To The Moon

$Grab Holdings(GRAB)$ ’s stock has had a tough few months.

The market did not like the acquisition spree, with Superbank, Stash, Foodpanda Taiwan and now Atome, while fuel prices stayed high throughout the quarter.

On top of that, Grab suffered a setback in October, when Taiwan’s Fair Trade Commission rejected its $600M acquisition of Foodpanda Taiwan. Not only did Grab lose a deal I called smart in Q1, it will also have to pay a $60M break fee in Q4.

However, the Q3 2026 results were the strongest Grab has ever reported.

• Revenue growth accelerated to 27% Y/Y, or 30% FXN.

• Record ADJ EBITDA of $237M, +73% Y/Y.

• Financial services turned ADJ EBITDA profitable for the first time.

• Full-year guidance raised for the second time this year.

For the first time in Grab’s history, all three core segments, mobility, deliveries and financial services, are profitable on a segment ADJ EBITDA basis.

But not everything is great.

Net income was only $18M, reported GMV growth slowed down, and incentives hit another record. Also, a big part of the financial services growth came from acquisitions, not organic growth.

Overall Results

• Revenue $1.11B +27% Y/Y

• ADJ EBITDA $237M +73% Y/Y

• EBIT $108M vs $27M last year

• Net Income $18M vs $17M last year

• Basic EPS $0.01 vs $0.01 last year

• ADJ FCF $286M vs $203M last year

Revenue

Total revenue grew by 27% Y/Y and 30% FXN to $1.11B!

This quarter, the currency situation flipped compared to Q1. In Q1, the weak dollar helped Grab, while in Q3, Southeast Asian currencies were weaker against the dollar than a year ago, so local currency growth was higher than reported growth.

As you can see in the chart above, this was a strong acceleration from 22% growth in Q2 2026 and 24% growth in Q1 2026.

The total revenue of $1.11B beat Wall Street expectations of around $1.07B.

By segment, deliveries revenue grew 21%, mobility only 10%, while financial services revenue more than doubled, growing 124%.

However, we need to be careful here. Part of this growth is not organic. Grab started consolidating Superbank in June and Stash from this quarter, and both add revenue to the financial services segment. Without them, the acceleration would be much smaller. Grab does not disclose exactly how much they contributed, which I don’t like.

ADJ EBITDA and ADJ FCF

ADJ EBITDA was a record $237M, up from $136M in Q3 2025, a 73% increase Y/Y!

This is a huge jump from $168M in Q2 2026, +41% Q/Q.

Moreover, the ADJ EBITDA margin expanded to 21.3% from 15.6% a year ago and 16.9% last quarter.

Total segment ADJ EBITDA grew 46% to $338M, while regional corporate costs were $101M, falling to 9.2% of revenue from 10.8% a year ago. This is operating leverage at work, as revenue is growing much faster than corporate costs.

Financial services also swung from a $28M loss to a $12M profit, a $40M improvement, which explains a big part of the jump.

Furthermore, looking at cash flow, ADJ FCF was $286M vs $203M in Q3 2025.

On a trailing twelve-month basis, ADJ FCF reached $533M, nearly double the $283M a year ago, and up from $489M in Q1.

A reminder that Grab’s financial services segment distorts normal FCF, so Grab shows ADJ FCF, which removes changes in loan receivables and deposits. Starting from Q2 2026, Grab also excludes net changes in treasury liquidity positions in the financial services segment.

Normal operating cash flow was $55M vs -$127M last year. This was driven by a $404M bigger inflow of customer deposits at the digital banks, partially offset by $294M more loan disbursals.

Net Income and EPS

Grab reported a Net Income of $18M for the quarter, basically flat compared to $17M a year ago.

This looks strange, considering that operating profit more than tripled from $27M to $108M.

The gap is explained by non-operating items:

• $57M net fair value loss on financial assets and liabilities

• $12M impairment loss on an equity-accounted investee

• $12M lower finance income

• $5M higher income tax expense

In Q1, I explained how the $1.5B convertible note makes Grab’s net income volatile. In Q1, the share price fell, and Grab booked a gain. This quarter, the net fair value change was a loss. Grab doesn’t break down how much of the $57M comes from the convertible notes vs other financial assets.

