Discord

IWhat it does

Discord is a chat, voice and video app. People join communities called servers, and each server has text channels and voice channels. It began as a place for gamers to talk while they play. The company says more than 90 million people use it every day, as of the fourth quarter of 2025³. Press reports and third-party profiles put monthly users above 200 million⁶¹⁵. Those are company claims, not audited numbers, and the filing that would define them has not been made public.

Most of the time the app is not a public square. It is a set of mostly private rooms: friend groups, game clans, study groups, fan communities, creators and their followers. The company’s own page lists the paid features as Nitro subscriptions, Server Boosting and Orbs, which is a reward currency earned by completing Quests, plus a pilot of social commerce tied to a game³. A newer layer is public discovery. Discord says more than 63,000 servers qualify as Discoverable, which needs at least 1,000 members and an age of at least eight weeks⁵.

Discord is not publicly traded. It has not sold shares on any exchange, and EDGAR, the SEC’s public filing system, shows only six old Form D notices for Discord Inc. (dated 2019 to 2023) and no registration statement or prospectus¹². Form D is a short notice a private company files after selling shares to private investors. A confidential IPO filing, if one exists, does not appear on EDGAR until the company chooses to make it public. So this page is about a private company that has been reported to be preparing to list.

IIHow it makes money

Discord has never published audited financials. Everything below is therefore either a company statement about product, a press report, or a third-party estimate. I keep those labels separate.

Subscriptions come first. Nitro costs $9.99 a month and Nitro Basic costs $2.99 a month⁹. Discord’s chief technology officer has said Nitro is the biggest contributor to revenue and has grown year on year, but gave no figures⁹. In May 2026 Discord added the base tier of Xbox Game Pass to Nitro, but not to Nitro Basic, as a reason to pay for the higher tier⁹. The subscription has been the core for years: CB Insights described Nitro at $9.99 a month and a cheaper Nitro Classic at $4.99 in August 2021¹³.

Other lines are smaller or newer. Server Boosts, which let members pay to unlock perks for a server, cost $4.99 a month each in CB Insights’ August 2021 description¹³. The same report said Discord took a 10% cut of games that developers sold through verified servers¹³. I could not confirm that either is unchanged in 2026. Sacra, a research firm, also lists a 10% platform fee on in-app purchases through the Embedded App SDK¹⁵. Advertising is the newest line. Discord’s Quests let a game publisher pay to promote a game, and users earn Orbs for engaging. A Discord product executive said more than 70 Quests reached a 10% acceptance rate over the prior year, per TechCrunch in May 2025⁸. That is a conversion signal, not a revenue number.

On size, the only reported revenue figure I could find is old: $130 million for 2020, up from about $45 million in 2019, which the Wall Street Journal reported in March 2021 and trade press repeated¹⁴¹³. It is press-reported, not a filed number. Sacra estimates about $600 million in 2023 and about $725 million of annualised revenue at the end of 2024, up about 21%¹⁵. This is an outside estimate, not a filed number. Sacra also says Discord reported positive adjusted EBITDA for five straight quarters as of April 2025¹⁵. Adjusted EBITDA is earnings before interest, tax, depreciation and amortisation, with the company’s own add-backs, so it is a softer number than a profit under accounting rules. I did not find a primary source for it.

Here is how I think about the unit economics, with the caveat that these are my assumptions. A subscription app with a free tier has a simple shape. A small share of users pays every month, and the rest provide the audience that makes the paid features worth having. If I take the company’s 90 million daily users³ and Sacra’s roughly $725 million run rate¹⁵, the revenue per daily user is under $8 a year. That is low next to the public social companies in the comparison table, and it cuts both ways. It means the audience is large relative to the revenue, which is room to grow, and it also means the audience is mostly not paying and not yet worth much to advertisers. I cannot split the $725 million between subscriptions, boosts, commerce and advertising, and the split matters a great deal for the multiple a public investor would pay.

One more point on quality of revenue. Subscription revenue from a social product is only as steady as the habit behind it. A user who stops opening the app stops renewing. Advertising revenue depends on advertisers trusting the audience and on the platform being able to target and measure, which gets harder if users are younger, pseudonymous and anxious about data. Both lines therefore depend on the same thing, which is attention, and I would want to see retention data in any prospectus before leaning on either.

Revenue and user figures, with how reliable each is

IIIThe market and the competition

Discord competes for time and for community. It does not have one clean rival. Sacra names Steam chat and console voice systems, low-latency voice tools such as TeamSpeak and Mumble, Roblox’s Guilded, and broader tools such as Slack and Telegram¹⁵.

