TL;DR — Key Takeaways
- Anaconda’s acquisition of Kilo Code strikes the corporate past Python tooling and into the AI coding-agent layer. Kilo provides Anaconda direct entry to the developer interface the place fashions are chosen, brokers are directed and organizational information is routed.
- The deal types a part of a broader platform technique. Anaconda’s bundle and setting administration, Outerbounds’ workflow orchestration and Kilo’s agentic improvement instruments may turn into an end-to-end enterprise AI improvement platform.
- The chance is critical, however integration and developer belief will likely be vital. Anaconda should present enterprise governance, safety and visibility with out undermining Kilo’s open-source flexibility, mannequin neutrality or developer expertise.
The Kilo Code acquisition enhances Anaconda’s Python roots, but it surely additionally reveals a bigger ambition: Transferring up the AI stack earlier than the worth—and the developer relationship—strikes past it.
Anaconda introduced this week that it’s buying Kilo Code, and the announcement arrived wrapped in sufficient AI advertising and marketing language to fill a number of context home windows.
There’s a “tokenpocalypse.” Enterprises are “token-maxxing.” CIOs are being requested whether or not they know the place their information is. Anaconda and Kilo, in the meantime, are promising “AI by yourself phrases.”
Let’s strip all of that away. The deal is essential sufficient with out the promotional wrapping.
Kilo Code is an open-source, model-agnostic coding agent utilized by greater than 3 million builders. In accordance with the businesses, it orchestrates practically 10 trillion tokens per 30 days and might route work throughout greater than 500 fashions. It operates inside VS Code, JetBrains and the command line, putting it straight the place builders and AI brokers more and more carry out their work.
That makes Kilo a major addition to Anaconda. However the extra fascinating query is what the acquisition says about Anaconda itself.
Most of us nonetheless consider Anaconda as the corporate behind the Anaconda Python distribution. That identification has served it nicely. Python grew to become the lingua franca of information science, machine studying and far of synthetic intelligence, whereas Anaconda made the language and its sprawling bundle ecosystem manageable for tens of millions of builders and 1000’s of enterprises.
Now, Anaconda seems decided to not stay merely the Python firm.
Extra Than a Python Distribution
The simple rationalization is that Kilo enhances Anaconda’s present enterprise. That’s true, but it surely doesn’t totally seize what is occurring.
Anaconda’s worth was by no means restricted to packaging Python and making it simpler to put in. It supplied environments, dependency administration, curated packages, reproducibility and a stage of safety and governance round an open supply ecosystem that might in any other case turn into tough for enterprises to handle.
Anaconda didn’t create Python, nor did it personal all of the software program its clients used. It made that software program consumable and reliable. It grew to become an essential layer between a sprawling open supply ecosystem and organizations that wanted some assurance about what their builders and information scientists had been truly working.
That very same downside is reappearing at a a lot bigger scale with AI.
Builders are not selecting solely packages and libraries. They’re choosing fashions, connecting brokers to inside programs, passing organizational information by exterior providers and permitting software program to take actions on their behalf. One developer could use a frontier mannequin for a fancy reasoning activity, an open-weight mannequin for routine coding and a regionally hosted mannequin for work involving delicate information. Brokers could set up packages, name MCP servers, entry databases and generate code that finally reaches manufacturing.
The Python setting stays essential, however it’s not your complete setting that must be managed.
Kilo provides Anaconda a place on this expanded improvement floor. It strikes the corporate from managing what sits beneath the developer to collaborating within the interface straight in entrance of the developer.
That’s complementary, however it is usually transformational.
The Threat of Remaining Indispensable
Anaconda may have chosen to stay centered on its conventional place. Python will not be disappearing. Neither is the necessity for safe packages, reproducible environments and dependency administration. If something, AI-generated software program makes these capabilities extra essential.
However significance doesn’t essentially translate into management of the shopper relationship or seize of the financial worth.