Basic EPS was $0.01, below the $0.02 that Wall Street expected. Excluding the non-cash items, Grab says profit would have been $95M or $0.02 per share.

Year-to-date, Grab’s net profit is $373M. However, $307M of that is a one-time, non-cash gain from remeasuring its previously held stake in Superbank in Q2.

This is exactly why I said in Q1 that P/E is not a usable metric for Grab. In Q1, net income was boosted by a convertible note gain, in Q2 by the Superbank revaluation, and in Q3 it was dragged down by fair value losses. None of these says anything about how the business is doing.

The good news is that EBIT, which was only $22M in Q1, is now $108M. EBIT is a much cleaner number, and it’s moving in the right direction.

Subscribe

Guidance

Grab raised its full-year guidance for the second time this year!

The company now expects revenue to grow between 23% and 25% to $4.15-4.20B, up from $4.04-4.10B at the start of the year.

They also expect ADJ EBITDA to grow by 58% to 60% to $790-800M, up from $720-740M after Q2 and $700-720M originally.

However, if we do the math, Grab has made $3.06B in revenue and $559M in ADJ EBITDA in the first three quarters. That implies Q4 ADJ EBITDA of $231-241M, essentially flat vs Q3.

Q4 is usually one of Grab’s strongest quarters, so analysts asked about this on the call. The CFO explained that Q4 is a “high return investment window” with the festive season and tourist travel, and that with elevated fuel prices, Grab needs more drivers on the platform. Essentially, Grab plans to spend more on incentives.

I think management is being conservative here, leaving room to beat. But it also tells us that incentives will stay high.

Grab also reiterated its 2028 targets of 30%+ revenue CAGR and $1.7B in ADJ EBITDA. These include the proposed acquisition of Atome, which I will discuss in the M&A chapter.

GMV

In Q3 2026, On-Demand GMV grew by 18% Y/Y to reach $6.8B.

On FXN basis, the growth was 22%, in line with Wall Street’s ~$6.81B estimate.

Reported GMV growth slowed from 21% in Q2 2026 and 24% in Q1 2026. However, FXN growth was stable at 22% vs Q2 and up from 21% in Q1. So the slowdown is mostly a currency effect, not weaker demand.

The growth in GMV was again driven by two things.

First, Group monthly transacting users reached a record 57M, up from 54M in Q2 and 51.6M in Q1. On-Demand MTUs grew 16% Y/Y.

Second, GMV per MTU grew only 2% Y/Y, but on FXN basis it grew 5%, slightly better than the 4% FXN growth in Q1.

Grab’s loyalty programme is a big part of this. More than 1 in 5 deliveries users is now a GrabUnlimited subscriber, these subscribers drive 35% of deliveries GMV, and they order almost 3x more often than non-subscribers.

Grab is also bringing in new users with cheaper products. In mobility, transactions grew 21%, while GMV grew only 14%, which means the average ride is getting cheaper. More users, but at lower prices.

Delivery and Mobility

Deliveries GMV grew by 21% or 24% FXN to $4.5B, while revenue grew 21% or 22% FXN to $561M.

Mobility GMV grew by 14% or 17% FXN to $2.3B, while revenue grew 10% or 12% FXN to $349M.

Deliveries revenue beat Wall Street estimates of around $540M, while mobility revenue came in slightly below the ~$353M expected.

The deliveries take-rate stayed at about 12.5%, roughly the same as last year, but lower than 13.1% in Q1. Meanwhile, the mobility take-rate fell from about 15.5% to 14.9% Y/Y.

This was the first quarter with the new Indonesian commission cap for two-wheelers, which took effect in July, combined with higher driver incentives. Both pushed the mobility take-rate down.

Looking at Segment ADJ EBITDA, deliveries grew by 58% Y/Y to $123M, while mobility grew by 15% to $207M!