The better way to think about it is as three markets. The first is gaming voice and chat, where Discord is the default for many groups and rivals are mostly features inside other products. The second is community hosting for creators and fan groups, where Telegram, WhatsApp groups, Reddit and Slack overlap. The third is advertising and commerce around those communities, where Discord is a small and young seller next to Reddit, Snap, Pinterest and Roblox. For scale, those public companies trail twelve-month revenue of $3.1 billion (Reddit), $6.4 billion (Snap), $4.7 billion (Pinterest) and $5.9 billion (Roblox)^16171819. My estimate of Discord’s 2026 revenue, roughly $1.0 billion, would be a fraction of any of them. That is my estimate, not a reported figure.

The IPO list on StockAnalysis shows that the market for new listings is open in 2026: SpaceX, Cerebras, SK hynix and PayPay all listed, though several recent listings trade well below their offer price²¹. Discord does not appear in that list of recent listings²¹.

A fair question is who would pay for a Discord share and why. A public investor would likely compare it with Reddit, which also hosts communities and has built an advertising business, and with Roblox, which hosts young users and games. Reddit’s trailing revenue is about $3.1 billion and Roblox’s about $5.9 billion¹⁶¹⁹. Discord would arrive as a smaller company than both, with a more private product and less proven ad revenue. In its favour, it reports a very large daily audience³, it has a paid tier that was already well established by 2021¹³, and Sacra says it has had positive adjusted EBITDA¹⁵. Roblox shows a gross margin of 27% in the data I pulled¹⁹, which is a reminder that user-generated platforms can carry heavy payout costs. Discord’s costs of running voice and video at scale are not public, and I cannot say which end of that range it sits at.

IVThe moat and how it could be disrupted

Discord’s advantage is that groups are already there. A friend group or a clan has its channels, roles, bots, history and habits in one place. Moving means rebuilding that and persuading every member to follow. That is a switching cost for the group, and the group is what holds each member. It is also a network effect, but a narrow one: it works within each community, not across the whole network the way a phone network does.

I rate the moat contestable, not strong. The reasons are plain. The product is free for most people, so the lock-in is social, not financial. Voice and chat are well understood and any big platform can copy them. Many servers are run by volunteer moderators, and if the habit moves, the revenue goes with it. Discord’s own recent decisions show how thin the loyalty can be: when the company announced global age verification in February 2026, the backlash was heavy and the company delayed the rollout¹⁰. TechCrunch ran a roundup of alternatives at the time²².

There are three ways the moat could break. One is a platform owner bundling a good enough alternative: a console maker, a game store or an operating-system company can build voice and groups into the place players already are. Another is trust. Discord has had a breach at a third-party vendor that may have exposed government ID photos of about 70,000 users¹⁰, and it is now asking more users to prove their age¹¹. The third is the cost of safety rules, which are heavier on a platform with private rooms and young users than on a broadcast site.

What helps the moat is that the product keeps adding reasons to stay: end-to-end encrypted voice and video for all users, announced in May 2026²², a Social SDK for games, launched in spring 2025³, and rewards partnerships such as Xbox Game Pass⁹. None of these has been shown, in public data, to lift retention or revenue.

VThe numbers we can see

There is very little. A private company that has not published a prospectus gives the public only user counts, price lists and a few press-reported items. I looked for audited financials, a public S-1 and a company statement on revenue. I did not find them. EDGAR lists no S-1 or similar filing for Discord Inc. as of 3 November 2026¹².

The last priced funding round is from 2021: about $500 million at a reported valuation of $14.7 billion, per TechCrunch citing earlier reports⁶⁷. Bloomberg’s headline in September 2021 said $15 billion (I could not open the article)²⁷, and Sacra’s profile shows both $14.7 billion in its text and $15 billion in its summary box¹⁵. Two figures, a gap of about 2%. Neither says whether the figure is before or after the new money. Sacra puts total funding near $1 billion¹⁵. Separately, Microsoft was reported in 2021 to be in talks to buy Discord for about $10 billion, and Discord ended those talks, according to Wall Street Journal reporting summarised by Business Today²⁶. That is a press report, and it is a price that was offered, not a price that was paid.

Headcount is a clue, but I cannot pin it down. TechCrunch reported that Discord cut about 4% of staff in August 2023, nearly 40 people²⁴, which implies a workforce near 1,000 then (my arithmetic). Yahoo Finance’s profile lists 501 full-time employees with no date²⁵. The two figures do not sit well together and I cannot tell which is current. A company that trimmed staff and then reported positive adjusted EBITDA looks to have focused on cost¹⁵. That is my reading, not a company statement.

The comparison table below shows what public investors pay for comparable consumer internet companies. It is a data pull from 3 November 2026, using the prior day’s price^1617181920.

