The timing of the sacking was terrible, given an upcoming share sale, which if successful would give the owner of ChatGpt much needed funds at a purported $86bn valuation. 

Key $13bn investor Microsoft was reportedly told just an hour before of the decision to sack the CEO. On Sunday, Altman and several key OpenAI employees then briefly adopted Office 365 as their operating platform. Some commentators suggested Microsoft had played a blinder securing its key human assets for nothing, while still having access to the proprietary technology — in effect paying $13bn for a $86bn business. 

Others said frictions (and the sacking) resulted from Altman pushing harder on AI development, seeking to reach artificial general intelligence (a level above human intelligence) faster. 

Plus there are big commercial pressures for the altruistic business. Despite the lofty valuation, the business burns a lot of cash. It reportedly costs around $700,000 per day (or $0.36 per enquiry) to keep ChatGpt running, requiring huge computing power. It is unclear how many punters are paying $20 a month for the Plus option, which would certainly help commercially, with Microsoft keen to wean it off expensive Nvidia chips by developing its own. 

There are several other interesting angles. Not least of these is the cult of the charismatic CEO, a trait almost seen as a prerequisite for tech company leaders. 95% of OpenAI employees signed a letter asking the board to resign, and many swiftly followed him to Microsoft. 

But charisma is not always the best for a business — as Christian Stadler from Warwick University famously said: “For every Steve Jobs there is a Dick Fuld.” (or should that be a Sam Bankman-Fried). Or, in real estate, for every Sam Zell there is a Donald Trump (or Rene Benko). 

Their 2013 study of 100 European corporations suggested leaders of highly-performing companies were less likely to be charismatic: 

“The problem with charismatic leaders is that exceptional powers of persuasion make it easy for them to overcome resistance and opposition to their chosen course of action.”

Admittedly, the stage of the business is important. Steve Jobs worked well for Apple in the early days, but Tim Cook is the leader that has created the biggest value. Stadler recognises this: 

“If your company is heading in the right direction, a charismatic leader will get you there faster. Unfortunately, if you’re heading in the wrong direction, charisma will also get you there faster.”

Is this where X (nee Twitter) is heading? Hurtling to earth faster than a Starship rocket. 

But my main interest in OpenAI is cultural, and the relationships between boards, employees and investors. Bloomberg columnist Matt Levine nailed the issues on Monday: the board is a non-profit organisation, holds no shares, and appoint themselves. 

“They answer to their own consciences, not to any investors,” wrote Levine, citing OpenAI’s own materials:

“The Nonprofit’s principal beneficiary is humanity, not OpenAI investors.”

But to grow the loss-making business and fund development (as we’ve mentioned above) requires lots of money (as Matt’s amended org chart above shows). It also shows how much control was in the hands of the board.

The risk now is that the money (investors and employees are a powerful coalition) wins and OpenAI becomes more like a normal tech company, with a more powerful and emboldened founder CEO at the bridge. I won’t get into the ethical issues or threats to humanity debate here, but it does show conflicts between stakeholders and boards matter if interests are misaligned. 

Duties of care

Which brings me on to boards and company directors duties, an area I feel doesn’t always get the attention it should get, especially in stressed/distressed situations. 

Exactly when does the duty of care of companies and directors switch from owners/sponsors to concerned creditors? 

Directors should always act in interests of the company, but how does that work at somewhere such as X where all control is in the hands of its new owner/sponsor? What say do X directors and board members have? Should they push for control over the erratic posts of the owner, which recently led to advertisers leaving in droves and may place the solvency of the business at risk? Should they be worried about libel and deformation as key business risks after checks and balances were removed — or as a platform not a publisher, does X marks its posts? 

In distressed situations, as the situation worsens the directors duties dynamic should shift away from sponsor and shareholders towards creditors. But the tipping point is important, and is often later than in past years given the absence of covenants and more doc flexibility in recent vintages. 

Sponsors typically pick the legal and financial advisors to their portfolio companies at times of distress. But the new advisors’ duty is not to them but to the company itself. Yes, the sponsor has first mover advantage and could fix the problem via equity injections and other measures but, if this is not forthcoming, the duties of directors and advisors to the company should shift. 

There can also be further complications, with complex group (and cross-border) structures, leading to number of boards with common directors — in theory they should wear different hats on different boards to reflect this. 

This was a criticism made by a group of Adler Real Estate bondholders in a letter to directors (most of which also sat at Adler Group level) last year, concerned about a series of transactions which seemingly leaked value from the structurally senior German OpCo to the Luxembourg incorporated parent. 