Deliveries ADJ EBITDA margin improved to 2.7% of GMV from 2.1%, while mobility was stable at 8.9%. Both improved Q/Q.

In Q1, I said that deliveries is growing at almost double the pace of mobility and will likely overtake it in the long term. This quarter, deliveries ADJ EBITDA grew almost 4x faster than mobility. Deliveries ADJ EBITDA is now 59% of mobility’s, up from 44% in Q1.

The deliveries margin expansion came from operating leverage and advertising. Quarterly active advertisers grew 39% Y/Y, but their average spend grew only 13%, a big slowdown from the 44% growth in Q1.

GrabMart continues to do well. Groceries GMV grew 1.8 times faster than food, and 16% of food users also ordered through GrabMart. Grab also launched GrabMall to expand into branded retail, beauty, and consumer electronics.

In mobility, the number of active driver-partners grew 20% Y/Y to another all-time high. Over 70% of two-wheeler drivers now work across both mobility and deliveries, which helps keep them busy between peak periods.

Fuel Prices Update

In Q1, I wrote about the Iran War oil shock. Fuel prices stayed elevated throughout Q3, and management expects the same in Q4.

Grab continued with its countermeasures:

• Driver incentives: Grab continued paying extra incentives to make sure driver earnings keep pace with fuel costs.

• AI Driver Assistant: The tool now reaches nearly 500,000 drivers, and those who actively use it earn over 10% more per online hour.

• EV transition: EVs now account for 18% of four-wheel trips, up from 13% in Q1. Grab is financing EVs through its digital banks, including GXS Bank in Singapore and Superbank in Indonesia.

In Q1, management said that incentives would peak in Q1. This did not happen, and I will expand on this later.

Take-Rate Law Change in Indonesia Update

The new commission framework for two-wheeler drivers in Indonesia took effect in July. Grab has implemented it and says it continues to engage closely with the government.

Grab doesn’t disclose how much of the 60bps decline in the mobility take-rate came from Indonesia, but it certainly contributed.

What worries me more is that, as I predicted in Q1, the regulation is now spreading. The Indonesian government is now evaluating regulatory frameworks for on-demand logistics and food delivery, with the Ministry of Transport and the Ministry of Manpower involved.

According to Grab, these guidelines will focus on driver earnings and fee transparency. Management doesn’t expect them to materially impact guidance, but the guidance only covers Q4. The real question is what happens in 2027.

Fee transparency is the interesting part. In Q1, I said Grab could offset the commission cap by increasing platform or service fees. If the government forces “transparent” fees, this escape route could become much narrower.

Financial Services

Financial services revenue grew by 124% Y/Y to $201M, or 132% FXN!

We can see in the chart above that this is a massive acceleration from 43% in Q1 2026 and 59% in Q2 2026. It also beat Wall Street estimates of around $173M.

In Q1, I complained that at a $107M quarterly scale, this segment was growing too slowly. Well, Grab clearly heard me, but it got there mostly through acquisitions. Superbank has been consolidated since June, and Stash since this quarter. Grab doesn’t split out how much revenue they contributed.

The best organic indicator we have is the gross loan portfolio, which grew 200% Y/Y to $2.7B. Excluding Superbank, it still grew 115%. That is very strong organic growth.

Loan disbursals grew 115% to $1.9B, up from $1.05B in Q1. The gross loan portfolio is also well above the $2B year-end target Grab set in Q1, although that target was set before Superbank was consolidated.

Credit quality looks fine for now. Non-performing loans (90 days past due) were 2.6% of the portfolio and have stayed within 2-3% since the start of 2025. Expected credit loss reserves are 6.5% of the portfolio, covering NPLs about 2.4 times.

Grab books credit loss reserves when it disburses a loan, so fast growth front-loads provisioning. As the loan book matures, Grab expects its lending risk-adjusted margin to improve from 7% today to 10-12% at steady state.

Customer deposits grew to $2.7B from $1.6B in Q1, mostly because Superbank is now included.

Most importantly, the segment turned ADJ EBITDA positive for the first time, at $12M vs a $28M loss a year ago and a $17M loss in Q1.

