Skip to content
Deepak Gupta markTech Graveyard

Tech Graveyard/consumer

Skype (2003-2025)

When Skype launched in 2003, peer-to-peer was the future of consumer internet. By the time it died in 2025, every consumer app routed through centralized clouds.

Born 2003 · Died 2025 · Status: dead

Certificate of Death

Name of decedent

Skype

Born
2003
Died
2025
Age
22

Cause of death

Microsoft consolidated to Teams; consumer messaging fragmented to WhatsApp, FaceTime, Zoom

Survived by

Microsoft Teams, Zoom, WhatsApp, FaceTime, Google Meet, Discord

Invented by

Niklas Zennstrom and Janus Friis (Estonia engineering team)

Status: DeadFinal breath: 2025

Filed by D. Gupta · guptadeepak.com

The hook

Microsoft shut down Skype on May 5, 2025. The product had 300M monthly users at its peak. Its death is the official end of an architectural era for consumer internet.

Thesis. Skype died twice. First in 2017 when Microsoft rebuilt it on Azure infrastructure and killed the original peer-to-peer architecture. Then in 2025 when the brand ended. The deeper death: consumer P2P apps stopped winning.

The story

The origin

August 2003. Niklas Zennstrom and Janus Friis (the KaZaA team) launched Skype. Peer-to-peer voice calling. Free worldwide. The model was genuinely new: no central servers carrying voice traffic, just an overlay network of supernodes routing calls between users.

The peak

2013. 300M monthly users. eBay had acquired Skype for $2.6B in 2005, sold it to a private equity group in 2009, and Microsoft bought it again for $8.5B in 2011. The supernode topology was still doing real work on user machines around the world.

The architecture flip

2017. Microsoft replaced the P2P backbone with Azure-hosted cloud servers. The original technical magic was gone. Calls now ran through Microsoft data centers. The pretext was mobile reliability. The result was a normal SaaS calling app with a heritage logo.

The platform consolidation

2020 onwards. Microsoft pushed Teams hard for enterprise after the pandemic-era boom. Consumer messaging splintered to platform-specific apps: FaceTime, WhatsApp, iMessage, Google Meet, Discord. Skype had no home left to defend.

The death certificate

February 2025 Microsoft announced retirement. May 5, 2025 Skype shut down. Users were migrated to Teams or told to export their data. A 22-year run ended quietly.

Key data points

  • Skype launch: August 2003
  • eBay acquisition: October 2005, $2.6B
  • Microsoft acquisition: May 2011, $8.5B
  • Peak monthly users: 300M+ in 2013
  • Architecture migration from P2P to cloud: 2017
  • Final shutdown date: May 5, 2025

Contrarian angle

Every open or decentralized consumer protocol that competed with Skype in 2005 (Jabber, SIP, XMPP) is also functionally dead. The lesson is not that Skype lost. The lesson is that consumer P2P lost. Every messaging app you use is now a centralized service with a king, often running on the same three hyperscaler clouds.

The flip side

What replaces it

The paired prediction in Future Tech.

Read the prediction

FAQ

Why did peer-to-peer architecture matter for Skype?

It let Skype scale globally without buying massive server capacity. User machines did the routing. It also produced strong end-to-end paths that were hard to wiretap, which is part of why Skype's audio quality was unusually good for the era.

What replaced Skype for consumers vs enterprise?

Consumers moved to WhatsApp, FaceTime, iMessage, Google Meet, and Discord. Enterprise moved to Microsoft Teams, Zoom, and Slack Huddles. The category fragmented into specialized apps.

Are any consumer apps still peer-to-peer today?

BitTorrent for file transfer, some mesh messengers like Briar, and a few niche audio apps. The mainstream is centralized. The Matrix and ActivityPub protocols exist but have not reached Skype-scale adoption.

Why did Microsoft kill Skype instead of upgrading it?

Teams is more valuable to Microsoft per user because it bundles with Microsoft 365 subscriptions. Skype's consumer business was hard to monetize and ran on infrastructure that did not align with Microsoft's enterprise roadmap.

What does this mean for the future of decentralized protocols?

Decentralization at consumer scale is harder than the 2005 wave believed. Network effects, identity, abuse moderation, and onboarding all favor centralized providers. Bluesky and the fediverse are trying again with better answers, but the deck is still stacked.

More from guptadeepak.com

Want the technical deep-dive on what replaces this?

Read the companion article

Related obituaries

More from the consumer graveyard.