If builders more and more start their work inside an AI coding agent, the agent supplier can turn into the first interface. That supplier could resolve which fashions are introduced, which instruments are related, how context is assembled and the place workloads run. The bundle and setting layer beneath it might stay important whereas turning into much less seen and extra interchangeable.
This can be a model of what I name The Indispensability Lure: Foundational expertise can stay crucial whilst worth strikes farther up the stack. The businesses offering the muse do the laborious work of holding the whole lot working, whereas somebody working nearer to the consumer captures the connection, differentiation and margin.
Anaconda seems to know the hazard.
Python gave it a unprecedented place originally of the info science and AI improvement course of. Kilo is an try to comply with that course of upward as the start line strikes from opening a Python setting to instructing an agent.
The Items of a Bigger Platform
Kilo makes much more sense when seen alongside Anaconda’s acquisition of Outerbounds in April.
Outerbounds is the corporate behind Metaflow, the open-source framework initially developed at Netflix for constructing and working information science and machine studying workflows. The acquisition gave Anaconda extra capabilities round compute, workflow orchestration and shifting AI functions from improvement into manufacturing.
The items now kind a recognizable structure.
Conventional Anaconda provides trusted packages, environments, fashions, dependency administration and software program provide chain governance. Outerbounds and Metaflow prolong that basis into workflow orchestration, compute and manufacturing. Kilo provides the agentic engineering layer the place builders select fashions, direct coding brokers, join instruments and produce software program.
Anaconda is assembling the components for an end-to-end AI improvement platform.
That doesn’t imply the corporate has already created one. Buying a group of complementary applied sciences is simpler than integrating them right into a coherent expertise. Packages, environments, mannequin catalogs, manufacturing workflows, coding brokers, mannequin routing, coverage enforcement and utilization analytics could line up properly on a presentation slide. Making them function as one platform with out including friction for builders is significantly more durable.
Nonetheless, the route is turning into tough to overlook.
David DeSanto and I mentioned Anaconda’s evolution throughout an interview at AWS re:Invent in 2025. The dialog was already about Anaconda’s AI platform, open source growth and the future of secure software. It was not merely a dialogue about bettering Python distribution.
Extra just lately, we spoke about OpenAI’s acquisition of Astral, the corporate behind the more and more widespread Python instruments uv, Ruff and Pyright. We mentioned what that acquisition may imply for Python, open supply and belief in AI improvement.
Seen collectively, the Astral and Kilo acquisitions create an fascinating strategic mirror.
OpenAI, a frontier mannequin firm, moved down the stack into Python tooling. Anaconda, a trusted Python and bundle firm, is shifting up the stack into coding brokers, mannequin routing and the developer interface.
The mannequin suppliers are shifting towards Anaconda’s conventional territory. Anaconda is shifting towards theirs with out making an attempt to construct a frontier mannequin of its personal.
The GitLab Connection
There may be additionally a human connection behind this deal that shouldn’t be missed.
Kilo co-founder Sid Sijbrandij is greatest referred to as the co-founder and government chair of GitLab. DeSanto spent six years at GitLab and have become its longest-serving chief product officer, main the product group as GitLab expanded into an built-in, AI-enabled DevSecOps platform.
That historical past issues.
DeSanto and Sijbrandij didn’t want to start this dialog by explaining their respective views of builders, open supply or built-in platforms. They’d already labored collectively whereas GitLab expanded from an organization identified primarily for source-code administration into a wider platform spanning the software program improvement lifecycle. DeSanto performed an essential function in that transformation, together with the addition of safety and compliance capabilities.
We have no idea who positioned the primary name or precisely how the Kilo transaction got here collectively. Except one of many individuals tells us, it could be irresponsible to assert that the prior relationship produced the deal.
It will be equally naïve to faux the connection was irrelevant.