Directors duties (and the penalties for breaches) differ wildly between jurisdictions (another factor in the Adler spat). Germany is historically where duties were most stark — a rigid (personal criminal liability) regime for directors of companies in the zone of insolvency — the infamous 21-day rule (suspended during Covid). And in November 2022, these were relaxed (again) in a response to the effects of the energy crisis. 

For lenders, Germany had the spectre of equitable subordination — if a court appointee decided retrospectively that lending had been done at a time when the group was technically insolvent, their debt claim could be converted to equity! 

Last year, there was some welcome advice from the UK Supreme Court on directors duties to creditors, in the BTI versus Sequana case. This relates to €135m dividend paid to a sole shareholder despite contingent environmental liabilities which could threaten its solvency. As Greenberg Traurig explains in a client note:

“The Court confirmed that when a company is insolvent or bordering on insolvency, or when an insolvent liquidation or insolvent administration is probable, this fiduciary duty is modified by common law such that the company’s interests are taken to include the interests of its creditors as a whole (the Creditor Duty). The Court confirmed that imminent insolvency, or a probability of an insolvent liquidation or insolvent administration, are the trigger points for directors to consider the interests of creditors.”

But the court rejected the “real risk of insolvency” as the appropriate trigger, suggesting that the flip in duties is only at a relatively late stage. Appropriate weight when creditor duties are engaged should be given to creditors if their interests conflict with shareholders, the judges said. 

Greenberg Traurig adds:

“The majority of the Court held that the Creditor Duty is engaged when the directors knew or ought to have known about such insolvency or probable insolvency (i.e., this is a subjective and an objective test).

However, where the company is irretrievably insolvent, the interests of those creditors become a ‘paramount consideration’ in the director’s decision making. This is because at that point the shareholders no longer are considered to have any economic interest in the company.”

So, a sliding scale balancing interests appears to be the approach. But the judgment appears to say that directors must consider interests of creditors — which suggests all, not just those in-the-money — but should the juniors be considered on the same basis as those in the equity, or higher? Let us know your thoughts. 

And in France? Arguably in the too complicated pile, as Norton Rose explains:

“Judges are not obliged to base their decisions on a purely arithmetic approach as to profits and losses . For instance, abusively pursuing a loss-making activity has been recognised as willful mismanagement even if the company had some profitable years over a considered period of time and regardless of the fact that the company may not have been in suspension of payments. Generally, directors were found liable because they had not taken any steps to turn the situation around after the difficulties began.”

They add: 

“This approach could cause issues for enterprises pursuing an activity despite suffering losses, even if the losses are eventually covered by shareholders loans. Should a French subsidiary end up in liquidation, the directors could still be at risk for pursuing a loss-making activity with an abnormal use of intragroup indebtedness.”

Unacceptable in the 80s

It has been great to work again with a bunch of ex-colleagues from Creditflux, who recently joined 9fin to bolster our Structured Credit offering (launched this week). 

One of our shared focuses is mapping stress and distress (and early signs of) from our bond/loan screeners, Top of the Flops reports and Watchlist to CLO-heavy issuers. The idea is to look for early indications of potential downgrades to triple-hook (subscribers can already seen the latest downgrades in our Navigating the triple-C reports).

The reason that potential downgrades to CCC are an area of interest, is that managers are typically limited to 7.5% of their portfolio as CCC. Above this threshold they will be carried at a discount to par value, which at the extreme’s could trip some of the underlying CLO covenants, such as their overcollateralisation tests. 

Ironically after a series of high profile defaults earlier this year, the average size of CCC buckets has declined slightly in recent months. But as BofA notes in a recent research piece, there is a market difference for deals inside their reinvestment periods (RPs) and those outside RP (tend to be much shorter paper, more on this later). For those inside their RP, only 10% have elevated CCC buckets (6-7.5%).

But this could change quickly, as according to the ratings agencies (very little variance among them) around 70% of CLO collateral is rated at B2/B or B3/B-. BofA assumes in a benign scenario that there are 3% defaults and 30% of B3s are downgraded to triple-hook. 

Another area of interest is loans and bonds trading at discount to par. CLOs face the prospect of having to mark-to-market loans and bonds if they trade at a specified price discount — typically below 75-80 for a certain time period (at least 30-days).

With this in mind we ran a screener of loan prices trading between 75 and 85 (we also added prior month change column to see which are recent entrants). 

