven years out — when the cycle has long since turned.

The Second-Order Risk Nobody Is Pricing

The first-order consequence of stripped safeguards is obvious: investors have less protection if an AI issuer is acquired or forced into an early conversion. The second-order consequence is less discussed and more important: it changes the correlation structure of the asset class at exactly the moment investors think they are diversified.

A traditional convertible portfolio is attractive because the bond leg and the equity leg are negatively correlated in stress. When the stock falls, the bond floor holds; when credit spreads widen, the option still has time value. That is the "free lunch" that long-only investors think they are buying when they allocate to convertibles for AI exposure. But a portfolio of AI convertibles with no put right and no make-whole does not have that property. In the scenario that matters — an AI sentiment shock that hits both equity prices and credit spreads simultaneously — both legs of the instrument fall together. The bond floor is only as good as the issuer's ability to pay, and an AI infrastructure company that has burned cash building data centers on the assumption of perpetual demand growth is not a reliable floor in a downturn.

The transmission then runs to the broader market. Convertibles are a core holding of convertible-arbitrage hedge funds, which hedge their long bond positions by shorting the underlying stock. When a convertible de-rates on both legs at once, the hedge fund must deleverage: it sells the bond and covers the short, which pushes the stock higher — but only if liquidity holds. In a disorderly move, the forced selling of the bond leg widens spreads for the entire asset class, and the issuance window that AI companies depend on slams shut. Companies that priced their capital plans on the assumption of continuous access to cheap convertible funding suddenly face a refinancing cliff. The weakest credits — the WhiteFibers of the world, recently public companies that sold a $230 million five-year convertible in January to fund data-center expansion while carrying a negative forward earnings multiple — are the first to find that the market which funded them at zero coupon will not fund them at all.

There is also a subtler cross-market transmission. The same long-only investors buying convertibles for AI exposure are often the buyers of the straight bonds and leveraged loans of the same issuers. When the convertible sleeve of their portfolio disappoints, the natural response is to become more conservative across the whole credit relationship — tightening internal limits, demanding more protection, or simply stepping back. A loss of confidence in the convertible market can therefore contaminate the straight-debt market for the same names, raising the all-in cost of capital for the AI buildout at the moment it is least affordable.

The Counter-Argument: Investors Are Being Paid, Not Robbed

The strongest case against the "investors are being stripped" narrative is that convertible bonds have always been lightly covenanted, and that today's terms are simply the fair price of rich equity valuations and elevated volatility. "Traditional capital markets convertible notes typically have few covenants, but do require attention to compliance," noted a January 2026 analysis of the market by the law firm Cleary Gottlieb. From this perspective, nothing has been given away for free. Investors accept a lower coupon and thinner protections because the conversion premium — the 40%, 50%, even 150% overlay to the current share price — and the embedded volatility are compensation enough. If the stock doubles, the holder participates fully; the make-whole and put right are just expensive insurance on a house the investor believes will never burn down.

There is truth to this. Many of the headline deals in 2026 still carry the fundamental-change put. CoreWeave's notes include it. Penguin Solutions' notes include it. The erosion of protections is selective, concentrated among the most sought-after names where investors are willing to compete hardest for allocation — not a universal feature of the market.

And the historical analogy cuts both ways. The 2024 crypto convertible wave did end badly for many holders, but the AI narrative is materially stronger: it is backed by hyperscaler capital expenditure, enterprise adoption, and government policy, not by retail speculation in unproven tokens. A structural financing shift toward mandatory and low-coupon convertibles may simply reflect the market's rational conclusion that AI issuers are more likely to see their shares rise than to be taken private or to fail.

The rebuttal is that the compensation argument works only if the option is actually valuable — and the option is valuable only if the stock is volatile and the holder has time to realize it. A mandatory convertible that forces conversion at maturity, or a zero-coupon note with no make-whole, transfers the timing decision from the holder to the issuer. The investor is no longer long an option; the investor is long a promise that the issuer's capital plan will succeed. That is a different instrument with a different risk profile, and it deserves a different price. When the market prices both instruments as if they were the same, the cheaper one is being mispriced — and the mispricing is a bet that the cycle will not turn before maturity.

What to Watch: The Signals That Will Prove the Thesis Wrong

The judgment that protections are being stripped for the wrong reasons rests on a falsifiable premise: that today's terms are an abnormal concession to demand, not a rational repricing of risk. Three observable signals would prove that judgment wrong.

