Switchboard Technology Labs Inc. announced on September 19, 2026, that all oracle implementations are deprecated effective immediately, with support ending September 25. The shutdown leaves Solana's largest DeFi protocols — Kamino ($1.3–1.5B TVL), Drift, MarginFi, and Jito — with six days to migr...
"AI has lowered the cost of building one's own oracle." — Switchboard Foundation Group, Shutdown Statement (September 19, 2026)
Switchboard Technology Labs Inc. announced on September 19, 2026, that all oracle implementations are deprecated effective immediately, with support ending September 25. The shutdown leaves Solana's largest DeFi protocols — Kamino ($1.3–1.5B TVL), Drift, MarginFi, and Jito — with six days to migrate price feeds to Pyth Network or RedStone. Switchboard processed hundreds of millions of weekly data requests across 550+ feeds on 11 chains.
The company raised $11.2 million in total funding, including a $7.5 million Series A in May 2024 led by Tribe Capital and RockawayX. That capital was not enough to sustain operations through a prolonged bear market in which AI tools commoditized oracle construction and large protocols began sourcing data directly from providers like S&P. Switchboard's closure follows an August 29 key compromise on four Move-based chains that caused at least $546,000 in confirmed DeFi losses and forced one protocol to permanently shut down.
The event exposes a structural risk in DeFi infrastructure: Switchboard secured an estimated 100% of Solana's lending TVL as an oracle provider, and its departure concentrates even more market share in the hands of Chainlink (~$33B in total value secured, approximately 59% of the tracked oracle market) and Pyth ($3.1B TVS). More than 260 crypto projects have shut down in 2026, according to RootData, but few carry the systemic weight of an oracle network embedded across billions in lending exposure.
Switchboard Technology Labs Inc. issued a single announcement on September 19, 2026, stating that "all implementations are deprecated effective immediately." The company set September 25 as the hard deadline for support termination, giving dependent protocols a six-day migration window.
The shutdown is a full corporate wind-down, not a temporary suspension. Switchboard's SWTCH token fell 18.7% on the announcement. The company named Pyth Network and RedStone as its recommended migration targets.
Key parameters of the shutdown:
| Item | Detail | |------|--------| | Announcement date | September 19, 2026 | | Effective deprecation | Immediate | | Support end | September 25, 2026 | | Migration window | 6 days | | Lifetime funding | $11.2 million | | Weekly data requests | Hundreds of millions | | Active feeds | 550+ across 53 Solana programs and 10 other chains | | SWTCH token reaction | -18.7% |
The Switchboard Foundation Group cited three forces that eroded the company's business model:
1. AI commoditized oracle construction. The cost of building a proprietary oracle dropped as AI tools became widely available. Protocols that previously outsourced data ingestion to third-party oracle networks found they could build and maintain feeds in-house for a fraction of the prior cost. This eliminated a core value proposition: that oracle networks provide specialized infrastructure too complex or expensive for individual protocols to replicate.
2. The bear market compressed budgets. New chain launches slowed. Existing protocols cut infrastructure spending. Switchboard's revenue — derived from chain partnerships, grants, and data request fees — contracted as the total addressable market shrank. The $7.5 million Series A raised in May 2024 did not provide a long enough runway.
3. Direct data partnerships bypassed oracle middleware. The Switchboard Foundation Group specifically cited Hyperliquid's direct collaboration with S&P as an example of the trend. Rather than routing market data through an intermediary oracle network, large protocols began signing direct agreements with data providers. This disintermediates the oracle layer entirely for high-value data feeds.
The combination is significant. Switchboard was not a failing product — it processed hundreds of millions of data requests weekly. It was a viable product whose pricing power evaporated. The oracle business model depends on being a necessary intermediary; when the necessity fades, margins collapse.
Switchboard's footprint on Solana was extensive. According to Solana Compass, the oracle secured nearly 100% of Solana's lending TVL, powering the following protocols:
| Protocol | Role | Estimated TVL/Activity | |----------|------|----------------------| | Kamino Finance | Lending (80% of Solana lending TVL) | $1.3–1.5B TVL | | MarginFi | Lending | Multi-hundred million | | Drift Protocol | Perpetuals / Lending | Significant open interest | | Jito | Liquid staking / MEV | Major Solana infrastructure | | Save (formerly Solend) | Lending | Active lending protocol | | Rain.fi | Lending against NFTs | Niche lending |
Oracle migration is not a trivial operation. It requires:
Six days is a compressed timeline for protocols carrying billions in open positions. Any oracle switch-over introduces a window of elevated risk: if the new feed reports a price that differs meaningfully from the old one, it can trigger unintended liquidations or create arbitrage opportunities.
Kamino Finance had already integrated Chainlink Data Streams and built a Multi-Price Oracle System that aggregates pricing from multiple oracle sources and cross-verifies in real time. This positions Kamino better than protocols that relied solely on Switchboard. Drift Protocol's documentation also references Pyth as a supported oracle, suggesting partial preparedness.