There are some interesting names and sectors in the list below. 

We’ve recently written about Huw’s Gray and soon to report on fellow building products sector issuer Xella. Our team are chasing down latest developments at Alloheim and Iberconsa, and are aware of poor numbers at Loparex. Praesidad’s English Scheme handing over the keys to creditors was sanctioned yesterday. However, we are very keen to know the trigger for the recent downward move in price for Veritas.

Bond buckets still remain a small proportion of CLO assets, but are rising sharply, as BofA notes fixed-rate bonds were just 4-5% of CLO collateral in early 2020. But this jumped to over 10% for 2022 vintages, and our Structured Credit team are seeing new deals with buckets in excess of 15%. 

Regular readers of the Flops report won’t be surprised by the whopping 181 EHY bonds from 115 borrowers quoted between 75 and 85, a function of duration and rises in government bond rates. BofA notes that the weighted average cash price of B3 loans is 95.6, whereas B1-rated bonds (two notches higher) is just 87.6. 

To take advantage CLO managers can reduce WARF by switching into bonds trading in the eighties and can even realise some par build too. 

But there is also downside risk. 

High profile CLO names with bonds trading in the low 80s, include Virgin Media, Altice International and Upfield Flora. And 30 of these bonds are due before end-June 2026 — notable high profile names here are Lowell, Talk Talk,and Kloeckner Pentaplast. 

Staying Relevant 

I had a rude awakening on my return from the US, with most of my distressed editorial team on holiday, especially painful in a period of interesting earnings releases and breaking news. I was back to being in full reporter mode. 

We had already received some inbound over the weekend and on Monday relating to SBB’s tender for its bonds via Dutch auction. The number of enquiries increased to our legal team following Fir Tree’s open letter to the company on Monday (more below). 

To recap, the Swedish property management firm announced a voluntary tender last week (16 November 2023), using €600m to retire notes across its maturity curve for prices ranging from 15 cents on the euro for its subordinated hybrid notes to 95 cents for its near-dated floating rate notes. This appeared to be a departure from its previous tactics of using spare cash to redeem near-term maturities. 

Fir Tree issued an open letter on Monday (20 November), flagging its three main concerns with the tender offer:

• That noteholders who take part in the tender offer may be waiving and releasing claims against the company — Fir Tree pointed out ambiguities in the drafting of the tender offer against a “typical tender offer”, which would offer a “reasonable interpretation” of the notes affected by the waiver
• That SBB was focused on preserving equity value — the hedge fund questioned the reasoning behind including the hybrid notes in the tender offer, given that these notes “can be deferred indefinitely” if the company does not pay dividends
• That a tender offer was unnecessary to buy back unsecured bonds — Fir Tree said the company could have just repurchased the notes through the open market rather than via the tender offer

The key point was the wording of the clause — did it apply just to those securities tendered and accepted, or under certain interpretations whether it was still binding if not accepted?

It makes sense to waive and release claims against the company if your debts are repurchased, that is standard in these types of documents. But the wording and the length of the clause was unusual (and included a number of ‘and’s’ to confuse) — the key question was if there was a waiver for ‘all securities’ tendered or just the ‘relevant securities’ — as some parts of the docs referred to ‘the securities’ and some to the ‘relevant securities’. 

Our internal view is that this was probably an unintended drafting error from boilerplate language used for a tender of one or two securities rather the whole debt stack. But we can understand a degree of distrust amongst investors given purported covenant breaches (Fir Tree accelerated on 8 November), changes in accounting treatment and preference share transactions irking bondholders who complained by letter of being primed. 

We did ping the company on Tuesday to find out if they would clarify the language and amend the docs. No response, but our sources told us that SBB did issue a statement via the clearing system clarifying that the terms of the “Renunciation of title and claims [section of the tender offer memorandum]… only applies to those Securities validly tendered or purchased by the Offeror, in line with customary market practice”.

No extension was made to the Wednesday (22 November) deadline — despite the negative noise around the request, and late clarification. 

This morning it announced that it had accepted tenders for €417.247m of its €700m FRNs due 8 February 2024, the shortest-dated notes of those on the list, for a consideration of €403.796m, to settle on 27 November. 

So, less than the €600m maximum, and no deviation from the plan to repay the shorter-dated maturities first, and no repurchases of hybrids. 

