NEAR projects 1B AI agents

NEAR Protocol and Concordium are making a similar argument from different ends of the stack: if AI agents move from demos to payments, trading and software execution, the bottlenecks will be throughput, privacy and identity. IDC said in a recent market forecast that actively deployed AI agents will top 1 billion by 2029 and execute more than 217 billion actions a day, a volume that would require new infrastructure for execution, verification and cost control. NEAR has used that projection to frame its own pitch around transaction rails. On its public site, NEAR says it is “the open infrastructure powering the agent economy,” unifying liquidity across more than 35 chains, keeping execution and inference confidential, and scaling to more than 1 million transactions per second. A February company announcement described the same stack as a mix of NEAR Intents for cross-chain execution, confidential transaction flows and hardware-secured AI agents. (idc.com) ### Why is the IDC forecast getting attention now? IDC’s forecast is notable because it attaches concrete operating numbers to the agent thesis. The firm said agents will execute more than 217 billion actions per day by 2029, consume 3.7 trillion tokens or calls daily, and drive more than $68 billion in annual token delivery costs worldwide. IDC also said the cost to complete an individual action should fall sharply as usage scales. (near.org) Those figures give blockchain and infrastructure companies a way to argue that model quality alone will not decide the market. If daily agent activity rises to IDC’s projected levels, the winning platforms would need to handle very high action volume, low-cost settlement and some way to manage identity and permissions across systems. That is an inference from IDC’s usage figures and the product claims now being made by NEAR and Concordium. (idc.com) ### What exactly is NEAR saying it can provide? NEAR’s public materials emphasize three pieces: scale, privacy and cross-chain execution. The company says its sharded architecture can scale above 1 million TPS, and a separate NEAR blog post described a publicly verifiable benchmark at that level using 70 shards. A February release from NEAR said its product stack combines NEAR Intents for one-click cross-chain execution, confidential transaction flows and a secure runtime for AI agents. (idc.com) The company said the system was designed to let users and institutions opt into confidentiality while preserving onchain verification. (near.org) ### Where does Concordium fit into this? Concordium is approaching the same market from identity rather than throughput. Its website says the protocol is built so verified humans and verified AI agents operate on the same identity layer, and recent company materials describe that as trust infrastructure for “agentic commerce.” The “Verified by Concordium” badge, as described in company materials and related coverage, is meant to let an agent carry verifiable backing from a human or business even when that agent operates outside Concordium’s own chain. (prnewswire.com) Concordium has also described an Agent Registry and portable identity tooling aimed at making agent actions traceable to a verified party without exposing personal data publicly. (concordium.com) ### Why are privacy and portable identity being grouped together? Concordium Chief Executive Boris Bohrer-Bilowitzki told Biometric Update last month, “Without identity, autonomous action is just autonomous risk.” That framing matches NEAR’s argument that confidential execution matters if agents are going to handle trading, payments or other sensitive actions across chains. (medium.com) The combined pitch is that agents need both discretion and accountability: privacy for the data and instructions they handle, and portable identity so a counterparty can verify who or what stands behind the agent. Both companies are presenting those features as infrastructure, not add-ons. ### What should readers watch next? IDC’s 2029 forecast gives the timeline, but the nearer test is product adoption. (biometricupdate.com) NEAR’s cross-chain and confidential execution products are already listed on its public site, and Concordium is continuing to roll out identity and registry tools for agent use cases. The next concrete signals are likely to come from transaction volumes, third-party integrations and whether developers adopt portable verification for agents operating across multiple chains. (concordium.com) (near.org)