Uber's €14.2bn Delivery Hero bridge: the clock starts 13 Nov

Uber has committed more money to a single acquisition than it currently owes in total, and the agreement records that none of it had been drawn.

On 16 July it agreed to buy Delivery Hero SE — the Berlin food delivery company — for €41.50 a share in cash, and the same day signed a bridge credit agreement for €14,200,000,000. The filing is blunt about the state of it: “No borrowing has been drawn on the Effective Date.” The offer is not expected to complete until the second half of 2027.

That gap between signing and needing the money is the whole story. Bridges are built to be abandoned, and the terms are written to make sure you abandon them. Three weeks later Uber started.

What Uber actually signed

Two filings, four agreements. The 8-K of 7 August is the interesting one, because it is where the July structure begins coming apart on purpose.

The euro term loan did not add money. The 8-K says so in one line: entering into it “reduced the commitments under the Bridge Credit Agreement by €4,000,000,000.” Same total, different lenders’ patience — 28.2% of the bridge became something with a longer fuse.

The revolver is the part that genuinely grew, from $5.0 billion to $7.7 billion, a 54% increase, with the maturity pushed from September 2029 out to August 2031.

Put plainly: the acquisition facilities alone are three times the unrestricted cash and short-term investments Uber reported at the end of June.

The date that did not move

Here is the detail you only get by holding the two filings side by side.

The bridge charges a commitment fee “commencing 120 days after the Effective Date.” The effective date is 16 July 2026. Count it out.

16 July → 31 July      15 days
August                 31 days
September              30 days
October                31 days
1 → 13 November        13 days
──────────
120 days   →  13 November 2026

Now read the term loan signed on 6 August. Its commitment fee commences — the agreement names the date rather than a period — on 13 November 2026.

The same day. Uber moved €4.0 billion out of a 364-day bridge and into an instrument that runs three years, and the meter still starts on the same Friday. The term loan inherited the bridge’s clock instead of resetting it.

What Uber did buy with the exchange is the back end. The bridge charges duration fees on the 90th, 180th and 270th days after closing if commitments are still outstanding — an escalating penalty for staying. The term loan’s disclosure carries no such fee, and its Tranche B does not come due until three years after a closing that is itself more than a year away.

There is a second concession in Amendment No. 1, and it is not cosmetic. The bridge’s cross-default event of default was replaced with a narrower cross-payment-default-and-acceleration trigger, and the threshold rose from $300 million to $500 million. A borrower about to carry a great deal of debt asked for more room before somebody else’s default becomes its own, and got it.

One ratio, three agreements

Read the covenant paragraph in the bridge, then in the term loan, then in the revolver. It is the same sentence three times: maintain a ratio of consolidated adjusted EBITDA to consolidated interest expense of not less than 3.00 to 1.00.

That is worth sitting with. A €14.2 billion acquisition facility, a €4.0 billion euro term loan and a $7.7 billion revolver are all governed by one number, and it is the same number Uber was already living under before any of this.

So is 3.00 to 1.00 tight? Run it on the reported first half.

Operating income              $3,813M
Depreciation and amortisation   +$372M
─────────
EBITDA proxy                   $4,185M
Interest expense                 $235M
Ratio                             17.8   (covenant floor: 3.00)

Nearly six times the floor. And that proxy is the conservative one — add back the $1,023 million of stock-based compensation that credit agreements almost always permit and the ratio goes to 22.2.

Turn it around and ask what would actually breach it. Annualise the proxy at $8,370 million; three times coverage permits $2,790 million of annual interest expense; Uber’s current run rate is about $470 million. That leaves roughly $2,320 million of annual interest to spare.

Permitted interest at 3.00×    $2,790M
Current run rate                 -$470M
─────────
Headroom                       $2,320M
$2,320M ÷ $16,360M  =  about 14.2% a year

Euro-denominated investment grade paper is not priced anywhere near 14%. The covenant is not the constraint here — which is precisely why it survived unchanged across three agreements. It is a tripwire for a different kind of trouble than the one this deal presents.

The constraint is the calendar. Fees begin on 13 November 2026, the acceptance condition and the regulatory approvals sit somewhere in 2027, duration fees start biting 90 days after a closing that has not happened, and the bridge must be gone 364 days after it. Meanwhile Uber’s own $2.0 billion 2026 Term Loan matures in December 2026, four months after the August filings — a maturity that predates the Delivery Hero deal entirely and has to be dealt with anyway.

None of that means the bridge gets used. Baker Hughes committed to one of up to $14.9 billion for its purchase of Chart Industries in July 2025, replaced it with bonds and term loans over the following eight months, and terminated the remaining commitments four months before the deal closed — which is what a bridge is designed to do when the calendar cooperates.