Marco Polo dials-up A&E

We expected this Wednesday to be eventful, with Altice France Q3 numbers around midday followed by its conference call. But it became even more lively with the release at 9.40am of Tele Columbus’s restructuring/refinancing/A&E — delete what you think is most appropriate — with a call for investors to be hosted just 20 mins later (without dial-in details — don’t worry, we got in with seconds to spare). 

The distressed German cable and fibre network operator had already flagged a €300m injection from shareholders and their intention to extend maturities to 2028 a week before, but there were still a few surprises for us in the full reveal of Project Marco Polo. 

The company has entered into a lock-up agreement with a majority of its €1.1bn worth of creditors — a €462m term loan due October 2024 and €650m of SSNs due May 2025. The money will not go towards deleveraging, but instead will be invested in the business, mostly for capex. To ease the hefty cash burn, the SSNs will have a 10% PIK-only coupon while the term loan will have a E+400bps (6% floor) margin with a minimum of 50bps cash interest and the remainder as PIK. 

Creditors will see their maturities harmonised, with the TLBs losing their temporal seniority. We assume that the differences in cash pay are a compensation for the TLBs losing their maturity advantage and the SSNs while losing a payment default (as fully PIK) are likely to have a cross-default to mitigate. 

It is interesting that Morgan Stanley is providing the funds rather than United Internet (which owns 40% of the Kublai JV). We and some investors we’ve spoken to have doubts whether this is enough, and most wanted a bigger injection of funds. However, we note there is a mechanism to inject more funds (if needed) — for every €25m of equity provided over the €300m the term loan and SSNs coupon will reduce by 0.5%, with a capped reduction of 2%. 

With capex expected to be hefty — close to €300m per year until 2028 — the company needs to deliver the sharp turnaround in performance promised in its business plan that sees EBITDA doubling from €209m in 2024 to €426m in 2028, and leverage falling from 6.6x to 4.4x.

Our initial reaction was surprise that holders had agreed to an A&E for so little in return, especially given the hefty capex spending needs and the high execution risk on its fibre rollout as its legacy TV business declines. 

We don’t really view this as a par exchange (can’t see how the new notes trade up from the 60s where the existing paper is quoted — they failed to jump on the news). 

That said, there is a recognition among lenders that they don’t really want to own this business and many desperately wanted to avoid a hard restructuring, and therefore were happy to give the sponsor some additional runway. We are aware of several advisor and management changes in recent months that hampered negotiations. 

We suspect they would have liked a higher proportion of holders onboard at launch — the ad hoc group comprises 61% of the term loan and 52% of the senior secured notes— but time is running out as the loans are now current, adding to nervousness of Tele Columbus’ German directors. 

But there could be an earlier exit/deleveraging event, providing better upside. 

We note the separation of the business into a ServeCo / NetCo structure as mentioned in the financial covenants (Find more details on the conditions and other financial covenants here). This would follow an industry wide telco movement towards this business structure. In 2-3 years NetCo could fetch a much higher multiple (into the mid teens) than the ServeCo (5-6x) noted one of the lenders and a source close to the deal. 

But there is still a lot to do in order for the proposed deal roll-out. 

The barriers to entry for a superfast A&E are high, needing 90% agreement from the bonds and 100% from the lenders. Otherwise we are looking at an English Scheme as the preferred route — which requires 75% from those voting (and a numerical majority) to pass — we assume a single-class for bonds and loans if they can argue enough commonality of economic interests. 

Otherwise, they could go down the UK Restructuring Plan or German STARUG routes which have cross class cram down if individual class thresholds are not met and/or an opposing group arises. Look out for a more in-depth analysis in the coming days. 

In brief

Quite a few other notable events and articles over the past fortnight, which I will mention only in brief here but are well worth clicking through and reading more on our platform. 

Its been quite a few weeks for Graanul, the Estonia-based wood pellets manufacturer. The troubles at its US peer Enviva — where my US colleagues first highlighted its potential as a distressed candidate, then a disastrous Q3 after a price bet with its largest customer Drax went wrong — caused a sharp drop in Graanul bond prices earlier this month. 

It also has suffered from a dispute with its largest customer (unnamed but widely known to be Drax), which last week Graanul said had now been resolved (without giving details of the settlement), resulting in a seven point pop in the bonds. The Q3 release yesterday showed further poor performance, and next Monday’s conference call (1pm GMT) shouldn’t be missed. 

Altice France (SFR) announced the sale of its Data Centre business a day before its earnings release. As 9fin’s Yusuf Sule and Nathan Mitchell reported, on the Q3 23 conference call mangement explained that the upcoming €1.05bn worth of early 2025 SSNs maturities are covered with data centre proceeds as well as RCF availability.