In Q1, I questioned whether prioritising break-even over deposit and market share growth was the right strategy. Now we have the answer: Grab got both profitability and growth, but it partly bought the growth.

And it’s not done buying. In September, Grab agreed to buy a 60% stake in Atome, a Southeast Asian buy-now-pay-later and consumer lender, for $1.49B in cash. I will explain this deal in the M&A chapter.

Incentives

In Q3 2026, total incentives reached a new record of $737M, compared to $650M in Q1!

On-Demand incentives as a percentage of On-Demand GMV rose to 10.7%, up 64bps Y/Y, but down 17bps Q/Q.

A reminder that in Q1, management said that Q1 would be the peak for incentives and that they would go down in the next quarters.

This didn’t happen. In dollar terms, incentives kept rising from $650M in Q1 to $737M in Q3. As a percentage of GMV, they rose to around 10.9% in Q2 and only slightly decreased to 10.7% in Q3, still above the 10.5% in Q1.

Grab says that the increase was due to continued investment in driver supply, as fuel prices stayed high, and to drive the adoption of new products.

Also a reminder that under IFRS accounting, incentives are recorded as a reduction of revenue. This is one of the reasons why mobility revenue grew only 10% while mobility transactions grew 21%.

And based on the Q4 guidance, incentives will likely stay high. I don’t love this, but with ADJ EBITDA growing 73%, Grab can clearly afford it.

M&A

This year, Grab has been much more active in M&A than ever before. It consolidated Superbank, bought Stash, tried to buy Foodpanda Taiwan, and now agreed to buy Atome.

Two of these deals had big news this quarter, so let’s go through them.

Atome

On 15 September 2026, Grab agreed to buy a 60% controlling stake in Atome Financial for $1.49B in cash.

Atome is a Southeast Asian buy-now-pay-later and consumer lending platform. It offers BNPL loans, consumer cash loans, BNPL cards and other digital lending products. It operates in 5 markets, and according to reports, it generates around $470M in revenue, growing at about 80%.

The deal is structured in two phases:

• Phase 1: Grab buys 60% for $1.49B in cash. This is expected to close by Q3 2027, subject to regulatory approvals.

• Phase 2: Grab reportedly has an agreement to buy the remaining 40% in 2029, at a valuation between $2B and $4.5B for the whole company, depending on Atome’s performance.

$1.49B for 60% values Atome at about $2.5B, or roughly 5.3x revenue. If the second phase lands at the top of the range, the remaining 40% would cost another ~$1.8B, bringing the total to over $3B.

Why does Grab want Atome?

• Lending scale: Building a consumer lending business from scratch takes years and comes with large credit losses while you learn. Atome already has a proven underwriting engine.

• New users and merchants: Atome brings a large user base and a network of over 30,000 brands across Grab’s core markets.

• Faster expansion: Atome speeds up Grab’s lending expansion into the Philippines, Indonesia and Thailand.

Grab is already working with Atome commercially. In October, it embedded Atome’s cash lending into the Grab app in Indonesia.

Atome is also a big part of Grab’s 2028 targets. Grab expects financial services to reach $500M in ADJ EBITDA in 2028, with about $300M of that coming from Atome and Superbank. The loan book is expected to exceed $6B, including Atome. The group target of $1.7B ADJ EBITDA in 2028 also includes Atome.

This means that a big part of Grab’s 2028 story now depends on a deal that hasn’t closed yet. If regulators block it, as happened in Taiwan, the 2028 targets will need to be cut.

I understand the strategic logic, and Atome fits well into Grab’s ecosystem. But I don’t love the price. Paying ~5x revenue for a consumer lender, with a potential second payment of up to $1.8B, is expensive. Consumer lending in emerging markets is also risky, as credit losses can rise quickly when the economy slows down, and the region is already dealing with high fuel prices and inflation.

The market clearly shares these concerns, as the stock fell after the deal was announced.

Foodpanda Taiwan

In Q1, I called the Foodpanda Taiwan acquisition a smart deal and an interesting opportunity.