Worth knowing that a covenant framed on EBITDA can look comfortable while the cash picture is tighter, because EBITDA excludes exactly the two costs that debt-funded expansion creates — depreciation on what the money bought, and interest on the money itself. Uber’s case is unusual in that its capital spending is tiny, $135 million in the first half, so the depreciation half of that objection barely applies. The interest half is the entire subject of this article.

Uber already owns a quarter of the target

The takeover is not starting from zero, and the 10-Q is where you find that out rather than the 8-K.

Uber acquired 8.4 million Delivery Hero ordinary shares in May 2024 and added to the position during the second quarter of 2026, crossing into significant-influence territory at 24.99% — a hair under the quarter mark. On top of the shares it holds swaps referencing the same stock, giving it economic exposure of another $1.5 billion at quarter end.

Set that against the offer’s minimum condition: acceptance by holders of at least 50% of the shares plus one, counting shares held by or attributed to the bidder and its affiliates. Uber is starting a quarter of the way there. The tender has to deliver a bit over 25%, not 50%.

There is a mirror to this. The transition to equity method accounting produced a $1.1 billion gain in the second quarter, which flowed through other income and helped lift quarterly net income to $2.4 billion. The accounting gain on the stake is not cash and does nothing for the covenant, which is measured on EBITDA and sits below that line.

What the filings do not tell you

The pricing. Both euro facilities bear interest at EURIBOR plus a margin that “will fluctuate based upon the Debt Rating,” and neither 8-K gives the grid. The commitment fee, the funding fee and the duration fees are all described the same way — named, tied to the rating, unquantified. The revolver’s margin over term SOFR and its commitment fee are likewise rating-linked and undisclosed.

So we can tell you when the fees start and cannot tell you what they are. That is the filing’s limit, not an omission on our part, and the numbers will surface in the interest expense line of a future 10-Q rather than in any announcement. When they do, they land in a market where the median large US borrower’s interest costs have been rising even as policy rates came down, and where the instrument a company picks decides which line of the financial statements the cost shows up on.

What would make this wrong

• Our covenant ratio is a proxy. “Consolidated adjusted EBITDA” and “consolidated interest expense” are defined in credit agreements filed as exhibits, which we did not read. Real definitions include add-backs and our figure is the conservative version. The direction of the error is toward more headroom, not less.
• The 14.2% break-even is illustrative arithmetic, not a scenario. It holds EBITDA flat, draws the entire bridge, and ignores Delivery Hero’s own earnings and its own debt — some of which the term loan proceeds are earmarked to refinance. Post-closing the ratio is tested on a combined company we have no consolidated figures for.
• Currency. Every euro conversion here uses one dated rate, 1.1521 dollars per euro on 6 August 2026 from FRED. The debt is euro-denominated and most of Uber’s cash flow is not; a euro acquisition funded with euro debt is a defensible hedge, and we take no view on it.
• “Not drawn” describes the filing dates, 16 July and 6 August 2026. Nothing here is a statement about the position today, which will not be visible until the next 10-Q.
• Upfront and arrangement fees are not disclosed in either 8-K. The 13 November date is when the commitment fee begins, not necessarily when the first dollar of cost was incurred.
• The deal may not complete. Completion needs competition and financial-services approvals and the minimum acceptance condition, and the agreement contemplates termination scenarios including a regulatory long-stop of 10 May 2028.

One piece of that package was not about this deal

The $7.7 billion revolver in the 7 August filing replaced an agreement dated 26 September 2024 and matures on 6 August 2031. That date puts it in company: between 31 July and 17 August, five other large US filers replaced revolving credit lines, and every one of the six new facilities matures in 2031 — mostly at the same size as the line it replaced. We went through all six separately. Uber’s is the outlier in that group, because its revolver arrived attached to acquisition financing rather than as a routine renewal.

Check it yourself

Both 8-Ks are short and neither hides anything. The bridge amount, the 120-day fee period, the 90/180/270-day duration fees, the mandatory prepayment sweep of net proceeds from equity, debt or asset sales, and the 3.00 to 1.00 covenant are all in Item 1.01 of the 16 July filing. The €4.0 billion reduction, the two tranches, the 13 November date, the $300m-to-$500m threshold change and the $7.7 billion revolver are all in Item 1.01 of the 7 August filing.

The 24.99% stake, the total return swaps and the $2.0 billion term loan maturing in December 2026 are in Notes 3 and 5 of the 10-Q for the quarter ended 30 June 2026. Operating income, depreciation, interest expense and stock-based compensation are one API call to SEC company facts. The exchange rate is one series at FRED.

Every calculation in this article is written out in full above and can be redone on paper.

If a figure here does not match a filing, tell us and it will be corrected on the article and on the corrections log, with the date.