First, if the share of new convertible deals issued without a fundamental-change put right or make-whole premium falls back below roughly 15% to 20% of volume within the next two quarters, the erosion is cyclical noise, not a trend. Deal-term surveys and indenture reviews would show protections returning as issuance normalizes.

Second, if conversion premiums on AI-linked convertibles re-widen above 40% to 50% for sub-investment-grade issuers — back toward the five-year average — investors are once again demanding to be paid for optionality, and the frenzy has cooled.

Third, and most simply, if the ICE BofA US Convertible index — up more than 20% this year, ahead of both the S&P 500 and the Nasdaq — suffers a drawdown of more than 10% from its peak without a corresponding surge in new issuance, the market is repricing risk rather than absorbing supply, and the terms being written today will look generous in hindsight.

Outlook: Three Scenarios, Split by Time Horizon

Short term (next 3–6 months): issuance stays hot, terms stay issuer-friendly. The backlog of AI capital projects is large enough to keep companies coming to market through the summer, despite the seasonal slowdown that usually accompanies corporate blackout periods.

"Typically we'd expect some slowdown in the summer," said Ivan Nikolov, head of convertible bonds at Fisch Asset Management. "But given the pace of these capex programs, I don't think this will be a typical summer."
In this window, the marginal deal will continue to be priced for access, and protections will remain thin.

Medium term (6–18 months): the first test. As 2026 vintages approach their first call dates and early refinancing windows, the market will learn which AI issuers can actually fund their plans from cash flow and which depend on perpetual access to cheap capital. This is where the stripped safeguards bite: a company that needs to refinance a zero-coupon convertible in a wider-spread environment, without the ability to offer holders a make-whole or a put, will have to pay up in coupon or dilution. Expect dispersion to widen sharply between the Alphas of the world and the WhiteFibers.

Long term (beyond 18 months): a new normal, or a cautionary tale. If the AI capex cycle delivers the productivity gains investors expect, today's convertibles will convert into valuable equity and the stripped protections will be remembered as a clever negotiation. If the cycle disappoints, the 2026 vintage will join the 2024 crypto cohort as a reminder that documentation quality is a lagging indicator of market tops. Either way, the structural shift toward mandatory and equity-linked financing for large technology capex is likely to persist — the question is only whether future investors will be paid appropriately for accepting it.

The base case is a bumpy middle path: issuance remains elevated but terms gradually firm as the easiest money has been raised and the marginal issuer gets weaker. The upside case is that AI earnings compound fast enough to make every 2026 convertible a profitable conversion, rendering the debate academic. The downside case is a sentiment shock that forces both legs of the trade lower at once and reveals how much of the market's "bond floor" was really just confidence in a narrative.

The bottom line: the convertible market is not just financing the AI boom — it is underwriting its assumptions. Investors who surrender the put right and the make-whole are not merely accepting a lower yield; they are betting that no AI issuer they own will ever be acquired on unfavorable terms, forced into a restructuring, or fail to reach the conversion price. That is a bolder bet than the "conservative fixed-income allocation" label suggests, and when the cycle turns, the contracts written in this window will reveal exactly who understood the instrument they bought.

Explore more exclusive insights at nextfin.ai.

Insights

What is a convertible bond and how does it function as a hybrid instrument?

What are the fundamental change put right and make-whole premium protections?

Why have convertible bonds historically included these investor safeguards?

How much convertible bond issuance occurred in the first four months of 2026?

What percentage of the global convertible bond market is now tied to AI issuers?

Why are long-only investors becoming the marginal buyers of AI convertibles?

How do hedge funds view convertible protections compared to long-only investors?

What specific terms did CoreWeave and Tempus AI accept in their recent convertible deals?

How did Alphabet and Super Micro Computer utilize mandatory convertible structures in 2026?

What data shows the 2026 convertible market is on track to break records?

What are the three time-horizon scenarios for the AI convertible market outlook?

How might stripped safeguards affect issuers during a future refinancing cliff?

What signals would indicate the erosion of protections is reversing?

Could mandatory convertibles become a permanent structural shift in tech financing?

Why does removing protections change the correlation structure of convertible portfolios?

What is the second-order risk nobody is pricing in the AI convertible market?

How could losses in convertibles contaminate the straight debt market for AI issuers?

Is the current erosion of safeguards cyclical noise or a structural regime change?

How does the 2026 AI convertible wave compare to the 2024 crypto convertible episode?

Why are the weakest credits like WhiteFiber most vulnerable to these market changes?

Search
NextFinNextFin
NextFin.Al
No Noise, only Signal.
Open App