Public comparables, data as of 3 November 2026 (price at 2 November close)

VIWho runs it and who owns it

Humam Sakhnini has been chief executive since April 2025. He was previously Vice Chairman of Activision Blizzard³. Co-founder Jason Citron moved to the board and an advisor role. Co-founder Stanislav Vishnevskiy is chief technology officer³. In announcing the change, Citron said Sakhnini would lead Discord through its next chapter of growth and "someday becoming a public company"²³.

On ownership, I could not open a cap table. Sacra lists investors including Tencent, Sony, Greylock, Accel, Dragoneer, Fidelity and Franklin Templeton, with total funding of about $1 billion¹⁵. I cannot say who owns what, whether any investors hold preferences that rank ahead of ordinary shares, or how much stock the founders retain. Those terms matter at an IPO, because preferred shares usually convert, but protective terms and a ratchet can change what each class receives.

Leadership has also been visible in public. The House Oversight Committee invited Sakhnini, with the heads of Steam, Twitch and Reddit, to a hearing on online radicalisation on 8 October 2025²⁸. I did not confirm whether he appeared. On age verification, the chief technology officer wrote after the February 2026 backlash that the company should have given more detail about its intentions¹⁰. I read that as a communication failure and not necessarily a product failure, but it is a data point on how the company handles trust with its users.

Two further incidents are worth knowing about. In July 2026 TechCrunch reported that Discord admitted an AI moderation bug had wrongly banned users over harmless images²². And in October 2026 a third-party Discord app called Double Counter was hacked. Its own incident report, as relayed by Insider Gaming, said about 28 million Discord user IDs and usernames were partly copied and about 1 million email addresses were copied, and Discord said the incident was not a breach of its own systems³⁵. Have I Been Pwned lists about 275,000 email addresses from the data that was later published³⁴. I did not open the original TechCrunch report on the moderation bug. The point is not any single incident. A company about to ask public investors for money has to show that its safety and security systems work at the scale it claims, and these are the events a prospectus risk-factor section would be expected to discuss.

VIIValuation: is the last price supported?

Equity value in US$ billions. Not a share price: a private company has none. Latest mark: primary round (about $500 million raised), press-reported; Bloomberg reported $15 billion, TechCrunch $14.7 billion; pre- or post-money not stated.

What the last valuation assumes

The last priced round, about $14.7 billion in 2021⁶, was set in a different market. On my base revenue estimate of about $1.0 billion for 2026, it equals roughly 14.7 times sales. None of the five public comparables trades above 5.0x^1617181920. To justify $14.7 billion today, Discord would need revenue well above $2.5 billion at a 5x to 6x multiple, soon, with strong margins. Nothing public shows that.

Bear, base and bull

• Bear. My estimate of 2026 revenue is $0.85 billion, growing 8%, 8%, 7% and 6% to $1.12 billion in 2030. Terminal multiple 2.4x sales, near Pinterest¹⁸. Discount rate 14% for an illiquid, thinly documented private company. Dilution 8%. Equity value about $1.4 billion. Net cash is not included because the balance sheet is not public.
• Base. My estimate of 2026 revenue is $1.0 billion, from Sacra’s $725 million end-2024 run rate¹⁵, growing 20%, 18%, 16% and 14% to $1.87 billion in 2030. Terminal multiple 4.0x sales, between Reddit at 4.7x¹⁶, Duolingo at 4.5x²⁰ and Pinterest at 2.4x¹⁸, reflecting a smaller scale and thin proven advertising. Discount rate 12%, a cost of equity basis for a small consumer internet company. Dilution 5%. Equity value about $4.4 billion, before net cash.
• Bull. My estimate of 2026 revenue is $1.1 billion, growing 30%, 27%, 24% and 21% to $2.72 billion in 2030, as advertising and commerce become a large second line. Terminal multiple 5.5x sales, slightly above Roblox at 5.0x¹⁹ and Reddit at 4.7x¹⁶. Discount rate 10%. Dilution 4%. Equity value about $9.6 billion, before net cash.

Comparables used

Where I come down

The 2021 valuation looks well above my base case of $4.4 billion and above my bull case of $9.6 billion. It is a stale mark: more than five years old, set in a different rate and multiple environment, and no new round, tender or IPO price has been published. The market might still be right to pay more than my base. My revenue base is a third-party estimate, not a filing. The company may have grown faster than 20% in 2025, advertising may be scaling faster than I assume, and a profitable business with 90 million daily users could earn a scarcity premium at IPO. The cash on its balance sheet is also unknown and would add to equity value. The honest reading is that the price from 2021 is not supported by public data in 2026, while the true figure could sit well above my base if the filings show faster growth and strong margins.