Solana's total DeFi TVL stood at approximately $5.92 billion as of September 6, 2026, up 25.46% over the prior 30 days. The migration event is occurring during a period of rising activity, which increases the stakes of any feed disruption.
Switchboard's exit further concentrates an already top-heavy market. As of June 2026, the oracle landscape by total value secured (TVS):
| Oracle Provider | TVS | Market Share | |----------------|-----|-------------| | Chainlink | ~$33B | ~59% | | Chronicle | ~$7.5B | ~13% | | RedStone | ~$3.6B | ~6.4% | | Pyth | ~$3.1B | ~5.5% | | Others | ~$9B | ~16% |
Chainlink is integrated with more than 2,400 projects. Some estimates place its broader value secured above $100 billion when including indirect integrations and cross-chain deployments. Pyth Network has expanded to 100+ chains and added equity, commodity, and futures data feeds, including CME index futures coverage.
RedStone's pull-model architecture — where protocols request data on-demand rather than receiving continuous push-based updates — has gained traction for cost efficiency. The protocol has been the fastest-growing oracle by integration count in 2025–2026.
Switchboard's permissionless architecture — which allowed anyone to create custom feeds without gatekeeping — was a differentiator that Pyth and Chainlink do not fully replicate. Protocols that relied on niche or custom Switchboard feeds may find no direct equivalent available from the recommended replacements.
The structural implication is clear: oracle infrastructure is consolidating toward fewer, larger providers. This reduces single-provider risk (the scenario that played out in August when Switchboard's key was compromised) but increases systemic concentration risk. If Chainlink or Pyth experienced a comparable failure, the blast radius would be orders of magnitude larger.
Switchboard's shutdown did not occur in isolation. On August 29, 2026, the company suspended oracle services on four Move-based chains — Aptos, Sui, IOTA, and Movement — after detecting a key compromise. The attacker gained control of 14 oracle signing keys on the IOTA mainnet queue and wrote arbitrary price data directly into feeds.
Confirmed damage from the August incident:
| Protocol | Chain | Loss | Outcome | |----------|-------|------|---------| | Virtue Money | IOTA | ~$455,000 | All functions frozen; VUSD undercollateralized | | Full Sail | Sui | ~$91,000 | Permanent shutdown | | Volo | Sui | $0 (precautionary pause) | Resumed operations |
The attacker set IOTA's price to $10 million, minted approximately 4.94 million VUSD stablecoins, then crashed the feed to trigger 47 liquidations across 45 users. Full Sail requested financial support from Mysten Labs (Sui's core developer); Mysten declined. Full Sail ceased operations permanently.
Total confirmed losses: $546,000. Switchboard's Solana deployment, built on a different codebase, was unaffected.
The August breach and the September shutdown are officially separate decisions, according to Switchboard. But the sequence is relevant: the security incident likely accelerated contributor departures and eroded protocol confidence, narrowing the company's options.
Switchboard's shutdown arrived one day after Linera, an a16z-backed Layer 1 blockchain, announced its own closure on September 18. Linera had raised $12 million across two seed rounds (a16z crypto in 2022, Borderless Capital in 2023) but failed to complete a community token sale — its LNRA round attracted $848,271 from 617 participants, falling short of the $1.5 million minimum threshold. All contributions were refunded.
RootData's 2026 dead projects list has surpassed 260 entries. The figure places 2026 among the most challenging years for crypto project viability since the 2018–2019 downturn. The pattern is consistent: venture-backed projects that raised during 2022–2024's funding cycle are running out of runway in a market where token launches generate insufficient revenue and grant funding has dried up.
The distinguishing factor with Switchboard is systemic importance. Most of the 260+ shuttered projects were applications, tokens, or experimental chains with limited downstream dependencies. An oracle network embedded across billions in lending TVL carries fundamentally different risk. Its failure mode is not simply lost investment — it is potential cascading liquidations across every protocol that depends on its feeds.
Switchboard's shutdown is a case study in infrastructure economics. The product worked — hundreds of millions of weekly data requests across 550+ feeds attest to that. But the business could not survive the convergence of three forces: AI lowering build costs, a bear market compressing revenue, and large protocols cutting out the middleman.
The six-day migration window is the immediate risk. Protocols like Kamino, which had already diversified to multi-oracle architectures, are better positioned than those running single-feed dependencies on Switchboard. The market will learn within a week whether the migration proceeds smoothly or produces feed disruptions.
The longer-term question is structural. Oracle infrastructure is consolidating. Chainlink's dominant position grows larger with each competitor that exits. That concentration is efficient — it reduces integration complexity and standardizes data feeds — but it is also a single point of systemic failure if the dominant provider ever faces its own operational crisis. DeFi's dependency on a small number of oracle networks is the infrastructure equivalent of too-big-to-fail, without any of the backstops that label implies.