A quick reminder of the deal. In March 2026, Grab agreed to buy Delivery Hero’s Foodpanda business in Taiwan for $600M in cash. It would have been Grab’s first expansion outside of Southeast Asia.

Taiwan’s food delivery market is basically a duopoly between Foodpanda and Uber Eats. Foodpanda Taiwan reportedly generated $1.8B in GMV last year, and the deal would have given Grab slightly more than half of the market. Morningstar estimated it would have increased Grab’s deliveries GMV by about 13% and added about $60M to 2028 ADJ EBITDA.

Unfortunately, on 21 October, Taiwan’s Fair Trade Commission rejected the acquisition.

What makes this decision interesting is that Grab doesn’t operate in Taiwan at all, so this was not a classic case of two competitors merging. Instead, the regulator focused on Grab’s relationship with Uber.

Uber owns about 13.1% of Grab’s shares, a stake it received when it sold its Southeast Asian business to Grab in 2018. Even though Grab’s dual-class share structure limits Uber’s voting power to 3.9%, the regulator said that the equity ties, overlapping board membership and history of business cooperation would reduce Grab’s incentive and ability to compete with Uber Eats.

The logic is that if Grab cut prices to win market share from Uber Eats, it would hurt Uber Eats, but if Uber Eats fought back with discounts, it would hurt Grab’s profits and therefore Uber’s investment. So neither would compete aggressively.

This is the second time the regulator blocked a sale of Foodpanda Taiwan. In December 2024, it blocked Uber Eats’ own $950M acquisition of the business, and Uber abandoned the deal in March 2025.

Honestly, I think this decision is debatable. Without a strong buyer, Foodpanda Taiwan stays with Delivery Hero, which has been under pressure to sell assets. A weaker Foodpanda could end up strengthening Uber Eats even more. The regulator said it doesn’t make decisions based on hypotheticals.

Grab and Delivery Hero have the right to reapply with changes to the deal. However, the main issue is Uber’s stake in Grab, which Grab can’t easily change, so I don’t expect a second attempt soon.

The financial impact is limited. Grab will pay a $60M break fee, which will be recorded in Q4 2026. It will hit IFRS net income, but it is excluded from ADJ EBITDA and has no impact on the guidance.

On the call, the CFO said Grab was disappointed but “anticipating it”. He also pointed out that the rejection adds about $550M of cash back to the balance sheet, which can now be used for buybacks, Atome, or organic investments.

The bigger lesson for me is about the Uber relationship. In Q1, I saw Uber’s stake as mostly neutral. Now it has cost Grab its first international expansion, and it could become an obstacle for future deals outside Southeast Asia as well.

So, while this is disappointing, it is not a disaster. Grab loses a growth opportunity, but keeps the cash.

Capital Allocation

Grab repurchased $400M worth of shares in September and October, and it still has $500M remaining under its share repurchase programmes.

Once completed, at the current share price, Grab’s cumulative buybacks since the start of 2024 will reach $1.75B, retiring more than 10% of its shares outstanding.

The company ended the quarter with $7.1B in gross cash liquidity and $5.0B in net cash liquidity.

CEO Anthony Tan also said he personally bought additional Grab shares. Insider buying is usually a good signal.

Conclusion

Grab delivered its best quarter ever!

The company reached a record $237M in ADJ EBITDA, saw its revenue grow by 27%, and for the first time, all three core segments are profitable.

Despite elevated fuel prices and the new commission cap in Indonesia, Grab kept growing its users and drivers to all-time highs, and raised its guidance for the second time this year.

However, there are a few things I don’t like.

First, a big part of the revenue and financial services acceleration came from acquisitions, and Grab doesn’t disclose how much.

Second, incentives didn’t peak in Q1 as management promised, and the Q4 guidance implies they will stay high.

Third, the regulatory risk in Indonesia is spreading from two-wheeler mobility to food delivery, just as I feared in Q1.

Lastly, the acquisition spree, especially the $1.49B Atome deal, raises questions about capital allocation.