Mentioned in This Article
1. What is Space and Time Enterprise?
Space and Time, a blockchain data project backed by Microsoft, announced Space and Time Enterprise on Monday, November 9. The product lets businesses prove that calculations run on their private data were performed correctly, without exposing the underlying records to the people, or the AI agents, that rely on the results. Companies can run the software on their own servers and do not need to operate a blockchain to use it.
The launch targets a problem that has grown as organizations hand more analytical work to automated systems: how to let those systems act on figures that can be trusted, without granting them open access to sensitive databases.
Verifiable computation is attractive in regulated industries because auditors and counterparties often need assurance about a figure without being allowed to view the confidential records behind it. A proof gives them a way to check the arithmetic while the data stays where it is.
The announcement frames the product as infrastructure for companies rather than for crypto traders. Its pitch is that a business can gain the assurance usually associated with blockchains, which is the ability to check that a record or calculation has not been tampered with, while keeping its data inside its own environment.
2. Why AI agents and private data are a concern
Nate Holiday, chief executive of MakeInfinite Labs, the original developer of Space and Time, argues that AI agents do not need unrestricted access to the records behind every calculation. In his view, giving AI tools broad access to company data creates an additional avenue of exposure, because every system that can read raw records is a system that could leak or misuse them.
AI agents are software programs that carry out tasks, such as producing reports or answering questions, with some degree of autonomy. When they are connected to a company's databases, they can query large volumes of records. The more freely they can do so, the greater the amount of information that could be exposed through errors, misconfiguration or an attack on the agent itself.
The alternative the platform proposes is to separate the answer from the data. An agent receives a result together with proof that the result was computed correctly from the underlying records, rather than receiving the records themselves.
Security teams typically respond to this kind of risk by limiting access, logging queries and separating duties. Cryptographic proofs add a different kind of control, since they let a recipient confirm a result without being trusted with the inputs, which narrows how much any single system needs to see.
Regulators and corporate security teams have both paid growing attention to how automated tools handle confidential data. An agent that can read everything in a database is convenient, but it widens the group of systems whose compromise would matter. Reducing what each tool can see is a long-standing principle in security, often described as giving a system only the access it needs.
3. How Proof of SQL works
The system relies on a technique called Proof of SQL, which Space and Time describes as patented. It is a cryptographic method for proving that a database query was computed correctly. SQL, short for Structured Query Language, is the standard language businesses use to ask questions of their databases, so a proof tied to SQL queries fits directly into existing workflows.
The software works with Google Cloud BigQuery, a widely used cloud data warehouse. Holiday says that GPU acceleration can prove supported queries across millions of rows in under a second, while noting that performance varies depending on the query, the hardware and the dataset.
In practical terms, a business could run a calculation, such as a total across a large ledger, and attach a proof that the figure was derived correctly. A reporting tool, an AI agent or a compliance team can then check the proof instead of examining the raw records.
A cryptographic proof of this kind is compact and can be checked much faster than redoing the full computation, which is what makes it useful as a trust mechanism. The recipient verifies a short proof and does not have to rerun the query or access the data.
Compatibility with an established warehouse such as BigQuery matters for adoption. Companies are generally reluctant to move their data into a new system, so a verification layer that sits on top of tools they already use faces fewer obstacles than one that asks them to migrate.
4. Who the product is aimed at
Space and Time says its target customers are firms in construction, engineering and manufacturing that are rebuilding their systems around AI. In those organizations, agents, reporting tools and compliance teams would be able to share the same verified records, which reduces disagreements over which figures are authoritative.
The company named several launch partners: Maple Finance, an onchain asset manager; Indomobil, an Indonesian vehicle maker and distributor; Trenez, a consultancy that works with Oracle software; and The Data Foundation, an AI data network. The list mixes crypto-native and traditional businesses, which matches the way the product is meant to operate outside of a blockchain.
Industries such as construction, engineering and manufacturing generate large volumes of operational records, from project costs to parts inventories, and many are modernizing their reporting. A shared verified source helps when finance, operations and compliance teams each rely on automated tools that may otherwise draw on different copies of the same numbers.
The partners also reflect different uses. An onchain asset manager has an interest in provable calculations for reporting to investors, a vehicle distributor holds large operational datasets, and a software consultancy can help clients integrate the product into existing systems. The Data Foundation, as an AI data network, connects the offering to the way models and agents obtain information.
5. What proofs do not guarantee
The announcement acknowledges a limit. A proof confirms that a calculation was performed correctly on the data provided, but it does not establish that the input data was accurate in the first place. If the records in a database are wrong, a verified calculation over them will faithfully produce a wrong answer.
For that reason, businesses still need their own data controls, including processes for entering, validating and auditing the information that feeds their systems. The platform addresses the integrity of computation, which is one layer of trust, and not the quality of the data going into it.
This distinction is common in verifiable computing and is worth keeping in mind when comparing the claims of different vendors. Proving that a query ran as specified is different from proving that the data it ran on is complete or truthful.
Buyers evaluating the platform will therefore need to ask two separate questions. The first is whether a given calculation can be proven, and the second is whether the data feeding it is controlled well enough to trust. A mature deployment would combine the proof layer with access controls, audit trails and validation at the point where data is entered.
6. The company's background and the SXT Chain
Space and Time raised $20 million in 2022 in a funding round led by M12, Microsoft's venture fund. Last year, Grayscale launched a trust for SXT, the token behind the project's blockchain. Earlier this year, the company introduced Dreamspace, an AI-powered tool for building onchain applications without writing code.
The SXT Chain is a decentralized network that supplies verifiable data to decentralized finance, tokenized assets and stablecoins. Holiday says that enterprise workloads would be paid for through Space and Time, and that the network becomes relevant when applications need to connect to onchain markets. A company could therefore use the product entirely on its own servers, and tie into the blockchain only if it wants to interact with those markets.
Holiday also pointed to the technology's reach in the wider cryptography field. Microsoft's Nova project includes Blitzar, Space and Time's acceleration library, and the team's Dory commitment scheme contributed to the Jolt proof system from a16z crypto. Those examples are offered as evidence that the underlying research has been adopted beyond the company's own products.
Taken together, the company's funding history, its token trust and its application tools show a project that has built several products around the same data verification core, with the enterprise offering extending that core to customers who do not want a blockchain at all.
The mention of Grayscale is a useful marker of how the token is viewed by institutional investors. A trust is a vehicle that lets investors gain exposure to an asset through a traditional product, and its launch for SXT last year indicates that the asset attracted interest from a large digital asset manager.
7. What to watch next
The announcement did not include customer deployment dates or pricing, so the first test is whether the named partners move from announcing involvement to running production workloads. Evidence of actual use, such as published case studies or deployments inside construction and manufacturing firms, would show whether the pitch resonates outside the crypto industry.
Performance claims are also worth tracking. Sub-second proofs over millions of rows are described as dependent on query type, hardware and dataset, so independent measurements on real business workloads will matter. Finally, the interplay between enterprise usage and the SXT Chain remains to be seen, since the company says the network matters when applications connect to onchain markets rather than for every deployment.
Waiting for those milestones is the sensible approach, because the announcement describes capabilities and partners rather than results. The company's own caveats about proofs and data quality suggest that adoption will depend on how well the product fits with the controls businesses already operate.
A dual structure, with an enterprise product that runs independently of any blockchain and a decentralized network for onchain use cases, gives the company two routes to customers. The question for the coming quarters is how revenue and activity divide between them.