Acquisitions contain greater than evaluating product options. The customer is betting on the individuals, tradition, neighborhood and expertise it’s absorbing. The vendor is deciding whether or not the customer might be trusted with what it constructed. A protracted working relationship between Anaconda’s CEO and one in all Kilo’s founders would give each side information that doesn’t seem in a regular due-diligence report.
Sijbrandij’s outstanding function within the acquisition announcement reinforces the connection. He described Kilo and Anaconda as a uncommon match with “nearly no overlap,” saying that what one lacked, the opposite already possessed. That’s greater than the customary congratulatory quote from a monetary backer. It’s the evaluation of a Kilo co-founder who already is aware of DeSanto and has beforehand labored with him whereas constructing a developer platform.
The connection additionally gives one other clue about DeSanto’s ambitions for Anaconda.
At GitLab, DeSanto participated within the firm’s evolution from a product related to a selected a part of the event course of right into a platform supposed to handle way more of the software program lifecycle. At Anaconda, he seems to be making use of a associated playbook: Begin with a robust open supply basis and a big developer neighborhood, then add adjoining capabilities till the unique product turns into the middle of a broader platform.
The comparability shouldn’t be pushed too far. Anaconda will not be GitLab, and AI improvement will not be merely the following model of DevSecOps. However DeSanto has seen one of these platform enlargement earlier than. Extra importantly, he has labored with Sijbrandij whereas doing it.
Kilo could due to this fact signify greater than a complementary acquisition. It could even be a reunion of people that share a view of how developer platforms are constructed.
Custody of the Developer—and the Information
The strategic prize will not be merely possession of one other AI coding assistant.
Kilo helps a whole bunch of fashions moderately than requiring builders to commit to at least one supplier. That probably provides Anaconda a spot between enterprises and the mannequin firms competing for his or her workloads. Anaconda can supply entry to frontier fashions for complicated work, open-weight fashions for lower-cost duties, native fashions for delicate use instances and air-gapped deployments the place information can’t go away the group.
The worth lies in governing these selections.
Which fashions are authorized? What information is being despatched to them? Which brokers can entry a source-code repository, database or manufacturing system? How a lot is the corporate spending throughout totally different instruments and accounts? Which packages did an agent introduce? Can a company reconstruct how a bit of AI-generated software program was produced?
These usually are not theoretical questions. They’re turning into a part of on a regular basis software program improvement, usually earlier than enterprise governance programs are able to reply them.
“The competition in enterprise AI improvement is shortly shifting to the interface the place builders direct brokers, select fashions, and route organizational information. Anaconda’s Kilo acquisition positions it with a belief layer to control that floor. Consumers are locking in agentic structure earlier than governance can reply which fashions brokers use, what information they ship, and which programs they contact. Coding agent choice is now a control-plane determination that can not be deferred till adoption hardens”, Mitch Ashley, VP and apply lead, CIO tech purchaser and software program lifecycle engineering at The Futurum Group.
Mannequin suppliers wish to personal that relationship. So do cloud suppliers, IDE distributors and agent platforms. Every needs to turn into the place the place enterprises select fashions, route workloads and apply coverage.
Anaconda enters that contest with two benefits: belief and distribution. It already claims greater than 52 million customers and a presence inside 95% of the Fortune 500. Kilo provides it a solution to prolong these relationships into the agentic improvement expertise.
The corporate is attempting to maneuver from being the setting builders depend upon to turning into the management aircraft enterprises use to handle AI improvement.
Developer Freedom Meets Enterprise Management
There may be an apparent pressure on this technique.
Kilo grew as a result of builders discovered it helpful, open and versatile. It’s model-agnostic and obtainable throughout the instruments builders already use. Its open supply basis provides customers a level of visibility and management that proprietary coding assistants don’t at all times present.
Anaconda is shopping for Kilo partly due to these qualities. It may additionally harm them.