What I cannot judge

I cannot see audited revenue, gross margin, cash, stock-based pay, the share count or the terms of the preferred stock. I cannot say why a reported March 2026 listing did not happen, or whether the company has delayed on purpose. I did not open court filings in the Texas and Brazil cases. The model rests on one third-party revenue estimate and a handful of consumer internet comparables, so treat the range of $1.4 billion to $9.6 billion as wide, and a public prospectus could move it a lot.

VIIIThe IPO

IPO filed confidentially. Press-reported, not confirmed by the company. TechCrunch, citing Bloomberg, reported on 7 January 2026 that Discord confidentially filed for a US IPO, with Goldman Sachs and JPMorgan as underwriters and a possible debut in March 2026⁶. Earlier, the New York Times reported exploratory IPO talks in March 2025, and Discord’s spokesperson said the company does not comment on rumours⁷. In April 2025 Discord’s co-founder said the incoming CEO would lead the company through growth and an eventual move to being a public company²³. As of 3 November 2026, EDGAR lists no S-1 or other registration statement for Discord Inc., only Form D notices from 2019 to 2023¹², and Discord is absent from the list of recent IPOs²¹. I found no exchange, ticker or offering size, and I found no reporting that explains the missing March listing. I could not open Bloomberg or Reuters directly.

What to watch

• A public S-1 on EDGAR, which would show audited revenue, margins, cash and share terms
• Reported revenue and adjusted EBITDA against the Sacra estimate of about $725 million annualised in 2024¹⁵
• Daily and monthly active user definitions and any change in engagement after age verification
• Revenue split between Nitro, advertising and commerce
• Outcomes in the Texas lawsuit and the Brazil order³⁰³¹
• Whether trading in recent tech listings, several of which sit well below offer price²¹, changes the appetite for a consumer internet listing

IXWhat could go wrong

Start with what most likely destroys value. For a social product, it is a slow loss of attention that shows up as falling engagement before it shows up in revenue. The company reports daily users, not time spent, and the best figure I could source is 90 million-plus daily users³. If that stalls, the 200 million monthly claim⁶ and the ad plans both suffer.

Regulation is the second risk, and it is specific. Texas sued Discord on 22 May 2026 in Collin County, citing the SCOPE Act and its deceptive trade practices law²⁹. By 3 August 2026 Discord had agreed to a temporary injunction, without admitting wrongdoing, that gives it 90 days to bring the age-assurance and default safety settings it uses in the UK to Texas users; the lawsuit itself was still pending as of early August 2026³⁰. In Brazil, the data protection authority ordered Discord on 12 August 2026 to suspend Go Live and similar video tools after the death of a 13-year-old girl that authorities linked to a Discord livestream, and Discord called the order premature³¹³². In the Philippines, the government’s cybercrime centre had Discord blocked from the evening of 23 September 2026 and said on 24 September that the block was being lifted after Discord agreed to strengthen child protection measures³³. I did not open the court documents, and the Texas injunction terms come from press coverage. Rules on age checks and on content for minors can raise costs and cut engagement at the same time.

Third, the monetisation plan has an awkward fit with the product. The new public-discovery features make selected messages searchable on the web starting mid-December 2026, with the setting off by default in the EU, UK, Canada, Brazil, Australia and South Korea⁵. Users have objected and some worry about AI training use¹². That could raise traffic and ad inventory. It could also upset the volunteers who run the communities.

Fourth, multiply the odds. For the 2021 valuation to be right in 2026, several things have to hold together: users keep using it daily, paid conversion holds or grows, advertising becomes a large second line, regulation costs stay contained, and the IPO market pays a software-like multiple for a business with consumer-internet churn. If each of these is a bit better than even, say 65% each, the joint chance of all five is about 12% (0.65 to the fifth power). That is my illustration, not a measured probability, and the factors are not truly independent.

Finally, there is the next best use of the same dollar. A buyer who wants consumer internet exposure can already own Reddit at 4.7x sales or Pinterest at 2.4x sales in public markets, with audited numbers¹⁶¹⁸. A private holder gets less information and less liquidity, so the price should reflect that. This is research and not investment advice.

What would change my view? In the favourable direction: a public S-1 showing 2025 revenue well above my $1.0 billion 2026 estimate, advertising growing quickly, stable daily users after the age checks, and a large cash balance. In the unfavourable direction: flat or falling engagement, a regulatory order that forces a costly change in Texas or elsewhere, a second data breach, or an IPO that is priced below the 2021 round. A down-round IPO would not mean the business is failing, but it would be the market’s own number replacing mine.

A last point on what an owner can take out. A shareholder in a private company cannot take anything out until the company pays a dividend, is bought, or lists. Discord has no public record of paying dividends and I would not expect a young growth company to. An investor’s return would therefore come from a sale or listing at a higher price, which depends on the market’s mood on the day as much as on the business. That is another reason I widen the range instead of trusting one number.