SFR announced the sale of a 70% majority stake in UltraEdge (is there someone in the finance team which is a cricket fan?) which holds 257 data centres plus office space, at a price that values the entity at an enterprise value of €764m with gross proceeds of €535m (70% stake sale). The EV is 29.4x UltraEdge’s pro forma EBITDA of €26m. 

But this is well below the €1bn flagged in the press and also includes office space. As 9fin’s Owen Sanderson suggested in a desk conversation, it is effectively a sale/leaseback and reduces secured collateral. Management expects to complete the deal in Q2 24 and that, should an opportunistic refinancing fall short, they would use the €850m availability under the RCF to cover the outstanding balance of the 2025 SSNs

Demire has failed to execute on its ambitious plans to offload €500m of property assets by the end of 2023. In a note yesterday it said it now expects a lower number of property sales than planned at the start of this year. With €499m of SUNs due next October, we expect either a stressed A&E (similar to Accentro) or a full blown restructuring in the coming months. 

As the Swedish online beauty retailer tries to keep the lights-on with an operational restructuring, Oriflame bondholders have organised and appoint advisors. The bonds languish in the 30s. 

Another zombie with no immediate triggers is Consolis, the France-based pre-formed concrete business which posted disappointing Q3 numbers on 16 November, the call transcript is available here. Their May 2026 SSNs continue to drift lower, 47.75-mid at time of writing. 

Victoria Plc’s half-year report also disappointed, with the UK-based flooring manufacturer’s bonds falling around four points on the news. For more detail our earnings write-up is here.

And finally a couple of English Schemes were sanctioned by Justice Norris this week. 

The first was for Praesidiad which amends and extends the Spanish perimeter fencing group’s existing indebtedness into a super senior TLB (maturing June 2026), a TLB (maturing December 2027) and a senior TLB (maturing September 2027). There is a transfer of ownership of the group from Carlyle to the existing SFA lenders.

The second is for Lecta, in which the Spanish paper manufacturer will exchange its two FRN indentures into a single tranche of new senior secured notes (Euribor +650bps) maturing in Q3 2028 and €78m in new senior secured notes (Euribor +800bps) maturing in Q1 2028, ranking senior to the new senior secured indenture.

What we are reading/watching this week

After the events of the past 18 months it is probably no surprise that another real estate company has come into the cross-hairs of short sellers. 

Muddy Waters’ CPI Property Group report has a lot of similarities to other situations, such as Aggregate and Adler, with allegations of related party transactions paid via receivables and then sold back to the company. 

This includes undeveloped plots of land in Prague bought for less than €20m which yielded a profit of €50.9m, an equity squeeze out at an alleged inflated price which benefited CPI founder Radovan Vitek’s son to the tune of €52.1m and the use of the property group to finance a new super yacht for Vitek’s personal use. 

We will be looking closer into the allegations in the coming days. The company has issued an initial response:

This suggests there are more allegations to come from MW. I suspect my weekend reading will mostly comprise CPI Property Group financial statements. 

A lot of assumptions have been made about the effects of liquidity on risk assets and markets in general. Here is a great post by Macro Alf (highly recommend you follow) on how it exactly works. 

I spent a lot of time this week in conversations about hidden stress in Private Credit and how to identify it. But most of the column inches were about banks seeking to move into this space. Am I the only one that she’s the irony that banks want to use private credit for relationship lending? Or in the increased willingness for direct lenders to work in club transactions — with Adevinta back on —  talk about going back to the future!

I would highly recommend The Athletic’s series on “Crisis Clubs” examining the financial states of five European Football clubs. They have already published on Everton, Barcelona, and Inter Milan — but the best so far is for Herta Berlin — it has everything, KKR, Lars Windhorst, private debt, terrible governance and 777 Investments the latest buyer (Everton too) with opaque funding.

Since the last Workout, Brighton had another disappointing 1-1 draw in a game that they dominated against Sheffield United until a red card — watched from my hotel lobby on an NBC stream (its incredible how many games you can watch in the US for $5.99 per month). 

Yet more injuries too, we are down to one fit left back and that’s Pascal Gross!

But my disappointment was tempered by front row seats at the Brooklyn Nets — which beat the Wizards in a tight game. I particularly liked Mr Whammy, the elderly superfan, who seeks to put the hex on opposition free throw shooters