If Kilo turns into a funnel right into a heavy enterprise platform, builders could look elsewhere. If governance interprets into restrictions, delays and company approval gates, the product will lose among the freedom that drove its adoption. If mannequin neutrality quietly turns into a most well-liked set of business relationships, the anti-lock-in place will likely be more durable to defend.
Anaconda’s alternative is to make enterprise management function across the developer moderately than in opposition to the developer. Insurance policies ought to decide which fashions and assets can be found with out requiring a developer to submit a ticket every time an agent must carry out helpful work. Safety ought to be embedded within the setting. Governance ought to present visibility with out turning AI improvement right into a checkpoint marathon.
That’s the promise. The acquisition doesn’t show that Anaconda can ship it.
Kilo Could Not Be the Final Deal
DeSanto and Anaconda have proven they don’t seem to be shy about utilizing acquisitions to speed up the corporate’s evolution.
Outerbounds and Kilo arrived inside a number of months of one another. That appears much less like opportunistic dealmaking than a deliberate effort to assemble a platform whereas the AI improvement market remains to be taking form.
There are items Anaconda may nonetheless add.
AI analysis and observability would give enterprises higher methods to measure the standard, price and reliability of agent habits. Mannequin and information lineage may present a verifiable report of the prompts, packages, datasets and fashions concerned in producing an software or determination. Agent identification and authorization would assist management which assets autonomous software program can entry and which actions it might carry out.
Safety is one other logical space. Immediate-injection defenses, mannequin scanning, secrets and techniques administration for brokers and runtime safety may all complement Anaconda’s software program provide chain place. Agent deployment, monitoring and rollback stay immature sufficient {that a} centered firm in manufacturing agent administration may fill a spot between Kilo’s improvement expertise and Outerbounds’ orchestration capabilities.
This doesn’t imply Anaconda will purchase an organization in every class. Nor ought to we flip the train right into a fantasy buying checklist stuffed with multibillion-dollar firms Anaconda is unlikely to accumulate. The extra believable targets can be centered developer-tool startups and open supply tasks that present essential expertise, neighborhood adoption and expertise.
The bigger level is that Anaconda could not imagine its platform is completed.
DeSanto will not be studying the platform-consolidation playbook on the job. He helped execute one at GitLab. Now he has acquired an organization co-founded by the individual with whom he helped construct that platform.
Is DeSanto making an attempt to construct the GitLab of enterprise AI improvement, with Python moderately than source-code administration as the unique beachhead?
It’s too early to make that declaration. However it’s not too early to ask the query.
Anaconda will not be abandoning Python. It’s utilizing Python as its beachhead, its supply of credibility and the muse beneath a bigger platform. Outerbounds prolonged that basis towards manufacturing. Kilo brings Anaconda into the agentic workspace. Future acquisitions could add extra of the management, safety and visibility required to attach the 2.
Python gave Anaconda its place within the AI stack. DeSanto is now attempting to climb that stack earlier than the worth and the developer relationship climbs away with out it.
The query is not whether or not Anaconda needs to be greater than the Python firm.
It’s how a lot of the AI improvement stack DeSanto finally intends to personal.
Regularly Requested Questions
What’s Kilo Code?
Kilo Code is an open-source, model-agnostic AI coding agent that works inside VS Code, JetBrains IDEs and the command line. It permits builders to route duties throughout a whole bunch of AI fashions moderately than being locked right into a single mannequin supplier.
Why did Anaconda purchase Kilo Code?
The acquisition provides Anaconda a place nearer to the developer’s day by day workflow. As a substitute of managing solely Python packages, dependencies and environments, Anaconda can now take part in mannequin choice, agent exercise, software connections and AI-generated software program improvement.
Is Anaconda attempting to turn into an enterprise AI platform?
The acquisition means that route. Anaconda now has capabilities spanning trusted packages and environments, workflow orchestration, manufacturing infrastructure and AI coding brokers. Nonetheless, turning these applied sciences right into a unified platform would require profitable integration, robust governance and a easy developer expertise.